how to write essay on passive voice

Passive Voice

What this handout is about.

This handout will help you understand what the passive voice is, why many professors and writing instructors frown upon it, and how you can revise your paper to achieve greater clarity. Some things here may surprise you. We hope this handout will help you to understand the passive voice and allow you to make more informed choices as you write.

So what is the passive voice? First, let’s be clear on what the passive voice isn’t. Below, we’ll list some common myths about the passive voice:

1. Myth: Use of the passive voice constitutes a grammatical error.

Use of the passive voice is not a grammatical error. It’s a stylistic issue that pertains to clarity—that is, there are times when using the passive voice can prevent a reader from understanding what you mean.

2. Myth: Any use of “to be” (in any form) constitutes the passive voice.

The passive voice entails more than just using a being verb. Using “to be” can weaken the impact of your writing, but it is occasionally necessary and does not by itself constitute the passive voice.

3. Myth: The passive voice always avoids the first person; if something is in first person (“I” or “we”) it’s also in the active voice.

On the contrary, you can very easily use the passive voice in the first person. Here’s an example: “I was hit by the dodgeball.”

4. Myth: You should never use the passive voice.

While the passive voice can weaken the clarity of your writing, there are times when the passive voice is OK and even preferable.

5. Myth: I can rely on my grammar checker to catch the passive voice.

See Myth #1. Since the passive voice isn’t a grammar error, it’s not always caught. Typically, grammar checkers catch only a fraction of passive voice usage.

Do any of these misunderstandings sound familiar? If so, you’re not alone. That’s why we wrote this handout. It discusses how to recognize the passive voice, when you should avoid it, and when it’s OK.

Defining the passive voice

A passive construction occurs when you make the object of an action into the subject of a sentence. That is, whoever or whatever is performing the action is not the grammatical subject of the sentence. Take a look at this passive rephrasing of a familiar joke:

Why was the road crossed by the chicken?

Who is doing the action in this sentence? The chicken is the one doing the action in this sentence, but the chicken is not in the spot where you would expect the grammatical subject to be. Instead, the road is the grammatical subject. The more familiar phrasing (why did the chicken cross the road?) puts the actor in the subject position, the position of doing something—the chicken (the actor/doer) crosses the road (the object). We use active verbs to represent that “doing,” whether it be crossing roads, proposing ideas, making arguments, or invading houses (more on that shortly).

Once you know what to look for, passive constructions are easy to spot. Look for a form of “to be” (is, are, am, was, were, has been, have been, had been, will be, will have been, being) followed by a past participle. (The past participle is a form of the verb that typically, but not always, ends in “-ed.” Some exceptions to the “-ed” rule are words like “paid” (not “payed”) and “driven.” (not “drived”).

Here’s a sure-fire formula for identifying the passive voice:

form of “to be” + past participle = passive voice

For example:

The metropolis has been scorched by the dragon’s fiery breath.

When her house was invaded, Penelope had to think of ways to delay her remarriage.

Not every sentence that contains a form of “have” or “be” is passive! Forms of the word “have” can do several different things in English. For example, in the sentence “John has to study all afternoon,” “has” is not part of a past-tense verb. It’s a modal verb, like “must,” “can,” or “may”—these verbs tell how necessary it is to do something (compare “I have to study” versus “I may study”). And forms of “be” are not always passive, either—”be” can be the main verb of a sentence that describes a state of being, rather than an action. For example, the sentence “John is a good student” is not passive; “is” is simply describing John’s state of being. The moral of the story: don’t assume that any time you see a form of “have” and a form of “to be” together, you are looking at a passive sentence.

Need more help deciding whether a sentence is passive? Ask yourself whether there is an action going on in the sentence. If so, what is at the front of the sentence? Is it the person or thing that does the action? Or is it the person or thing that has the action done to it? In a passive sentence, the object of the action will be in the subject position at the front of the sentence. As discussed above, the sentence will also contain a form of be and a past participle. If the subject appears at all, it will usually be at the end of the sentence, often in a phrase that starts with “by.” Take a look at this example:

The fish was caught by the seagull.

If we ask ourselves whether there’s an action, the answer is yes: a fish is being caught. If we ask what’s at the front of the sentence, the actor or the object of the action, it’s the object: the fish, unfortunately for it, got caught, and there it is at the front of the sentence. The thing that did the catching—the seagull—is at the end, after “by.” There’s a form of be (was) and a past participle (caught). This sentence is passive.

Let’s briefly look at how to change passive constructions into active ones. You can usually just switch the word order, making the actor and subject one by putting the actor up front:

The dragon has scorched the metropolis with his fiery breath.

After suitors invaded her house, Penelope had to think of ways to delay her remarriage.

To repeat, the key to identifying the passive voice is to look for both a form of “to be” and a past participle, which usually, but not always, ends in “-ed.”

Clarity and meaning

The primary reason why your instructors frown on the passive voice is that they often have to guess what you mean. Sometimes, the confusion is minor. Let’s look again at that sentence from a student’s paper on Homer’s The Odyssey:

Like many passive constructions, this sentence lacks explicit reference to the actor—it doesn’t tell the reader who or what invaded Penelope’s house. The active voice clarifies things:

After suitors invaded Penelope’s house, she had to think of ways to fend them off.

Thus many instructors—the readers making sense of your writing—prefer that you use the active voice. They want you to specify who or what is doing the action. Compare the following two examples from an anthropology paper on a Laotian village to see if you agree.

(passive)  A new system of drug control laws was set up. (By whom?)

(active)  The Lao People’s Revolutionary Party set up a new system of drug control laws.

Here’s another example, from the same paper, that illustrates the lack of precision that can accompany the passive voice:

Gender training was conducted in six villages, thus affecting social relationships.

And a few pages later:

Plus, marketing links were being established.

In both paragraphs, the writer never specifies the actors for those two actions (Who did the gender training? Who established marketing links?). Thus the reader has trouble appreciating the dynamics of these social interactions, which depend upon the actors conducting and establishing these things.

The following example, once again from that paper on The Odyssey, typifies another instance where an instructor might desire more precision and clarity:

Although Penelope shares heroic characteristics with her husband, Odysseus, she is not considered a hero.

Who does not consider Penelope a hero? It’s difficult to tell, but the rest of that paragraph suggests that the student does not consider Penelope a hero (the topic of the paper). The reader might also conceivably think that the student is referring to critics, scholars, or modern readers of The Odyssey. One might argue that the meaning comes through here—the problem is merely stylistic. Yet style affects how your reader understands your argument and content. Awkward or unclear style prevents your reader from appreciating the ideas that are so clear to you when you write. Thus knowing how your reader might react enables you to make more effective choices when you revise. So after you identify instances of the passive, you should consider whether your use of the passive inhibits clear understanding of what you mean.

Summarizing history or literary plots with the passive voice: don’t be a lazy thinker or writer!

With the previous section in mind, you should also know that some instructors proclaim that the passive voice signals sloppy, lazy thinking. These instructors argue that writers who overuse the passive voice have not fully thought through what they are discussing and that this makes for imprecise arguments. Consider these sentences from papers on American history:

The working class was marginalized. African Americans were discriminated against. Women were not treated as equals.

Such sentences lack the precision and connection to context and causes that mark rigorous thinking. The reader learns little about the systems, conditions, human decisions, and contradictions that produced these groups’ experiences of oppression. And so the reader—the instructor—questions the writer’s understanding of these things.

It is especially important to be sure that your thesis statement is clear and precise, so think twice before using the passive voice in your thesis.

In papers where you discuss the work of an author—e.g., a historian or writer of literature—you can also strengthen your writing by not relying on the passive as a crutch when summarizing plots or arguments. Instead of writing:

It is argued that… or  Tom and Huck are portrayed as… or  And then the link between X and Y is made, showing that…

you can heighten the level of your analysis by explicitly connecting an author with these statements:

Anderson argues that… Twain portrays Tom and Huck as… Ishiguro draws a link between X and Y to show that…

By avoiding passive constructions in these situations, you can demonstrate a more thorough understanding of the material you discuss.

Scientific writing

All this advice works for papers in the humanities, you might note—but what about technical or scientific papers, including lab reports? Many instructors recommend or even require the passive voice in such writing. The rationale for using the passive voice in scientific writing is that it achieves “an objective tone”—for example, by avoiding the first person. To consider scientific writing, let’s break it up into two main types: lab reports and writing about a scientific topic or literature.

Lab reports

Although more and more scientific journals accept or even prefer first-person active voice (e.g., “then we sequenced the human genome”), some of your instructors may want you to remove yourself from your lab report by using the passive voice (e.g., “then the human genome was sequenced” rather than “then we sequenced the human genome”). Such advice particularly applies to the section on Materials and Methods, where a procedure “is followed.” (For a fuller discussion on writing lab reports, see our handout on writing lab reports .)

While you might employ the passive voice to retain objectivity, you can still use active constructions in some instances and retain your objective stance. Thus it’s useful to keep in mind the sort of active verbs you might use in lab reports. Examples include: support, indicate, suggest, correspond, challenge, yield, show.

Thus instead of writing:

A number of things are indicated by these results.

you could write:

These results indicate a number of things . or Further analysis showed/suggested/yielded…

Ultimately, you should find out your instructor’s preference regarding your use of the passive in lab reports.

Writing about scientific topics

In some assignments, rather than reporting the results of your own scientific work, you will be writing about the work of other scientists. Such assignments might include literature reviews and research reports on scientific topics. You have two main possible tasks in these assignments: reporting what other people have done (their research or experiments) or indicating general scientific knowledge (the body of knowledge coming out of others’ research). Often the two go together. In both instances, you can easily use active constructions even though you might be tempted by the passive—especially if you’re used to writing your own lab reports in the passive.

You decide: Which of these two examples is clearer?

(passive) Heart disease is considered the leading cause of death in the United States.

or (active)  Research points to heart disease as the leading cause of death in the United States.

Alternatively, you could write this sentence with human actors:

Researchers have concluded that heart disease is the leading cause of death in the United States.

The last two sentences illustrate a relationship that the first one lacks. The first example does not tell who or what leads us to accept this conclusion about heart disease.

Here’s one last example from a report that describes angioplasty. Which sounds better to you?

The balloon is positioned in an area of blockage and is inflated. or The surgeon positions the balloon in an area of blockage and inflates it.

You can improve your scientific writing by relying less on the passive. The advice we’ve given for papers on history or literature equally applies to papers in more “scientific” courses. No matter what field you’re writing in, when you use the passive voice, you risk conveying to your reader a sense of uncertainty and imprecision regarding your writing and thinking. The key is to know when your instructor wants you to use the passive voice. For a more general discussion of writing in the sciences , see our handout.

“Swindles and perversions”

Before we discuss a few instances when the passive might be preferable, we should mention one of the more political uses of the passive: to hide blame or obscure responsibility. You wouldn’t do this, but you can learn how to become a critic of those who exhibit what George Orwell included among the “swindles and perversions” of writing. For example:

Mistakes were made.

The Exxon Company accepts that a few gallons might have been spilled.

By becoming critically aware of how others use language to shape clarity and meaning, you can learn how better to revise your own work. Keep Orwell’s swindles and perversions in mind as you read other writers. Because it’s easy to leave the actor out of passive sentences, some people use the passive voice to avoid mentioning who is responsible for certain actions.

So when is it OK to use the passive?

Sometimes the passive voice is the best choice. Here are a few instances when the passive voice is quite useful:

1. To emphasize an object. Take a look at this example:

One hundred votes are required to pass the bill.

This passive sentence emphasizes the number of votes required. An active version of the sentence (“The bill requires 100 votes to pass”) would put the emphasis on the bill, which may be less dramatic.

2. To de-emphasize an unknown subject/actor. Consider this example:

Over 120 different contaminants have been dumped into the river.

If you don’t know who the actor is—in this case, if you don’t actually know who dumped all of those contaminants in the river—then you may need to write in the passive. But remember, if you do know the actor, and if the clarity and meaning of your writing would benefit from indicating him/her/it/them, then use an active construction. Yet consider the third case.

3. If your readers don’t need to know who’s responsible for the action.

Here’s where your choice can be difficult; some instances are less clear than others. Try to put yourself in your reader’s position to anticipate how they will react to the way you have phrased your thoughts. Here are two examples:

(passive)  Baby Sophia was delivered at 3:30 a.m. yesterday.

and (active)  Dr. Susan Jones delivered baby Sophia at 3:30 a.m. yesterday.

The first sentence might be more appropriate in a birth announcement sent to family and friends—they are not likely to know Dr. Jones and are much more interested in the “object”(the baby) than in the actor (the doctor). A hospital report of yesterday’s events might be more likely to focus on Dr. Jones’ role.

Summary of strategies

• Look for the passive voice: “to be” + a past participle (usually, but not always, ending in “-ed”)
• If you don’t see both components, move on.
• Does the sentence describe an action? If so, where is the actor? Is the he/she/they/it in the grammatical subject position (at the front of the sentence) or in the object position (at the end of the sentence, or missing entirely)?
• Does the sentence end with “by…”? Many passive sentences include the actor at the end of the sentence in a “by” phrase, like “The ball was hit by the player ” or “The shoe was chewed up by the dog .” “By” by itself isn’t a conclusive sign of the passive voice, but it can prompt you to take a closer look.
• Is the doer/actor indicated? Should you indicate him/her/them/it?
• Does it really matter who’s responsible for the action?
• Would your reader ask you to clarify a sentence because of an issue related to your use of the passive?
• Do you use a passive construction in your thesis statement?
• Do you use the passive as a crutch in summarizing a plot or history, or in describing something?
• Do you want to emphasize the object?
• If you decide that your sentence would be clearer in the active voice, switch the sentence around to make the subject and actor one. Put the actor (the one doing the action of the sentence) in front of the verb.

Towards active thinking and writing

We encourage you to keep these tips in mind as you revise. While you may be able to employ this advice as you write your first draft, that’s not necessarily always possible. In writing, clarity often comes when you revise, not on your first try. Don’t worry about the passive if that stress inhibits you in getting your ideas down on paper. But do look for it when you revise. Actively make choices about its proper place in your writing. There is nothing grammatically or otherwise “wrong” about using the passive voice. The key is to recognize when you should, when you shouldn’t, and when your instructor just doesn’t want you to. These choices are yours. We hope this handout helps you to make them.

Works consulted and suggested reading

We consulted these works while writing this handout. This is not a comprehensive list of resources on the handout’s topic, and we encourage you to do your own research to find additional publications. Please do not use this list as a model for the format of your own reference list, as it may not match the citation style you are using. For guidance on formatting citations, please see the UNC Libraries citation tutorial . We revise these tips periodically and welcome feedback.

Anson, Chris M., and Robert A. Schwegler. 2010. The Longman Handbook for Writers and Readers , 6th ed. New York: Longman.

Baron, Dennis E. 1989. “The Passive Voice Can Be Your Friend.” In Declining Grammar and Other Essays on the English Vocabulary , 17-22. Urbana, IL: National Council of Teachers.

Hjortshoj, Keith. 2001. The Transition to College Writing . New York: Bedford/St Martin’s.

Lanham, Richard A. 2006. Revising Prose , 5th ed. New York: Pearson Longman.

Orwell, George. 1968. “Politics and the English Language.” In The Collected Essays, Journalism and Letters of George Orwell , edited by Ian Angus and Sonia Orwell, 4: 127-140. New York: Harcourt, Brace, Javanovich.

Rosen, Leonard J., and Laurence Behrens. 2000. The Allyn and Bacon Handbook , 4th ed. Boston: Allyn and Bacon.

Strunk, William, and E.B. White. 2000. The Elements of Style , 4th ed. Boston: Allyn and Bacon.

Trimble, John R. 2000. Writing With Style , 2nd ed. Upper Saddle River, NJ: Prentice Hall.

Williams, Joseph, and Joseph Bizup. 2017. Style: Lessons in Clarity and Grace , 12th ed. Boston: Pearson.

You may reproduce it for non-commercial use if you use the entire handout and attribute the source: The Writing Center, University of North Carolina at Chapel Hill

Make a Gift

• Writing Home
• Writing Advice Home

Passive Voice: When to Use It and When to Avoid It

• Printable PDF Version
• Fair-Use Policy

What is passive voice?

In English, all sentences are in either “active” or “passive” voice:

active: Werner Heisenberg formulated the uncertainty principle in 1927. passive: The uncertainty principle was formulated by Werner Heisenberg in 1927.

In an active sentence, the person or thing responsible for the action in the sentence comes first. In a passive sentence, the person or thing acted on comes first, and the actor is added at the end, introduced with the preposition “by.” The passive form of the verb is signaled by a form of “to be”: in the sentence above, “was formulated” is in passive voice while “formulated” is in active.

In a passive sentence, we often omit the actor completely:

The uncertainty principle was formulated in 1927.

When do I use passive voice?

In some sentences, passive voice can be perfectly acceptable. You might use it in the following cases:

The cave paintings of Lascaux were made in the Upper Old Stone Age. [We don’t know who made them.]

An experimental solar power plant will be built in the Australian desert. [We are not interested in who is building it.]

Mistakes were made. [Common in bureaucratic writing!]

Rules are made to be broken. [By whomever, whenever.]

Insulin was first discovered in 1921 by researchers at the University of Toronto. It is still the only treatment available for diabetes.

The sodium hydroxide was dissolved in water. This solution was then titrated with hydrochloric acid.

In these sentences you can count on your reader to know that you are the one who did the dissolving and the titrating. The passive voice places the emphasis on your experiment rather than on you.

Note: Over the past several years, there has been a movement within many science disciplines away from passive voice. Scientists often now prefer active voice in most parts of their published reports, even occasionally using the subject “we” in the Materials and Methods section. Check with your instructor or TA whether you can use the first person “I” or “we” in your lab reports to help avoid the passive.

When should I avoid passive voice?

Passive sentences can get you into trouble in academic writing because they can be vague about who is responsible for the action:

Both Othello and Iago desire Desdemona. She is courted. [Who courts Desdemona? Othello? Iago? Both of them?]

Academic writing often focuses on differences between the ideas of different researchers, or between your own ideas and those of the researchers you are discussing. Too many passive sentences can create confusion:

Research has been done to discredit this theory. [Who did the research? You? Your professor? Another author?]

Some students use passive sentences to hide holes in their research:

The telephone was invented in the nineteenth century. [I couldn’t find out who invented the telephone!]

Finally, passive sentences often sound wordy and indirect. They can make the reader work unnecessarily hard. And since they are usually longer than active sentences, passive sentences take up precious room in your paper:

Since the car was being driven by Michael at the time of the accident, the damages should be paid for by him.

Weeding out passive sentences

If you now use a lot of passive sentences, you may not be able to catch all of the problematic cases in your first draft. But you can still go back through your essay hunting specifically for passive sentences. At first, you may want to ask for help from a writing instructor. The grammar checker in your word processor can help spot passive sentences, though grammar checkers should always be used with extreme caution since they can easily mislead you. To spot passive sentences, look for a form of the verb to be in your sentence, with the actor either missing or introduced after the verb using the word “by”:

Poland was invaded in 1939, thus initiating the Second World War. Genetic information is encoded by DNA. The possibility of cold fusion has been examined for many years.

Try turning each passive sentence you find into an active one. Start your new sentence with the actor. Sometimes you may find that need to do some extra research or thinking to figure out who the actor should be! You will likely find that your new sentence is stronger, shorter, and more precise:

Germany invaded Poland in 1939, thus initiating the Second World War. DNA encodes genetic information. Physicists have examined the possibility of cold fusion for many years.

Have a language expert improve your writing

Run a free plagiarism check in 10 minutes, generate accurate citations for free.

• Knowledge Base
• Academic writing

Active vs. Passive Constructions | When to Use the Passive Voice

Published on June 29, 2023 by Shona McCombes . Revised on August 23, 2023.

The passive voice occurs when the person or thing that performs an action is not the grammatical subject of the sentence. Instead, the person or thing that receives the action is placed before the verb . Passive sentences are formed using the verb  to be combined with a past participle.

Active voice

The dog bites the bone.

Passive voice

The bone is bitten by the dog.

In a passive construction, the actor does not have to be named at all.

Passive construction

The bone is bitten .

Writers are often advised to avoid the passive voice, but it is not a grammatical error. In academic writing , this type of sentence structure is sometimes useful or necessary. However, overusing it can make your writing unclear or convoluted.

Instantly correct all language mistakes in your text

Upload your document to correct all your mistakes in minutes

Table of contents

Avoiding the passive voice, using the passive voice, other interesting articles.

In most cases, it’s best to use active sentence constructions where possible. Sometimes the passive voice makes a sentence less clear by obscuring the actor.

Who made the decision? To properly understand what occurred, we need to know who was behind the action. This is possible in the passive voice, but the sentence becomes convoluted.

An active construction is preferable for clarity and concision .

If you write a passive sentence, consider carefully whether leading with the actor would strengthen your point.

This is also relevant when discussing previous research: active constructions that specify who is responsible for findings can make your writing more credible and convincing.

• Evidence   has been found of nonhuman primates engaging in ritualistic behaviour.
• Smith (2015)  found  evidence of nonhuman primates engaging in ritualistic behaviour.
• Several recent studies have found evidence of nonhuman primates engaging in ritualistic behaviour.

Here's why students love Scribbr's proofreading services

Discover proofreading & editing

Some types of academic writing do not permit the use of first-person pronouns . In these cases, the passive voice can be used for referring to your own actions.

If you use the passive voice in more complex sentences, make sure to avoid dangling modifiers .

The passive voice is often also appropriate when the subject of an action is unknown or unimportant to the meaning of the sentence.

In this case, the object of the action – the votes – is more important than who did the counting. Specifying the actor wouldn’t add any useful information to the sentence.

With verbs like require , there is often no particular actor who does the requiring , so the passive voice is used to state a general sense of necessity or obligation.

If you want to know more about AI for academic writing, AI tools, or writing rules make sure to check out some of our other articles with explanations and examples or go directly to our tools!

 Academic Writing

• Avoiding repetition
• Effective headings
• Taboo words
• Transition words
• Deep learning
• Generative AI
• Machine learning
• Reinforcement learning
• Supervised vs. unsupervised learning

 (AI) Tools

• Grammar Checker
• Paraphrasing Tool
• Text Summarizer
• AI Detector
• Plagiarism Checker
• Citation Generator

Cite this Scribbr article

If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.

McCombes, S. (2023, August 23). Active vs. Passive Constructions | When to Use the Passive Voice. Scribbr. Retrieved June 9, 2024, from https://www.scribbr.com/academic-writing/passive-voice/

Is this article helpful?

Shona McCombes

Other students also liked, how to fix dangling and misplaced modifiers (with examples), what is a dangling modifier | examples & how to eliminate, what is a pronoun | definition, types & examples, what is your plagiarism score.

Passive Voice – Usage, Misuses & Worksheet

| Candace Osmond

Candace Osmond

Candace Osmond studied Advanced Writing & Editing Essentials at MHC. She’s been an International and USA TODAY Bestselling Author for over a decade. And she’s worked as an Editor for several mid-sized publications. Candace has a keen eye for content editing and a high degree of expertise in Fiction.

To be a successful writer in any sense, it’s important to understand how to correctly use the different parts of speech. I’ve learned this first-hand as a writer and published author for over a decade. One of these parts is the passive voice, which can often be misused. I see it all the time. My guide will discuss passive voice and how to use it correctly. I’ve also included a worksheet for you to practice using the passive voice. Let’s get started!

The Use of Passive Voice

There is much debate surrounding the use of passive voice in writing. Some believe that using passive voice is inherently bad and can even negatively impact readers’ comprehension. Others, however, argue that there are certain instances where passive voice can effectively make writing more engaging or interesting.

What Is Passive Voice and When to Use It?

Passive voice is a grammatical construction in which the sentence’s subject performs an action that is not done directly by that subject. In passive-voice sentences, the object of the action becomes the sentence’s subject.

This can be used as a stylistic choice in writing to emphasize specific elements of a story or argument or to make a description more concise and avoid sounding repetitive.

However, it can sometimes mask the true actor initiating that action, so it should be used carefully and only when appropriate. When deciding whether to use passive voice in your writing, consider the function of each sentence and what you’re attempting to achieve with your style and content.

How Do You Identify a Passive Voice?

Passive voice is a grammatical construction in which the subject receives an action. In standard English, this is often indicated by using a form of “to be” plus a past participle. For example, in the sentence “The ball was hit by Mary,” the passive voice uses “was hit” instead of “hit.”

Because passive voice sentences often emphasize certain elements over others and may need to be clearer than active voice sentences, it can sometimes be difficult to identify them.

One reliable way to identify passive voice is to look for forms of “to be” that are not in their simple conjugation or infinitive forms (am, is, etc.).

Seeing the word “by” paired with a noun or pronoun could indicate passive voice construction. Finally, if the sentence feels unnatural or awkward and does not clearly name an agent performing the action (also known as a subject), it could be constructed using passive voice.

Recognizing when something is written in passive voice takes some practice, but paying attention to word choice, tense agreement, and structure can help you make accurate grammatical judgments. I’m guilty of slipping into the passive voice from time to time, but I know how to spot it now and fix it to flow better.

Why Is Passive Voice Used?

Passive voice is often used in writing for a variety of reasons.

Firstly, it can be used to place greater emphasis on the reader or listener. By shifting the focus from the subject to the object, passive voice can help to highlight important information and draw attention to specific details.

Passive voice effectively creates a detached tone, which can help convey an unbiased point of view in academic or formal writing.

Finally, passive voice can enable the writer or speaker to evade responsibility by obscuring the agent’s identity when making action happen.

Here are some specific situations that call for passive voice use:

• When reporting incidents and crimes with unknown authors
• When the emphasis falls on the action and the subject is not important
• In scientific papers and research

Here are correct passive voice usage examples:

• Two cars were stolen last night.
• Lady Gaga was awarded as the best singer in 2011.
• The cure for cancer was discovered.

What Are the Rules for Passive Voice?

When writing or speaking in passive voice, the clause’s subject is the one receiving the verb’s action. The correct passive voice formula is subject + “to be” verb form + transitive verb in past participle + prepositional phrase (optional).

Passive voice examples:

• The test was completed by everyone in the class.
• This essay wasn’t written by any of our students.

Now Let’s See Them in Active Sentences

• Everyone in the class completed the test.
• None of our students wrote the essay.

See how the active sentence structure is more concise and direct. 

The Difference Between Active and Passive Voice

There is a common misconception that the difference between active and passive voices is simply whether a sentence contains an action or not. However, this is only sometimes the case. Rather, active and passive voice refers to relationships between subjects and objects in a sentence.

With active voice, the subject acts on the object directly; passive voice is the other way around. In general, the active voice tends to be clearer and more concise than the passive voice, making it a better choice for most types of writing.

This is because, in sentences using passive voice, there is often an abundance of unnecessary words and phrases that take away from the original meaning of the sentence.

For example:

• Active voice: I kicked the ball.
• Passive voice: The ball was kicked (by me).

Creative Ways to Use the Passive Voice in Writing

There are many clever ways of using the passive voice in writing , and each method can bring unique benefits to your work. For example, one technique is to use passives as nominalizations or nouns created from verbs. Doing so can create concise, action-oriented phrases that pack a lot of punch.

In addition, you could also try drawing comparisons between two unrelated objects or ideas through passive construction. This approach can help to make complex ideas more accessible and understandable for your readers.

Passive Voice Misuse

Passive voice is often misused in writing, either due to a poor understanding of what this technique entails or simply because it sounds better. However, to get the most out of passive voice, you must use it appropriately and with caution.

In general, passive voice is best used to emphasize the receiver or result of an action instead of the agent performing that action. It can also help to make simple instructions sound more sophisticated, yet using passive voice can be tricky when it comes to complex actions or stating opinions.

When used correctly and sparingly, however, passive voice can add a level of nuance and complexity to your writing that will keep your readers engaged and interested.

When Should I Avoid Passive Voice?

It all comes down to one question: is passive voice bad?

If you ask my editor, she’ll say yes, but I don’t always agree. There are many situations where it is preferable to use active voice instead of passive voice. Generally, active voice is more efficient and immediate, as it involves focusing on the actor and making it clear who or what is responsible for creating a particular situation.

Additionally, the active voice tends to be more engaging and exciting, drawing the reader in with its immediacy.

Therefore, when writing about a process or action that you want to appear direct and efficient, it is best to refrain from using passive voice whenever possible.

Weeding Out Passive Sentences

When it comes to writing, the use of passive voice sentences can often leave your work feeling flat and uninteresting. Passive voice is characterized by a lack of agency, with the subject of the sentence appearing as though it is being acted upon rather than acting or taking an active role in the action.

While there are certain situations where passive voice may be appropriate, it should be avoided. Weeding out passive verbs and replacing them with more direct and engaging verbs can help you to give your sentences a more active and dynamic feel.

Whether you are writing an academic paper or a social media post, paying attention to your use of passive voice can help you craft more powerful and interesting content.

Final Words

While people consider passive voice incorrect, it’s a matter of style and knowing when to use it. Active voice makes everything sound more natural, whereas passive voice requires specific uses. Forming the passive voice is simple as long as you remember to use the verb “to be” and a past participle. Make sense?

Grammarist is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com. When you buy via the links on our site, we may earn an affiliate commission at no cost to you.

2024 © Grammarist, a Found First Marketing company. All rights reserved.

What is Passive Voice? Definition, Examples of Passive Sentences in Writing

Home » The Writer’s Dictionary » What is Passive Voice? Definition, Examples of Passive Sentences in Writing

Passive voice definition: The passive voice is a style of writing where what would be the object of a sentence becomes the subject of the sentence.

What is Passive Voice?

What does passive voice mean? The passive writing voice occurs when something that is ordinarily “done by” the subject of a sentence is “done by” the object of a sentence .

In the passive writing voice, whatever is doing the action of the sentence is not the grammatical subject of the sentence.

Passive Voice Examples:

• The work was completed by Jaime.

Even though “Jaime” completed the work, “Jaime” is not the grammatical subject of this sentence. The subject of this sentence is “work.”

• The ball was hit by Johnny.

Again, in this sentence, rather than say, “Johnny hit the ball.” the ball becomes the subject of the sentence.

How is the Passive Voice Formed?

Passive construction: The passive writing voice is formed when what should be the object of a sentence becomes the subject of a sentence.

Examples of Passive Voice:

• Austin bought clothes.
• “Austin” is subject; “clothes” is object
• The clothes
• The clothes were bought.
• The clothes were bought by Austin.

Passive vs. Active Voice: What’s the Difference?

What is a passive sentence? The passive writing voice occurs when the action is done by what seems like it should be the subject.

Active Voice Example:

• Shakespeare wrote the play.

In this sentence, Shakespeare is “doing” the action of the sentence.

Active vs. Passive Voice Examples:

• Shakespeare wrote the play. (active)
• The play was written by Shakespeare. (passive)

The subject is typically clearer in active voice whereas in passive voice it may seem like the object is the subject.

More Examples:

• President Barack Obama signed a rescue package on Thursday for financially strapped Puerto Rico, which is facing more than $70 billion in debt and a major payment due Friday. – ABC News (active)
• A bill designed to reclaim businesses that have left the state and better Missouri’s port infrastructure was signed into law by Gov. Jay Nixon Tuesday. – The Missouri Times (passive)

Should You Avoid Passive Voice in Writing?

The passive voice is used less frequently in writing. However, it should not be avoided altogether.

Some phrases in English are always stated in the passive voice (i.e. The book was written by Herman Melville).

Stylistically, passive voice can be used as well. The passive voice can add style when:

• the writer wants the “punch” to be at the end of a sentence;
• the agent is unknown or unimportant (i.e. The person doing the action is unknown.);
• the writer wants to hide the agent’s identity

Many will say that the passive voice is not permitted in good writing. This is untrue. Good writers know how to delicately blend the passive writing voice with active sentences. Like all writing techniques, passive voice should be used with intention and purpose.

Passive Voice vs. Active Voice Exercises

• This episode was brought to you by Coca-Cola.
• Children danced in the halls to celebrate the last day of school.
• The china was made in Japan.
• We bought a new entertainment unit.
• The fossil has been discovered.

See Answers Below.

Summary: What is a Passive Voice?

Define passive voice: The definition of passive voice is when the recipient of the verb’s action becomes the subject of a sentence .

The passive voice is not common in writing. However, it is a stylistic choice that writers will occasionally use for effect.

The passive writing voice occurs when what would be the object of a sentence becomes the subject.

For more information on English voice, see our full article on the active voice.

• Walden University
• Faculty Portal

Scholarly Voice: Active and Passive Voice

Active and passive voice.

Active voice and passive voice are grammatical constructions that communicate certain information about an action. Specifically, APA explains that voice shows relationships between the verb and the subject and/or object (see APA 7, Section 4.13). Writers need to be intentional about voice in order to ensure clarity. Using active voice often improves clarity, while passive voice can help avoid unnecessary repetition.  

Active voice can help ensure clarity by making it clear to the reader who is taking action in the sentence. In addition, the active voice stresses that the actor (or grammatical subject) precedes the verb, again, putting emphasis on the subject. Passive voice construction leaves out the actor (subject) and focuses on the relationship between the verb and object.

The order of words in a sentence with active voice is subject, verb, object.

• Active voice example : I conducted a study of elementary school teachers.
• This sentence structure puts the emphasis of the sentence on the subject, clarifying who conducted the study. 
• Passive voice example : A study was conducted of elementary school teachers.
• In this sentence, it is not clear who conducted this study. 

Generally, in scholarly writing, with its emphasis on precision and clarity, the active voice is preferred. However, the passive voice is acceptable in some instances, for example:

• if the reader is aware of who the actor is;
• in expository writing, where the goal of the discussion is to provide background, context, or an in-depth explanation;
• if the writer wants to focus on the object or the implications of the actor’s action; or
• to vary sentence structure.  

Also, much like for anthropomorphism , different writing styles have different preferences. So, though you may see the passive voice used heavily in articles that you read for your courses and study, it does not mean that APA style advocates the same usage.

Examples of Writing in the Active Voice

Here are some examples of scholarly writing in the active voice:

• This is active voice because the subject in the sentence precedes the verb, clearly indicating who (I) will take the action (present).

Example : Teachers conducted a pilot study addressing the validity of the TAKS exam.

• Similarly, teachers (subject) clearly took the action (conducted) in this sentence.

Recognizing the Passive Voice

According to APA, writers should select verb tenses and voice carefully. Consider these examples to help determine which form of the verb is most appropriate:

Example : A study was conducted of job satisfaction and turnover.

• Here, it is not clear who did the conducting. In this case, if the context of the paragraph does not clarify who did the action, the writer should revise this sentence to clarify who conducted the study. 

Example : I conducted a study of job satisfaction and turnover.

• This revised sentence clearly indicates the action taker. Using “I” to identify the writer’s role in the research process is often a solution to the passive voice and is encouraged by APA style (see APA 7, Section 4.16).

Using the past tense of the verb “to be” and the past participle of a verb together is often an indication of the passive voice. Here are some signs to look for in your paper:

• Example : This study was conducted.
• Example : Findings were distributed.

Another indication of passive voice is when the verb precedes the actor in the sentence. Even if the action taker is clearly identified in a passive voice construction, the sentence is usually wordier. Making the actor the grammatical subject that comes before the verb helps to streamline the sentence.

• Issue : Though the verb and the actor (action taker) are clearly identified here, to improve clarity and word economy, the writer could place that actor, Rogers, before the verb.
• More concise active voice revision : Rogers (2016) conducted a study on nursing and turnover.  
• Issue : Here, the actor follows the verb, which reduces emphasis and clarity.
• This revised sentence is in the active voice and makes the actor the subject of the sentence.

Intentional Use of the Passive Voice

Sometimes, even in scholarly writing, the passive voice may be used intentionally and strategically. A writer may intentionally include the subject later in the sentence so as to reduce the emphasis and/or importance of the subject in the sentence. See the following examples of intentional passive voice to indicate emphasis:

Example : Schools not meeting AYP for 2 consecutive years will be placed on a “needs improvement” list by the State’s Department of Education.

• Here, all actors taking actions are identified, but this is in the passive voice as the State’s Department of Education is the actor doing the placing, but this verb precedes the actor. This may be an intentional use of the passive voice, to highlight schools not meeting AYP.
• To write this in the active voice, it would be phrased: “The State’s Department of Education will place schools not meeting AYP for 2 consecutive years on a “needs improvement” list. This sentence places the focus on the State’s Department of Education, not the schools.

Example : Participants in the study were incentivized with a $5 coffee gift card, which I gave them upon completion of their interview.

• As the writer and researcher, I may want to vary my sentence structure in order to avoid beginning several sentences with “I provided…” This example is written in the passive voice, but the meaning is clear.

Using Passive Voice in Scholarly Writing

As noted before, passive voice is allowed in APA style and can be quite appropriate, especially when writing about methods and data collection. However, students often overuse the passive voice in their writing, which means their emphasis in the sentence is not on the action taker. Their writing is also at risk of being repetitive. Consider the following paragraph in which the passive voice is used in each sentence:

A survey was administered . Using a convenience sample, 68 teachers were invited to participate in the survey by emailing them an invitation. E-mail addresses of teachers who fit the requirements for participation were provided by the principal of the school . The teachers were e-mailed an information sheet and a consent form. Responses were collected from 45 teachers… As you can see, the reader has no idea who is performing these actions, which makes the research process unclear. This is at odds with the goal of the methods discussion, which is to be clear and succinct regarding the process of data collection and analysis.

However, if translated entirely to the active voice, clearly indicating the researcher’s role, “I” becomes redundant and repetitive, interrupting the flow of the paragraph:

In this study, I administered a survey. I created a convenience sample of 68 teachers. I invited them to participate in the survey by emailing them an invitation. I obtained e-mail addresses from the principal of the school… “I” is quite redundant here and repetitive for the reader.

The Walden Writing Center suggests that students use “I” in the first sentence of the paragraph . Then, as long as it is clear to the reader that the student (writer) is the actor in the remaining sentences, use the active and passive voices appropriately to achieve precision and clarity (where applicable):

In this study, I administered a survey using a convenience sample. Sixty-eight teachers were invited to participate in the survey. The principal of the school provided me with the e-mail addresses of teachers who fit the requirements for participation. I e-mailed the teachers an information sheet and a consent form. A total of 45 teachers responded …

The use of the passive voice is complicated and requires careful attention and skill. There are no hard-and-fast rules. Using these guidelines, however, should help writers be clearer and more engaging in their writing, as well as achieving the intended purposes.

Remember, use voice strategically. APA recommends the active voice for clarity. However, the passive voice may be used, with intention, to remove the emphasis on the subject and also as a method for varying sentence structure. So, generally write in the active voice, but consider some of the above examples and some uses of the passive voice that may be useful to implement in your writing. Just be sure that the reader is always aware of who is taking the action of the verb.

• For more practice, try our Clarifying the Actor module .

Related Resources

• Principles of Writing: Active and Passive Voice (blog post) APA Style Blog post.

Didn't find what you need? Email us at [email protected] .

• Previous Page: Avoiding Bias
• Next Page: Word Choice (Diction)
• Office of Student Disability Services

Walden Resources

Departments.

• Academic Residencies
• Academic Skills
• Career Planning and Development
• Customer Care Team
• Field Experience
• Military Services
• Student Success Advising
• Writing Skills

Centers and Offices

• Center for Social Change
• Office of Academic Support and Instructional Services
• Office of Degree Acceleration
• Office of Research and Doctoral Services
• Office of Student Affairs

Student Resources

• Doctoral Writing Assessment
• Form & Style Review
• Quick Answers
• ScholarWorks
• SKIL Courses and Workshops
• Walden Bookstore
• Walden Catalog & Student Handbook
• Student Safety/Title IX
• Legal & Consumer Information
• Website Terms and Conditions
• Cookie Policy
• Accessibility
• Accreditation
• State Authorization
• Net Price Calculator
• Contact Walden

Walden University is a member of Adtalem Global Education, Inc. www.adtalem.com Walden University is certified to operate by SCHEV © 2024 Walden University LLC. All rights reserved.

• Get started with computers
• Learn Microsoft Office
• Apply for a job
• Improve my work skills
• Design nice-looking docs
• Getting Started
• Smartphones & Tablets
• Typing Tutorial
• Online Learning
• Basic Internet Skills
• Online Safety
• Social Media
• Zoom Basics
• Google Docs
• Google Sheets
• Career Planning
• Resume Writing
• Cover Letters
• Job Search and Networking
• Business Communication
• Entrepreneurship 101
• Careers without College
• Job Hunt for Today
• 3D Printing
• Freelancing 101
• Personal Finance
• Sharing Economy
• Decision-Making
• Graphic Design
• Photography
• Image Editing
• Learning WordPress
• Language Learning
• Critical Thinking
• For Educators
• Translations
• Staff Picks
• English expand_more expand_less

Writing Skills  - Active vs. Passive Voice in Your Writing

Writing skills  -, active vs. passive voice in your writing, writing skills active vs. passive voice in your writing.

Writing Skills: Active vs. Passive Voice in Your Writing

Lesson 5: active vs. passive voice in your writing.

/en/writing-skills/how-to-write-an-effective-business-email/content/

Active vs. passive voice in your writing

Have you ever noticed how some parts of your writing seem to pop, while other parts don't? You can improve those dull sentences if you take a moment to consider the active and passive voices.

Learn more about the differences between active and passive voice in the video below.

The difference between active and passive

There are two voices in writing: Active and passive. In the active voice, the subject of a  sentence acts, like "Neil Armstrong walked on the moon." The active voice is direct, clear, and easy to read.

With the passive voice, the subject is acted upon, like "The moon was walked on by Neil Armstrong". Although the passive voice is still grammatically correct, it typically doesn't carry the same energy or clarity as the active voice. Its structure can feel clumsy and unnatural, which makes your writing harder to read. It also tends to use more words than the active voice. Over the course of a document, all those extra words can make your writing drag.

Overall, we recommend using the active voice more often than the passive. This will help keep your writing snappy and efficient.

Identifying the passive voice

Here's how to spot the passive voice:  

First, look for a phrase like "was visited", "has been cleaned", or "will be built".  Each one contains a "to be" verb , like "was", "has been", or "will be". 

That phrase is followed by an action that's already happened , like "visited", "cleaned", or "built".  Finally, the person or thing doing the action comes last , if they're mentioned at all. 

If you see these parts together, there's a good chance the sentence is in passive voice .

Changing passive into active

Let's change a sentence from passive into active voice.

Our sentence is, "The money was tossed into the air by Jacob". "Jacob" is our subject, and "tossed" is the verb. Move Jacob to the beginning of the sentence, cut out any unnecessary words, and rearrange a few others. Our passive example is now, "Jacob tossed the money into the air". The delivery is more brief, clear, and more immediate.

When passive is best

Although active voice is incredibly useful, the passive voice is occasionally the better choice. For instance, you may go passive if the actor of a sentence is unknown or irrelevant, like in the sentence, "The amendments will be approved after a discussion". In this case, we're interested in the amendments' approval, not who approved them.

Passive voice is also great for creating an authoritative tone , like on a sign requiring employees to wash their hands. It doesn't matter who requires employees to wash up; they just need to do it!

You may also want to go passive when you don't know who is responsible for the action , like in this example: "The mystery was never solved." 

The voice you use can make a big difference in your writing. The active voice will often add pep and clarity, but occasionally the passive voice will be your best option. Take some time to choose the voice that fits best, and your writing will almost certainly grow stronger.

• How It Works
• Prices & Discounts

Active Voice vs Passive Voice in Essay Writing: What's the Difference?

Table of contents

Every type of writing, spanning academic assignments, research proposals, movie or book reviews, newspaper articles, technical or scientific writing, and more, requires a verb in the sentence to express an action being taken.

Essentially, we know that there are two types of voices in writing – active voice vs. passive voice. Both voices have a different sentence structure, length, purpose, and tone of writing.

Now when you analyze your writing, you would be able to find specific sentences that pop out and leave a mark on the reader while some sentences remain bland and unengaging.

This will determine your active voice sentences and your passive voice sentences.

You think to yourself, “how do I choose the right voice for my writing?”

What is Active Voice in Essay Writing

In a sentence, the active voice is used when the subject or person in this specific sentence is the one who is carrying out an action that was represented by the verb. The subject is always a noun or a pronoun and this voice is used to express information in a stronger, more direct, clear, and easier-to-read way than passive voice sentences.

Active voice highlights a logical flow to your sentences and makes your writing feel alive and current – which is pivotal to use in your formal academic writing assignments to get top-scoring grades.

What is Passive Voice in Essay Writing

The passive voice, in a sentence, is used to emphasize the action taken place by the subject according to the verb. In this, the passive phrase always contains a conjugated form of ‘to be’ and the past participle of the main verb.

Due to this, passive sentences also include prepositions, which makes them longer and wordier than active voice sentences.

Active Voice vs. Passive Voice

Now, let us understand the difference between active voice vs. passive voice in writing.

The choice between using active voice vs. passive voice in writing always comes down to the requirements that are suitable for the type of sentences you choose to write.

For most of the writing that you do, be it blogs, emails, different types of academic essays, and more, an active voice is ideal to use for communicating and expressing your thoughts, facts, and ideas more clearly and efficiently. This way, your essay papers or other academic assessments stand out amongst the rest.

Use your judgment to write in an active voice if accuracy is not an important aspect, and always keep your readers in mind. In this case, academic writing teachers - ranging from middle school to college/universities, prefer reading your assignments in an active voice as it makes your arguments, thoughts, and sentence structures confident, brief, and compelling.

However, there are a few exceptions to using passive voice

• If the reader is aware of the subject;
• In expository writing (where the primary goal is to provide an explanation or a context);
• Crime reports, data analysis;
• Scientific and technical writing.

Passive voice is majorly used while writing assignments that direct the reader's focus onto the specific action taking place rather than the subject. It is also used when you need an authoritative tone, like on a banner or a sign on a bulletin board.

Passive voice is ideally used when the person involved in the action is not known and/or is insignificant. Similarly, if you are writing something that requires you to be objective with its solution and analysis – like a research paper, lab report , or newspaper article – using passive voice should be your go-to choice. This allows you to avoid personal pronouns, which in turn, helps you present your analysis or information in an unbiased and coherent way.

However, if your writing is meant to engage your target audience, such as a novel, then writing your sentences with a passive voice will not only flatten your content and make your writing clumsy to read, but your paper would also inherit all the extra words that would make your write-ups vague and too wordy.

2. Examples of active and passive voice

Every active voice sentence contains a form of action that is taking place by the subject. An interesting fact is that they can be written in any tense – past tense, present tense, past perfect tense, future tense, and more.

An active voice always emits a sense of agency and strength in your writing.

Here are some examples of active voice sentence structuring

• Kaitlyn worked on her upcoming novel all day long.
• Our professor will reveal this week’s surprise assignment.
• The police know that the accused is a flight risk.
• A baby monkey bit Sasha on her leg.
• I presented my research thesis to the class.
• Malek proposed the methods & principles by which each product could be analyzed.
• We will ride a train to go to Switzerland.
• I conducted a study of criminal psychology.
• The gardener was planting the Hydrangeas.

Whereas passive voice sentence structures are lengthier in words and are used when the subject or person is the recipient of an action. Passive voice in writing often conveys subtlety, submissiveness, and lack of engagement.

Moreover, just like active sentences, passive sentence structures do not need to be dependent on the verb, as they can occur in the past tense, the future tense, the subjunctive, etc.

Here are some examples of passive-voice sentences

• Our professor drove us to Universal Studios.
• Clara was persuaded to move to Toronto.
• Jack was given two choices for the presentation topic.
• The jobs were given to two people who had no experience in writing.
• An old bike and a gun were found in the toolshed.
• The moon was walked by Neil Armstrong.
• The candy was eaten by the lady in yellow.
• Ballet dancing is a beloved activity in our class.
• The concert will be enjoyed by us tomorrow.
• Some new books were bought by me.

3. Changing passive voice to active voice

Unless you’re required to use passive voice, it is always beneficial to use active voice in your writing. That’s why overusing and misusing passive sentences can make your writing look sloppy, wordy and non-informative and you may even end up with more grammatical errors.

Here are a few ways to change your passive sentence into an active sentence

a. Identify the passive voice

In writing, the writer should choose their verb tenses, word choice, and tonality of the content very carefully.

As you finish your draft, re-read it to identify sentences that could have been more concise, or framed in a better way to improve its readability. Ask yourself what the action of the specific sentence, who is perpetrating this action is.

That is your passive voice.

Passive voice or tone consists of a past, past participle or future tense and generally the auxiliary word ‘to be’ is an indication of a passive sentence. It always refers to action not being addressed directly.

b. Remove the auxiliary verb

It's best to remove the auxiliary verb from your sentences to change it into active voice sentences by adjusting the tense of the main verb. Generally, the tense of the main verb is in the past tense.

So, determine the correct tense and use it in your verb to create an active sentence. This in turn delivers your writing in a more clear, strong, concise and urgent way.

Here’s an interesting video by mmmEnglish that explains what auxiliary verbs are.

c. Change the subject of the sentence

The main difference between active voice and passive voice is that one performs a verb and the other is a recipient of an action.

For example, in a passive sentence, “The novel was drafted by the writer”, the ‘novel’ is the subject which had been actioned by the writer.

To change it into an active voice, restructure the verb that is taking place (drafted by the writer) with the subject (novel), thus structuring “The writer drafted the novel”.

Once you have mastered the technique of identifying the voice and tonality, you will discover the ease with which your communication takes on different textures, depending upon the context at hand. While the active voice remains the direct form of communication and has more mass appeal, it is the passive voice that assumes a less biased and more objective tone.

Make sure to embellish your written expression with the right voice and give it the power and authority it deserves. We hope the tips and suggestions given above will go a long way in giving weightage to every sentence that you write and strike the right chord with readers.

When you work with Writers Per Hour, you’ll be happy to know that our professional team of writers knows when to use active and passive voice correctly, which works best for the type of paper.

If you’re running short of time or are not confident about your English writing skills, reach out to us, and we’ll ensure you receive nothing short of professionally written, high-quality papers.

Share this article

Achieve Academic Success with Expert Assistance!

Crafted from Scratch for You.

Ensuring Your Work’s Originality.

Transform Your Draft into Excellence.

Perfecting Your Paper’s Grammar, Style, and Format (APA, MLA, etc.).

Calculate the cost of your paper

Get ideas for your essay

• Features for Creative Writers
• Features for Work
• Features for Higher Education
• Features for Teachers
• Features for Non-Native Speakers
• Learn Blog Grammar Guide Community Events FAQ
• Grammar Guide

Free Passive Voice Checker

Our passive voice detector automatically highlights passive voice in your writing and suggests stronger, active alternatives.

Start typing, paste, or use

Get more suggestions to enhance this text and all your future writing

Your suggestions will show once you've entered some text.

Great job! We didn't find any suggestions in your text.

Feel more confident in everything you write

Join over 3 million authors, students, and professionals improving their writing.

I am continually impressed with the positive input this program offers me every time I sit down to write. My skills have improved immensely since I bought it and heartily recommend it to anyone who wants to have more confidence in their own writing.

Ginger Wakem

I’ve tried every free and paid writing/editing/grammar extension out there and this by far is the best one my team and I have found. It’s fast, accurate and really helps improve your writing beyond simple grammar suggestions.

Joel Widmer

ProWritingAid has been a resource in my writer toolkit for many years. The program helps me to craft and clarify my stories for a better reader experience. Your editor will thank you for making their job easier.

Siera London

Power up your writing with ProWritingAid

Get an editing tool that goes beyond being a passive voice checker, with real-time suggestions to improve spelling, grammar, style, and more.

Check for passive voice

ProWritingAid’s passive voice checker highlights passive voice in your writing and gives you active voice options that keep your sentences clear and compelling.

Avoid grammar and spelling errors

Whether you’re working on a short email or a full-length novel, ProWritingAid will catch spelling and grammar errors as you write so that no pesky mistakes slip through.

Get real-time style suggestions

ProWritingAid’s in-depth writing reports give you real-time style suggestions, such as avoiding overused words and varying your sentence lengths.

Works wherever you do

Check for passive voice across all the apps you use, whether they’re on desktop or online.

What type of content needs a passive voice checker?

You should check any text that will be read by someone else.

• cover letters
• marketing copy
• social posts
• lab reports
• assignments
• applications

ProWritingAid’s Passive Voice Checker FAQs

1. what is passive voice.

Active voice and passive voice are two different grammatical voices in the English language. When a sentence is written in active voice, the subject performs the verb. When a sentence is written in passive voice, the subject gets acted upon by the verb. For example, “I’m writing a novel” is in active voice, whereas “A novel is being written by me” is in passive voice.

2. Why should I avoid passive voice?

Passive voice isn’t a grammatical error, but it’s typically weaker than active voice. Sentences written in passive voice tend to feel unnecessarily convoluted or indirect. Unless you have a strong stylistic reason for using the passive voice, it’s better to write your sentences using the active voice.

3. How does ProWritingAid’s passive voice detector beat Grammarly’s?

With 27 different writing reports, ProWritingAid offers a more detailed analysis of your writing than other passive voice checkers on the market, such as Grammarly. Plus, ProWritingAid’s premium package is cheaper and more affordable than Grammarly’s.

4. Can I detect passive voice in email? And on social media?

Yes! You can use our browser extensions (Chrome, Safari, Firefox, and Edge) to use our passive vs active voice checker on nearly every website out there, like Facebook, Twitter, and Medium, as well as on web-based email providers like Gmail and Yahoo.

Feel confident in everything you write

Drop us a line or let's stay in touch via :

• PRO Courses Guides New Tech Help Pro Expert Videos About wikiHow Pro Upgrade Sign In
• EDIT Edit this Article
• EXPLORE Tech Help Pro About Us Random Article Quizzes Request a New Article Community Dashboard This Or That Game Popular Categories Arts and Entertainment Artwork Books Movies Computers and Electronics Computers Phone Skills Technology Hacks Health Men's Health Mental Health Women's Health Relationships Dating Love Relationship Issues Hobbies and Crafts Crafts Drawing Games Education & Communication Communication Skills Personal Development Studying Personal Care and Style Fashion Hair Care Personal Hygiene Youth Personal Care School Stuff Dating All Categories Arts and Entertainment Finance and Business Home and Garden Relationship Quizzes Cars & Other Vehicles Food and Entertaining Personal Care and Style Sports and Fitness Computers and Electronics Health Pets and Animals Travel Education & Communication Hobbies and Crafts Philosophy and Religion Work World Family Life Holidays and Traditions Relationships Youth
• Browse Articles
• Learn Something New
• Quizzes Hot
• This Or That Game
• Train Your Brain
• Explore More
• Support wikiHow
• About wikiHow
• Log in / Sign up
• Education and Communications
• English Grammar

How to Avoid Using the Passive Voice

Last Updated: September 30, 2022 Fact Checked

This article was co-authored by Christopher Taylor, PhD . Christopher Taylor is an Adjunct Assistant Professor of English at Austin Community College in Texas. He received his PhD in English Literature and Medieval Studies from the University of Texas at Austin in 2014. There are 7 references cited in this article, which can be found at the bottom of the page. This article has been fact-checked, ensuring the accuracy of any cited facts and confirming the authority of its sources. This article has been viewed 221,942 times.

Passive voice is usually weaker writing since it makes it unclear who or what is doing the action. Do you find yourself using the passive voice too frequently in your writing? Do you want to make your writing active and stronger? A few simple rules will help you avoid using the passive voice. This will turn you into a stronger writer overnight.

Identifying the Different Parts of a Sentence

• The verb is the action word in the sentence, and it can be either active or passive. Most verbs are active, like kicked or graded, but the verb phrase "to be" and its conjugations (is, was) are passive. [1] X Research source
• For example, the sentence: Mark kicked the ball. The verb in the sentence is kicked. In this sentence it's past tense. It's the action. The act of kicking. The ball is passive because it's not doing anything; it's just being kicked. [2] X Research source

• The subject of the sentence in the active voice is who or what did the action. Let's go back to our example sentence: Mark kicked the ball. Kicked, as we have already established, is the verb. Ask yourself, who or what did the kicking? Answer: Mark. That's how you determine the subject in the active voice.
• Recognize that sometimes the subject is not a person. Example: The plane stopped the traffic. The action, and thus, the verb, is the act of stopping. Who or what did the stopping? The plane. Thus, the plane is the subject of the sentence.

• In the example, Mark kicked the ball, the word ball is the direct object. The ball isn't active. It's not the catalyst for movement. It's being acted upon. It's the direct object.
• Sometimes, people eliminate the direct object and just write something like, Mark kicked. Sometimes in the passive voice, the previous direct object becomes the subject: The ball was kicked.

Determining Active Versus Passive Voice

• We've established that the verb is the action word in the sentence. Make sure the “who or what did the action” is actually performing the action.
• The dog wagged its tail is active voice. That's because the act of wagging is the verb. Who or what did the wagging? The dog. The dog in the sentence is performing the verb, wagging. Thus, the sentence is in active voice.

• For example, The tail was wagged by the dog is passive voice. This is because the tail, the subject, is being acted upon by the verb.
• Sometimes people eliminate the prepositional phrase describing who the action was being done by, and write a sentence like: The tail was wagged. This is often done to eliminate responsibility or when people don't know who did the action.

Knowing When to Use Passive Voice

• Let's say you're a news writer. The mayor was arrested by the police. Active voice would be: The police arrested the mayor. The more important point in the sentence is to announce that the mayor, a very prominent person, was arrested! Thus, in this case, it's logical to write instead: The mayor was arrested by police.
• Similarly, in science, it might make more sense to put the object first, not the process. Instead of writing, I poured the hydrogen into the beaker, you might write, the hydrogen was poured into the beaker.
• Generally speaking, though, active voice is better because it's tighter and punchier writing. It's more dramatic, and it puts responsibility where it belongs.

• Get in the habit of re-reading and checking over all your work. Look at each sentence individually.
• Don't be afraid to ask for help. Ask an English teacher or professor to explain the difference between the active and passive voices. Get a grammatically-adept friend to help you look over your work. Don't be afraid to let others help you!

• The papers will be graded by the teachers is passive (it's future tense). The papers will have been graded by the teachers is also passive (it's future perfect). The papers are being graded by the teachers is present progressive tense and also passive.
• Again, the key point is that all of these sentences are passive because the direct object (papers) is in front of the verb, and the subject (who did the action) is located after the verb. The active voice version of this sentence is: The teachers graded the papers.

Expert Q&A

• Politicians and PR operatives sometimes use passive voice to confuse or eliminate responsibility. Thanks Helpful 5 Not Helpful 1
• It's possible to use first-person in the passive voice. I was hit by the ball is still passive because the person, the “I,” is not active in this sentence. Thanks Helpful 0 Not Helpful 0

You Might Also Like

• ↑ https://www.mckendree.edu/offices/writing-center/resources/grammar/passive-verbs/index.php
• ↑ https://www.american.edu/provost/academic-access/upload/active-and-passive-verbs.pdf
• ↑ https://owl.purdue.edu/owl/general_writing/academic_writing/active_and_passive_voice/changing_passive_to_active_voice.html
• ↑ https://webapps.towson.edu/ows/activepass.aspx
• ↑ https://writing.wisc.edu/handbook/style/ccs_activevoice/
• ↑ http://writingcenter.unc.edu/handouts/passive-voice/
• ↑ https://academicguides.waldenu.edu/writingcenter/scholarlyvoice/activepassive

About This Article

To avoid using the passive voice, make sure the subject of your sentence is doing the action. For example, "The tail was wagged by the dog" is passive, but "The dog wagged its tail" isn't because the subject of the sentence, the dog, is doing the action. In general, watch out for sentences where the subject is preceded by the prepositional phrase "by the" or "by a," which means the sentence was written using the passive voice. To learn how to tell when it's OK to use the passive voice, scroll down! Did this summary help you? Yes No

• Send fan mail to authors

Reader Success Stories

Jun 14, 2017

Did this article help you?

Aubrey Wright

Apr 27, 2017

Yagami Light

Jun 24, 2017

Apr 26, 2017

Dante Tomaino

Oct 5, 2017

Featured Articles

Trending Articles

Watch Articles

• Terms of Use
• Privacy Policy
• Do Not Sell or Share My Info
• Not Selling Info

Don’t miss out! Sign up for

wikiHow’s newsletter

• Order Content
• Write Content

Find out the latest developments here on our blog.

How to write in a passive voice.

posted by iWriter user iWriter Team

December 21, 2020

In order to write effective copy , you need to understand the core mechanics of grammar and sentence structure. A common issue that writers encounter involves the use of passive voice, and many people aren’t sure how to use it appropriately.

Not quite sure how to begin? We’ve got all the info you need.

Let’s explore everything you should keep in mind about how to write in a passive voice.

What Exactly Is Passive Voice?

As the name suggests, passive voice involves phrasing your sentence in a way that isn’t quite as strong or direct as ‘active voice.’ In general, the subject of a sentence in passive voice has something done to it as opposed to the subject performing the action.

Let’s take a look at a variation of the same sentence in both active voice and passive voice.

• Active:  Mom slammed the door.
• Passive:  The door was  slammed by Mom.

We previously established that, in the passive voice, the subject is the part of the sentence that the verb affects. This means that ‘Mom’ is no longer relevant once the sentence is changed from active to passive.

So, you could even phrase the sentence like this:

“The door was slammed.”

In this scenario, the above sentence in its ‘purest’ form of passive voice isn’t very interesting. It also doesn’t tell the reader the circumstances, such as that Mom was the one who slammed the door.

How Can I Recognize a Sentence in Passive Voice?

In general, it’s fairly easy to discern when you’re reading a sentence in the passive voice. More often than not, it will sound overly ‘wordy’ or awkward. But, it’s not always this obvious in every case.

You can quickly figure out whether or not a sentence is in passive voice by looking for specific keywords. Some of these include:

More often than not, the above words are used when writing a sentence in the passive voice. A few examples are:

• The ball was hit by Brady.
• The candy was eaten.
• Each year, food is given to the homeless.

Of course, you’ll still need to use your best judgment— not every passive sentence is quite this obvious.

Why Do People Advise Against Using Passive Voice?

When used incorrectly, passive voice easily detracts from the overall clarity of the sentence. As we saw with the sentence about slamming the door, the passive voice version removed the subject ‘Mom’ entirely.

In many scenarios, it could be difficult to determine  who slammed the door. Let’s give that particular sentence more context.:

Due to the overwhelming amounts of homework David had to deal with, he didn’t complete the chores like his mother had asked him to. When she arrived home at 6 PM, she entered his bedroom and the two got into a heated argument. The door was slammed .

As you can see, it’s unclear whether David or his mother was the one who slammed the door, resulting in a loss of overall clarity.

In copywriting , it’s best to do all that you can in order to establish a high-quality experience for the reader. This means that you should generally avoid using the passive voice  unless you’re able to do so effectively.

But, that doesn’t mean that you can’t ever use it.

When Is It Appropriate to Use?

There are a handful of situations where using passive voice can actually be more effectiveas opposed to active voice.

For example, it’s particularly useful for placing less emphasis on irrelevant actors . In turn, this places more emphasis on the subject of the sentence.

Example:  Over 100 fires were ignited last week throughout the entire state, and law enforcement is struggling to find an answer.

In this scenario, the perpetrator(s) who started the fires haven’t yet been identified, so the main focus is on the fires. The same use of passive voice is applicable in sentences where the reader doesn’t need to know who’s responsible for the action.

Passive voice can also be sued to  emphasize a particular object.

Example:  The clock displayed that there were 60 seconds of time left, and only two more points were needed to win the game.

PAssive voice helps make the scenario much more captivating to the reader.

How Can I Use It Creatively?

You can easily use passive voice to help deflect blame that’s associated with a statement. Unfortunately, this is something we often see with politicians.

Instead of saying “They were responsible for the accident,” you could say “an accident occurred.” This is ideal for acknowledging situations where you don’t want to place the blame directly on another party.

As previously mentioned, you can also use passive voice creatively in order to emphasize certain parts of a sentence. A common example would be when writing about an incident that resulted in the injury or death of another individual.

It’s far more common to write “(celebrity) was killed in a plane crash” than “there was a plane crash earlier this afternoon that killed (celebrity) and five others.”

As long as using passive voice provides extra value to your writing, it’s more than likely an appropriate option.

Understanding How to Write in a Passive Voice Can Seem Complicated

But the above information will make the process far smoother. From here, you’ll be able to know how to write in a passive voice and how to recognize the most appropriate times to do so.

Want to learn more about what we have to offer? Feel free to reach out to us today and see how we can help.

Comments are closed.

Free Lessons

Popular Post

• Creating a Style Guide 101: How to Improve Your Content and Branding
• FreeUp & iWriter: How to Take Your Content Further With FreeUp
• How a Free E-Book Can Boost Your Marketing Efforts
• How to Create an E-Book for Your Business

How to Bypass GPT Zero AI Content Detection?

• Smodin Editorial Team
• Updated: May 9, 2024

Bypassing GPTZero can be challenging but not impossible.

It involves turning AI-generated text into content that sounds more human-like. Some of these steps include using Smodin’s AI Rewriter tool, writing in a unique tone of voice (TOV), adopting active voice over passive voice, and sticking to good copywriting best practices.

We all know AI writing tools like ChatGPT tend to have the “same” TOV, sentence structure, and predictable flow. And this is often why (if you directly copy and paste from these platforms) your writing is given a high AI score.

AI detection tools like GPTZero have become invaluable in differentiating between AI-generated and human-written text and are often used in educational settings to check for plagiarism and copywriting infringements.

Fortunately, there are ways to bypass GPTZero’s AI detection tools, such as the Smodin AI Rewriter.

In this article, we’ll uncover how to bypass GPTZero AI detection and how to tweak your writing so you can create content that resonates with your intended audience. But before we begin, let’s take a closer look at GPTZero and how its artificial intelligence detection mechanisms work.

What Is GPTZero?

GPTZero was launched in January 2023 by Edward Tian, a Princeton University undergraduate student, to address the growing concerns about AI-written text in educational settings.

GPTZero is artificial intelligence detection software that detects AI-generated content primarily generated from large language models (LLMs) like ChatGPT.

Used by students, educational facilities, and business owners, it helps content creators make their copy sound more human and reduces the risk of being flagged for plagiarism or infringing copywriting laws.

How Does It Detect AI-Generated Content?

GPTZero scans a document to identify common AI patterns using natural language processing (NLP).

AI-generated text examples often display the following characteristics:

• Repetitive phrases
• Grammatical errors
• Use of jargon and complex words
• Defaulting to a formal TOV
• Lack of originality/personal anecdotes
• Predictable flow
• Long sentence structures

Why You Might Want To Bypass GPTZero AI Detector?

You’ll need to bypass GPTZero AI detectors to avoid plagiarism and AI-related penalties and minimize the risk of dropping in the search engine rankings.

Avoid Plagiarism

Plagiarism is a major concern, especially in educational settings which can result in penalties or expulsion. Plagiarism means copying someone else’s work and claiming it as your own.

To avoid this, you’d need a paraphrasing tool like Smodin AI Rewriter to express the content’s meaning differently and uniquely. This could involve changing the sentence structure, using synonyms, and rephrasing sentences to make it more original.

Minimize Declining Search Engine Rankings

Many business owners saw their search engine rankings plummet when Google released its March 2024 Core Update . Google aims to reduce the amount of “spammy” AI-generated content from the SERPs (Search Engine Results Pages) by at least 40%.

This is because content creators have abused artificial intelligence (AI), filling the web with wishy-washy content that lacks authenticity and depth. A trend in recent years has been to focus content on Google rankings instead of being valuable to readers.

The update penalizes “ unhelpful, poor user experience ” content.

You’d want to bypass GPTZero AI detection software to minimize declining search engine rankings.

Luckily, there are ways to bypass GPTZero AI detection, which we’ll cover in the next section.

9 Ways To Bypass GPTZero AI Detection Systems

How to bypass GPTZero AI detectors involves using Smodin AI detection remover, writing how you speak, and switching passive for active voice – amongst others:

1. Using Smodin AI Rewriter

Smodin’s AI Rewriter transforms AI-generated content into an original piece of writing using advanced algorithms to paraphrase and rephrase AI-generated content. This results in a more human-like text that will bypass AI detection tools like GPTZero.

The Smodin AI detection remover has a built-in plagiarism checker that ensures your content passes AI detection tools and provides insightful answers for every subject, like chemistry and history (citing MLA or APA formats).

Transform AI-generated content within minutes when using Smodin’s AI Rewriter feature.

2. Use Unique Writing Styles

Opting for your own unique writing style is crucial for bypassing GPTZero’s AI detectors.

No AI detection tool can flag you for authenticity. Be genuine, transparent, and honest in how you write.

Many AI writing programs have an inherent writing style that AI detection tools are accustomed to. Long, drawn-out explanations and repetitiveness are often major red flags in this department.

If you’re using AI writers to outline your blog post, essay, or research paper, adapt the robotic writing style to match how you speak. This includes adding personal anecdotes, removing jargon, and simplifying complex explanations in a way that’s unique to you.

Reading a mix of different writing styles – such as formal, persuasive, or casual – can broaden your horizons and allow you to find a tone of voice (TOV) best suited to you.

Other tips include:

• Use bullet points and white space to break up large bodies of text.
• Make use of bold, italics, and capitalization.
• Be as authentic as possible.
• Write like you’re explaining something for the first time.
• Use synonyms.
• Alternate and vary sentence lengths.

3. Alternating Sentences

One way to identify AI-generated content is by recognizing sentences with a consistent length and flow. AI writers lack the variation and rhythm of how humans naturally speak.

Change up your sentence structure to include long and short sentences; you can even use one-worded sentences like “ Yikes !” to add a bit of pizazz!

Use white space to add depth to your writing.

4. Make It Personal

Personal anecdotes and stories are effective tools to connect with your reader, build trust, and come across as an expert in your field.

Writing sounds more human when personal stories are added. This is where an AI model misses the mark (almost) every time!

The idea behind great copy is to:

Tips To Make Your Writing Sound Natural Include:

• Use natural language that everyone can understand.
• Use emotion.
• Be authentic and relatable.
• Don’t overcomplicate your writing.
• Use storytelling.
• Be concise.

For instance, if you’re writing a travel blog about fun things to do in Bali, you don’t want to bore your audience with a rigid, monotone TOV, right?

You want to create excitement, describe the Balinese sights and sounds, and explain how they’ll feel when swooshing down water slides at Waterbom in Kuta!

You want to transport your audience to Bali, not make them want to avoid the place like the plague!

5. Use Simple Expressions

AI-generated content is filled with technical jargon, overly complicated explanations, and longer sentence structures. Granted, there are times you’ll need industry jargon depending on the intended audience and subject matter.

For instance, marketers talk about CTAs, PPCs, and CRMs. If this is aimed at a beginner, your audience won’t have the foggiest clue as to what you are saying. It would, however, make more sense to senior management marketing executives.

See the difference?

What this means is you must explain complex terms simply—end of story.

And the trick to great writing is simple writing .

No need for big, ambiguous words, phrases, or overly complicated sentence structure.

Simple writing improves your content’s readability – and bypasses AI tools like GPTZero!

Use Hemingway Editor and/or Grammarly to improve your readability score. It should be “easy” enough for a 9th Grader to understand.

6. Use Both Active and Passive Voice

Most AI content generators default to passive voice.

It’s basically a roundabout way of saying something simple.

• Passive voice: A delicious meal was prepared by the chef.
• Active voice: The chef prepared a delicious meal.

You’ll be flagged by AI-detection mechanisms if your content is jam-packed with passive voice.

It’s good practice to keep passive voice phrases to a minimum (although, yes, there are times when you’ll need to use them).

7. Use Copywriting Techniques

One way to bypass GPTZero AI detection is to use tried and tested copywriting formulas such as:

• SPEAR framework for introductions.
• Echoing back subheadings in the opening paragraph.
• Remember how Steve Jobs explained the iPod as “ a thousand songs in your pocket “? He didn’t mention anything about the mechanical components!
• Include reputable external links to back up claims (i.e., cited sources).
• Mention unique information not found online.
• Play on readers’ emotions using trigger/action words to create excitement, anger, fear, or awe.

8. Use Unique Prompts

Another ‘trick’ to try when it comes to AI generators is to get better at your instruction prompts. Get this right, and you can shave off a lot of time when it comes to humanizing your content.

• Familiarize yourself with Google’s E-E-A-T guidelines where Experience, Expertise, Authoriveness, and Trustworthiness are key!
• Tell AI writers to use a specific tone of voice or act like a certain industry expert to give it a better understanding and foundation of what it is you’re looking for.
• Where possible, do your own research on the topic and feed datasets, research papers, or mathematical equations into your AI writing program to ‘upskill’ its knowledge base. Then, ask the AI writer to ‘echo back’ what you have given it (you know, to make sure they grasped the concept!)

9. Type Everything Manually

Use AI writing tools responsibly to avoid being penalized for AI-generated content.

This means using features like ChatGPT or Google Gemini to help create content outlines, summarize large datasets, and break down technical terms.

Piece this together in your own words, add a few suggestions from above, like personal stories, and switch to active voice wherever possible.

Don’t mindlessly copy and paste directly from AI writing tools without going through it with a fine-toothed comb. While AI-generated content is produced in a fraction of the time compared to human writing, it’s not always reliable in terms of fact-checking.

Bypass GPTZero AI detection systems by writing your content yourself, whether it’s a history essay, medical report, or your homework solution.

Bypassing an AI detector like GPTZero involves using Smodin’s AI Rewriter, developing your own writing style, adding personal anecdotes, and keeping passive voice to a minimum.

By focusing on human connection with the help of personal stories, natural language, and simplifying complex jargon, you’ll connect with your audience, build trust, and solidify your online credibility. Understanding and implementing the tips laid out in this guide will ensure you bypass AI detection algorithms in no time!

Bypass GPTZero with Smodin’s AI Rewriter tool today!

Purdue Online Writing Lab Purdue OWL® College of Liberal Arts

Welcome to the Purdue Online Writing Lab

Welcome to the Purdue OWL

This page is brought to you by the OWL at Purdue University. When printing this page, you must include the entire legal notice.

Copyright ©1995-2018 by The Writing Lab & The OWL at Purdue and Purdue University. All rights reserved. This material may not be published, reproduced, broadcast, rewritten, or redistributed without permission. Use of this site constitutes acceptance of our terms and conditions of fair use.

The Online Writing Lab at Purdue University houses writing resources and instructional material, and we provide these as a free service of the Writing Lab at Purdue. Students, members of the community, and users worldwide will find information to assist with many writing projects. Teachers and trainers may use this material for in-class and out-of-class instruction.

The Purdue On-Campus Writing Lab and Purdue Online Writing Lab assist clients in their development as writers—no matter what their skill level—with on-campus consultations, online participation, and community engagement. The Purdue Writing Lab serves the Purdue, West Lafayette, campus and coordinates with local literacy initiatives. The Purdue OWL offers global support through online reference materials and services.

A Message From the Assistant Director of Content Development 

The Purdue OWL® is committed to supporting  students, instructors, and writers by offering a wide range of resources that are developed and revised with them in mind. To do this, the OWL team is always exploring possibilties for a better design, allowing accessibility and user experience to guide our process. As the OWL undergoes some changes, we welcome your feedback and suggestions by email at any time.

Please don't hesitate to contact us via our contact page  if you have any questions or comments.

All the best,

Social Media

Facebook twitter.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 03 June 2024

Applying large language models for automated essay scoring for non-native Japanese

• Wenchao Li 1 &
• Haitao Liu 2  

Humanities and Social Sciences Communications volume  11 , Article number:  723 ( 2024 ) Cite this article

320 Accesses

2 Altmetric

Metrics details

• Language and linguistics

Recent advancements in artificial intelligence (AI) have led to an increased use of large language models (LLMs) for language assessment tasks such as automated essay scoring (AES), automated listening tests, and automated oral proficiency assessments. The application of LLMs for AES in the context of non-native Japanese, however, remains limited. This study explores the potential of LLM-based AES by comparing the efficiency of different models, i.e. two conventional machine training technology-based methods (Jess and JWriter), two LLMs (GPT and BERT), and one Japanese local LLM (Open-Calm large model). To conduct the evaluation, a dataset consisting of 1400 story-writing scripts authored by learners with 12 different first languages was used. Statistical analysis revealed that GPT-4 outperforms Jess and JWriter, BERT, and the Japanese language-specific trained Open-Calm large model in terms of annotation accuracy and predicting learning levels. Furthermore, by comparing 18 different models that utilize various prompts, the study emphasized the significance of prompts in achieving accurate and reliable evaluations using LLMs.

Similar content being viewed by others

Scoring method of English composition integrating deep learning in higher vocational colleges

ChatGPT-3.5 as writing assistance in students’ essays

Detecting contract cheating through linguistic fingerprint

Conventional machine learning technology in aes.

AES has experienced significant growth with the advancement of machine learning technologies in recent decades. In the earlier stages of AES development, conventional machine learning-based approaches were commonly used. These approaches involved the following procedures: a) feeding the machine with a dataset. In this step, a dataset of essays is provided to the machine learning system. The dataset serves as the basis for training the model and establishing patterns and correlations between linguistic features and human ratings. b) the machine learning model is trained using linguistic features that best represent human ratings and can effectively discriminate learners’ writing proficiency. These features include lexical richness (Lu, 2012 ; Kyle and Crossley, 2015 ; Kyle et al. 2021 ), syntactic complexity (Lu, 2010 ; Liu, 2008 ), text cohesion (Crossley and McNamara, 2016 ), and among others. Conventional machine learning approaches in AES require human intervention, such as manual correction and annotation of essays. This human involvement was necessary to create a labeled dataset for training the model. Several AES systems have been developed using conventional machine learning technologies. These include the Intelligent Essay Assessor (Landauer et al. 2003 ), the e-rater engine by Educational Testing Service (Attali and Burstein, 2006 ; Burstein, 2003 ), MyAccess with the InterlliMetric scoring engine by Vantage Learning (Elliot, 2003 ), and the Bayesian Essay Test Scoring system (Rudner and Liang, 2002 ). These systems have played a significant role in automating the essay scoring process and providing quick and consistent feedback to learners. However, as touched upon earlier, conventional machine learning approaches rely on predetermined linguistic features and often require manual intervention, making them less flexible and potentially limiting their generalizability to different contexts.

In the context of the Japanese language, conventional machine learning-incorporated AES tools include Jess (Ishioka and Kameda, 2006 ) and JWriter (Lee and Hasebe, 2017 ). Jess assesses essays by deducting points from the perfect score, utilizing the Mainichi Daily News newspaper as a database. The evaluation criteria employed by Jess encompass various aspects, such as rhetorical elements (e.g., reading comprehension, vocabulary diversity, percentage of complex words, and percentage of passive sentences), organizational structures (e.g., forward and reverse connection structures), and content analysis (e.g., latent semantic indexing). JWriter employs linear regression analysis to assign weights to various measurement indices, such as average sentence length and total number of characters. These weights are then combined to derive the overall score. A pilot study involving the Jess model was conducted on 1320 essays at different proficiency levels, including primary, intermediate, and advanced. However, the results indicated that the Jess model failed to significantly distinguish between these essay levels. Out of the 16 measures used, four measures, namely median sentence length, median clause length, median number of phrases, and maximum number of phrases, did not show statistically significant differences between the levels. Additionally, two measures exhibited between-level differences but lacked linear progression: the number of attributives declined words and the Kanji/kana ratio. On the other hand, the remaining measures, including maximum sentence length, maximum clause length, number of attributive conjugated words, maximum number of consecutive infinitive forms, maximum number of conjunctive-particle clauses, k characteristic value, percentage of big words, and percentage of passive sentences, demonstrated statistically significant between-level differences and displayed linear progression.

Both Jess and JWriter exhibit notable limitations, including the manual selection of feature parameters and weights, which can introduce biases into the scoring process. The reliance on human annotators to label non-native language essays also introduces potential noise and variability in the scoring. Furthermore, an important concern is the possibility of system manipulation and cheating by learners who are aware of the regression equation utilized by the models (Hirao et al. 2020 ). These limitations emphasize the need for further advancements in AES systems to address these challenges.

Deep learning technology in AES

Deep learning has emerged as one of the approaches for improving the accuracy and effectiveness of AES. Deep learning-based AES methods utilize artificial neural networks that mimic the human brain’s functioning through layered algorithms and computational units. Unlike conventional machine learning, deep learning autonomously learns from the environment and past errors without human intervention. This enables deep learning models to establish nonlinear correlations, resulting in higher accuracy. Recent advancements in deep learning have led to the development of transformers, which are particularly effective in learning text representations. Noteworthy examples include bidirectional encoder representations from transformers (BERT) (Devlin et al. 2019 ) and the generative pretrained transformer (GPT) (OpenAI).

BERT is a linguistic representation model that utilizes a transformer architecture and is trained on two tasks: masked linguistic modeling and next-sentence prediction (Hirao et al. 2020 ; Vaswani et al. 2017 ). In the context of AES, BERT follows specific procedures, as illustrated in Fig. 1 : (a) the tokenized prompts and essays are taken as input; (b) special tokens, such as [CLS] and [SEP], are added to mark the beginning and separation of prompts and essays; (c) the transformer encoder processes the prompt and essay sequences, resulting in hidden layer sequences; (d) the hidden layers corresponding to the [CLS] tokens (T[CLS]) represent distributed representations of the prompts and essays; and (e) a multilayer perceptron uses these distributed representations as input to obtain the final score (Hirao et al. 2020 ).

AES system with BERT (Hirao et al. 2020 ).

The training of BERT using a substantial amount of sentence data through the Masked Language Model (MLM) allows it to capture contextual information within the hidden layers. Consequently, BERT is expected to be capable of identifying artificial essays as invalid and assigning them lower scores (Mizumoto and Eguchi, 2023 ). In the context of AES for nonnative Japanese learners, Hirao et al. ( 2020 ) combined the long short-term memory (LSTM) model proposed by Hochreiter and Schmidhuber ( 1997 ) with BERT to develop a tailored automated Essay Scoring System. The findings of their study revealed that the BERT model outperformed both the conventional machine learning approach utilizing character-type features such as “kanji” and “hiragana”, as well as the standalone LSTM model. Takeuchi et al. ( 2021 ) presented an approach to Japanese AES that eliminates the requirement for pre-scored essays by relying solely on reference texts or a model answer for the essay task. They investigated multiple similarity evaluation methods, including frequency of morphemes, idf values calculated on Wikipedia, LSI, LDA, word-embedding vectors, and document vectors produced by BERT. The experimental findings revealed that the method utilizing the frequency of morphemes with idf values exhibited the strongest correlation with human-annotated scores across different essay tasks. The utilization of BERT in AES encounters several limitations. Firstly, essays often exceed the model’s maximum length limit. Second, only score labels are available for training, which restricts access to additional information.

Mizumoto and Eguchi ( 2023 ) were pioneers in employing the GPT model for AES in non-native English writing. Their study focused on evaluating the accuracy and reliability of AES using the GPT-3 text-davinci-003 model, analyzing a dataset of 12,100 essays from the corpus of nonnative written English (TOEFL11). The findings indicated that AES utilizing the GPT-3 model exhibited a certain degree of accuracy and reliability. They suggest that GPT-3-based AES systems hold the potential to provide support for human ratings. However, applying GPT model to AES presents a unique natural language processing (NLP) task that involves considerations such as nonnative language proficiency, the influence of the learner’s first language on the output in the target language, and identifying linguistic features that best indicate writing quality in a specific language. These linguistic features may differ morphologically or syntactically from those present in the learners’ first language, as observed in (1)–(3).

我-送了-他-一本-书

Wǒ-sòngle-tā-yī běn-shū

1 sg .-give. past- him-one .cl- book

“I gave him a book.”

Agglutinative

彼-に-本-を-あげ-まし-た

Kare-ni-hon-o-age-mashi-ta

3 sg .- dat -hon- acc- give.honorification. past

Inflectional

give, give-s, gave, given, giving

Additionally, the morphological agglutination and subject-object-verb (SOV) order in Japanese, along with its idiomatic expressions, pose additional challenges for applying language models in AES tasks (4).

足-が 棒-に なり-ました

Ashi-ga bo-ni nar-mashita

leg- nom stick- dat become- past

“My leg became like a stick (I am extremely tired).”

The example sentence provided demonstrates the morpho-syntactic structure of Japanese and the presence of an idiomatic expression. In this sentence, the verb “なる” (naru), meaning “to become”, appears at the end of the sentence. The verb stem “なり” (nari) is attached with morphemes indicating honorification (“ます” - mashu) and tense (“た” - ta), showcasing agglutination. While the sentence can be literally translated as “my leg became like a stick”, it carries an idiomatic interpretation that implies “I am extremely tired”.

To overcome this issue, CyberAgent Inc. ( 2023 ) has developed the Open-Calm series of language models specifically designed for Japanese. Open-Calm consists of pre-trained models available in various sizes, such as Small, Medium, Large, and 7b. Figure 2 depicts the fundamental structure of the Open-Calm model. A key feature of this architecture is the incorporation of the Lora Adapter and GPT-NeoX frameworks, which can enhance its language processing capabilities.

GPT-NeoX Model Architecture (Okgetheng and Takeuchi 2024 ).

In a recent study conducted by Okgetheng and Takeuchi ( 2024 ), they assessed the efficacy of Open-Calm language models in grading Japanese essays. The research utilized a dataset of approximately 300 essays, which were annotated by native Japanese educators. The findings of the study demonstrate the considerable potential of Open-Calm language models in automated Japanese essay scoring. Specifically, among the Open-Calm family, the Open-Calm Large model (referred to as OCLL) exhibited the highest performance. However, it is important to note that, as of the current date, the Open-Calm Large model does not offer public access to its server. Consequently, users are required to independently deploy and operate the environment for OCLL. In order to utilize OCLL, users must have a PC equipped with an NVIDIA GeForce RTX 3060 (8 or 12 GB VRAM).

In summary, while the potential of LLMs in automated scoring of nonnative Japanese essays has been demonstrated in two studies—BERT-driven AES (Hirao et al. 2020 ) and OCLL-based AES (Okgetheng and Takeuchi, 2024 )—the number of research efforts in this area remains limited.

Another significant challenge in applying LLMs to AES lies in prompt engineering and ensuring its reliability and effectiveness (Brown et al. 2020 ; Rae et al. 2021 ; Zhang et al. 2021 ). Various prompting strategies have been proposed, such as the zero-shot chain of thought (CoT) approach (Kojima et al. 2022 ), which involves manually crafting diverse and effective examples. However, manual efforts can lead to mistakes. To address this, Zhang et al. ( 2021 ) introduced an automatic CoT prompting method called Auto-CoT, which demonstrates matching or superior performance compared to the CoT paradigm. Another prompt framework is trees of thoughts, enabling a model to self-evaluate its progress at intermediate stages of problem-solving through deliberate reasoning (Yao et al. 2023 ).

Beyond linguistic studies, there has been a noticeable increase in the number of foreign workers in Japan and Japanese learners worldwide (Ministry of Health, Labor, and Welfare of Japan, 2022 ; Japan Foundation, 2021 ). However, existing assessment methods, such as the Japanese Language Proficiency Test (JLPT), J-CAT, and TTBJ Footnote 1 , primarily focus on reading, listening, vocabulary, and grammar skills, neglecting the evaluation of writing proficiency. As the number of workers and language learners continues to grow, there is a rising demand for an efficient AES system that can reduce costs and time for raters and be utilized for employment, examinations, and self-study purposes.

This study aims to explore the potential of LLM-based AES by comparing the effectiveness of five models: two LLMs (GPT Footnote 2 and BERT), one Japanese local LLM (OCLL), and two conventional machine learning-based methods (linguistic feature-based scoring tools - Jess and JWriter).

The research questions addressed in this study are as follows:

To what extent do the LLM-driven AES and linguistic feature-based AES, when used as automated tools to support human rating, accurately reflect test takers’ actual performance?

What influence does the prompt have on the accuracy and performance of LLM-based AES methods?

The subsequent sections of the manuscript cover the methodology, including the assessment measures for nonnative Japanese writing proficiency, criteria for prompts, and the dataset. The evaluation section focuses on the analysis of annotations and rating scores generated by LLM-driven and linguistic feature-based AES methods.

Methodology

The dataset utilized in this study was obtained from the International Corpus of Japanese as a Second Language (I-JAS) Footnote 3 . This corpus consisted of 1000 participants who represented 12 different first languages. For the study, the participants were given a story-writing task on a personal computer. They were required to write two stories based on the 4-panel illustrations titled “Picnic” and “The key” (see Appendix A). Background information for the participants was provided by the corpus, including their Japanese language proficiency levels assessed through two online tests: J-CAT and SPOT. These tests evaluated their reading, listening, vocabulary, and grammar abilities. The learners’ proficiency levels were categorized into six levels aligned with the Common European Framework of Reference for Languages (CEFR) and the Reference Framework for Japanese Language Education (RFJLE): A1, A2, B1, B2, C1, and C2. According to Lee et al. ( 2015 ), there is a high level of agreement (r = 0.86) between the J-CAT and SPOT assessments, indicating that the proficiency certifications provided by J-CAT are consistent with those of SPOT. However, it is important to note that the scores of J-CAT and SPOT do not have a one-to-one correspondence. In this study, the J-CAT scores were used as a benchmark to differentiate learners of different proficiency levels. A total of 1400 essays were utilized, representing the beginner (aligned with A1), A2, B1, B2, C1, and C2 levels based on the J-CAT scores. Table 1 provides information about the learners’ proficiency levels and their corresponding J-CAT and SPOT scores.

A dataset comprising a total of 1400 essays from the story writing tasks was collected. Among these, 714 essays were utilized to evaluate the reliability of the LLM-based AES method, while the remaining 686 essays were designated as development data to assess the LLM-based AES’s capability to distinguish participants with varying proficiency levels. The GPT 4 API was used in this study. A detailed explanation of the prompt-assessment criteria is provided in Section Prompt . All essays were sent to the model for measurement and scoring.

Measures of writing proficiency for nonnative Japanese

Japanese exhibits a morphologically agglutinative structure where morphemes are attached to the word stem to convey grammatical functions such as tense, aspect, voice, and honorifics, e.g. (5).

食べ-させ-られ-まし-た-か

tabe-sase-rare-mashi-ta-ka

[eat (stem)-causative-passive voice-honorification-tense. past-question marker]

Japanese employs nine case particles to indicate grammatical functions: the nominative case particle が (ga), the accusative case particle を (o), the genitive case particle の (no), the dative case particle に (ni), the locative/instrumental case particle で (de), the ablative case particle から (kara), the directional case particle へ (e), and the comitative case particle と (to). The agglutinative nature of the language, combined with the case particle system, provides an efficient means of distinguishing between active and passive voice, either through morphemes or case particles, e.g. 食べる taberu “eat concusive . ” (active voice); 食べられる taberareru “eat concusive . ” (passive voice). In the active voice, “パン を 食べる” (pan o taberu) translates to “to eat bread”. On the other hand, in the passive voice, it becomes “パン が 食べられた” (pan ga taberareta), which means “(the) bread was eaten”. Additionally, it is important to note that different conjugations of the same lemma are considered as one type in order to ensure a comprehensive assessment of the language features. For example, e.g., 食べる taberu “eat concusive . ”; 食べている tabeteiru “eat progress .”; 食べた tabeta “eat past . ” as one type.

To incorporate these features, previous research (Suzuki, 1999 ; Watanabe et al. 1988 ; Ishioka, 2001 ; Ishioka and Kameda, 2006 ; Hirao et al. 2020 ) has identified complexity, fluency, and accuracy as crucial factors for evaluating writing quality. These criteria are assessed through various aspects, including lexical richness (lexical density, diversity, and sophistication), syntactic complexity, and cohesion (Kyle et al. 2021 ; Mizumoto and Eguchi, 2023 ; Ure, 1971 ; Halliday, 1985 ; Barkaoui and Hadidi, 2020 ; Zenker and Kyle, 2021 ; Kim et al. 2018 ; Lu, 2017 ; Ortega, 2015 ). Therefore, this study proposes five scoring categories: lexical richness, syntactic complexity, cohesion, content elaboration, and grammatical accuracy. A total of 16 measures were employed to capture these categories. The calculation process and specific details of these measures can be found in Table 2 .

T-unit, first introduced by Hunt ( 1966 ), is a measure used for evaluating speech and composition. It serves as an indicator of syntactic development and represents the shortest units into which a piece of discourse can be divided without leaving any sentence fragments. In the context of Japanese language assessment, Sakoda and Hosoi ( 2020 ) utilized T-unit as the basic unit to assess the accuracy and complexity of Japanese learners’ speaking and storytelling. The calculation of T-units in Japanese follows the following principles:

A single main clause constitutes 1 T-unit, regardless of the presence or absence of dependent clauses, e.g. (6).

ケンとマリはピクニックに行きました (main clause): 1 T-unit.

If a sentence contains a main clause along with subclauses, each subclause is considered part of the same T-unit, e.g. (7).

天気が良かった の で (subclause)、ケンとマリはピクニックに行きました (main clause): 1 T-unit.

In the case of coordinate clauses, where multiple clauses are connected, each coordinated clause is counted separately. Thus, a sentence with coordinate clauses may have 2 T-units or more, e.g. (8).

ケンは地図で場所を探して (coordinate clause)、マリはサンドイッチを作りました (coordinate clause): 2 T-units.

Lexical diversity refers to the range of words used within a text (Engber, 1995 ; Kyle et al. 2021 ) and is considered a useful measure of the breadth of vocabulary in L n production (Jarvis, 2013a , 2013b ).

The type/token ratio (TTR) is widely recognized as a straightforward measure for calculating lexical diversity and has been employed in numerous studies. These studies have demonstrated a strong correlation between TTR and other methods of measuring lexical diversity (e.g., Bentz et al. 2016 ; Čech and Miroslav, 2018 ; Çöltekin and Taraka, 2018 ). TTR is computed by considering both the number of unique words (types) and the total number of words (tokens) in a given text. Given that the length of learners’ writing texts can vary, this study employs the moving average type-token ratio (MATTR) to mitigate the influence of text length. MATTR is calculated using a 50-word moving window. Initially, a TTR is determined for words 1–50 in an essay, followed by words 2–51, 3–52, and so on until the end of the essay is reached (Díez-Ortega and Kyle, 2023 ). The final MATTR scores were obtained by averaging the TTR scores for all 50-word windows. The following formula was employed to derive MATTR:

\({\rm{MATTR}}({\rm{W}})=\frac{{\sum }_{{\rm{i}}=1}^{{\rm{N}}-{\rm{W}}+1}{{\rm{F}}}_{{\rm{i}}}}{{\rm{W}}({\rm{N}}-{\rm{W}}+1)}\)

Here, N refers to the number of tokens in the corpus. W is the randomly selected token size (W < N). \({F}_{i}\) is the number of types in each window. The \({\rm{MATTR}}({\rm{W}})\) is the mean of a series of type-token ratios (TTRs) based on the word form for all windows. It is expected that individuals with higher language proficiency will produce texts with greater lexical diversity, as indicated by higher MATTR scores.

Lexical density was captured by the ratio of the number of lexical words to the total number of words (Lu, 2012 ). Lexical sophistication refers to the utilization of advanced vocabulary, often evaluated through word frequency indices (Crossley et al. 2013 ; Haberman, 2008 ; Kyle and Crossley, 2015 ; Laufer and Nation, 1995 ; Lu, 2012 ; Read, 2000 ). In line of writing, lexical sophistication can be interpreted as vocabulary breadth, which entails the appropriate usage of vocabulary items across various lexicon-grammatical contexts and registers (Garner et al. 2019 ; Kim et al. 2018 ; Kyle et al. 2018 ). In Japanese specifically, words are considered lexically sophisticated if they are not included in the “Japanese Education Vocabulary List Ver 1.0”. Footnote 4 Consequently, lexical sophistication was calculated by determining the number of sophisticated word types relative to the total number of words per essay. Furthermore, it has been suggested that, in Japanese writing, sentences should ideally have a length of no more than 40 to 50 characters, as this promotes readability. Therefore, the median and maximum sentence length can be considered as useful indices for assessment (Ishioka and Kameda, 2006 ).

Syntactic complexity was assessed based on several measures, including the mean length of clauses, verb phrases per T-unit, clauses per T-unit, dependent clauses per T-unit, complex nominals per clause, adverbial clauses per clause, coordinate phrases per clause, and mean dependency distance (MDD). The MDD reflects the distance between the governor and dependent positions in a sentence. A larger dependency distance indicates a higher cognitive load and greater complexity in syntactic processing (Liu, 2008 ; Liu et al. 2017 ). The MDD has been established as an efficient metric for measuring syntactic complexity (Jiang, Quyang, and Liu, 2019 ; Li and Yan, 2021 ). To calculate the MDD, the position numbers of the governor and dependent are subtracted, assuming that words in a sentence are assigned in a linear order, such as W1 … Wi … Wn. In any dependency relationship between words Wa and Wb, Wa is the governor and Wb is the dependent. The MDD of the entire sentence was obtained by taking the absolute value of governor – dependent:

MDD = \(\frac{1}{n}{\sum }_{i=1}^{n}|{\rm{D}}{{\rm{D}}}_{i}|\)

In this formula, \(n\) represents the number of words in the sentence, and \({DD}i\) is the dependency distance of the \({i}^{{th}}\) dependency relationship of a sentence. Building on this, the annotation of sentence ‘Mary-ga-John-ni-keshigomu-o-watashita was [Mary- top -John- dat -eraser- acc -give- past] ’. The sentence’s MDD would be 2. Table 3 provides the CSV file as a prompt for GPT 4.

Cohesion (semantic similarity) and content elaboration aim to capture the ideas presented in test taker’s essays. Cohesion was assessed using three measures: Synonym overlap/paragraph (topic), Synonym overlap/paragraph (keywords), and word2vec cosine similarity. Content elaboration and development were measured as the number of metadiscourse markers (type)/number of words. To capture content closely, this study proposed a novel-distance based representation, by encoding the cosine distance between the essay (by learner) and essay task’s (topic and keyword) i -vectors. The learner’s essay is decoded into a word sequence, and aligned to the essay task’ topic and keyword for log-likelihood measurement. The cosine distance reveals the content elaboration score in the leaners’ essay. The mathematical equation of cosine similarity between target-reference vectors is shown in (11), assuming there are i essays and ( L i , …. L n ) and ( N i , …. N n ) are the vectors representing the learner and task’s topic and keyword respectively. The content elaboration distance between L i and N i was calculated as follows:

\(\cos \left(\theta \right)=\frac{{\rm{L}}\,\cdot\, {\rm{N}}}{\left|{\rm{L}}\right|{\rm{|N|}}}=\frac{\mathop{\sum }\nolimits_{i=1}^{n}{L}_{i}{N}_{i}}{\sqrt{\mathop{\sum }\nolimits_{i=1}^{n}{L}_{i}^{2}}\sqrt{\mathop{\sum }\nolimits_{i=1}^{n}{N}_{i}^{2}}}\)

A high similarity value indicates a low difference between the two recognition outcomes, which in turn suggests a high level of proficiency in content elaboration.

To evaluate the effectiveness of the proposed measures in distinguishing different proficiency levels among nonnative Japanese speakers’ writing, we conducted a multi-faceted Rasch measurement analysis (Linacre, 1994 ). This approach applies measurement models to thoroughly analyze various factors that can influence test outcomes, including test takers’ proficiency, item difficulty, and rater severity, among others. The underlying principles and functionality of multi-faceted Rasch measurement are illustrated in (12).

\(\log \left(\frac{{P}_{{nijk}}}{{P}_{{nij}(k-1)}}\right)={B}_{n}-{D}_{i}-{C}_{j}-{F}_{k}\)

(12) defines the logarithmic transformation of the probability ratio ( P nijk /P nij(k-1) )) as a function of multiple parameters. Here, n represents the test taker, i denotes a writing proficiency measure, j corresponds to the human rater, and k represents the proficiency score. The parameter B n signifies the proficiency level of test taker n (where n ranges from 1 to N). D j represents the difficulty parameter of test item i (where i ranges from 1 to L), while C j represents the severity of rater j (where j ranges from 1 to J). Additionally, F k represents the step difficulty for a test taker to move from score ‘k-1’ to k . P nijk refers to the probability of rater j assigning score k to test taker n for test item i . P nij(k-1) represents the likelihood of test taker n being assigned score ‘k-1’ by rater j for test item i . Each facet within the test is treated as an independent parameter and estimated within the same reference framework. To evaluate the consistency of scores obtained through both human and computer analysis, we utilized the Infit mean-square statistic. This statistic is a chi-square measure divided by the degrees of freedom and is weighted with information. It demonstrates higher sensitivity to unexpected patterns in responses to items near a person’s proficiency level (Linacre, 2002 ). Fit statistics are assessed based on predefined thresholds for acceptable fit. For the Infit MNSQ, which has a mean of 1.00, different thresholds have been suggested. Some propose stricter thresholds ranging from 0.7 to 1.3 (Bond et al. 2021 ), while others suggest more lenient thresholds ranging from 0.5 to 1.5 (Eckes, 2009 ). In this study, we adopted the criterion of 0.70–1.30 for the Infit MNSQ.

Moving forward, we can now proceed to assess the effectiveness of the 16 proposed measures based on five criteria for accurately distinguishing various levels of writing proficiency among non-native Japanese speakers. To conduct this evaluation, we utilized the development dataset from the I-JAS corpus, as described in Section Dataset . Table 4 provides a measurement report that presents the performance details of the 14 metrics under consideration. The measure separation was found to be 4.02, indicating a clear differentiation among the measures. The reliability index for the measure separation was 0.891, suggesting consistency in the measurement. Similarly, the person separation reliability index was 0.802, indicating the accuracy of the assessment in distinguishing between individuals. All 16 measures demonstrated Infit mean squares within a reasonable range, ranging from 0.76 to 1.28. The Synonym overlap/paragraph (topic) measure exhibited a relatively high outfit mean square of 1.46, although the Infit mean square falls within an acceptable range. The standard error for the measures ranged from 0.13 to 0.28, indicating the precision of the estimates.

Table 5 further illustrated the weights assigned to different linguistic measures for score prediction, with higher weights indicating stronger correlations between those measures and higher scores. Specifically, the following measures exhibited higher weights compared to others: moving average type token ratio per essay has a weight of 0.0391. Mean dependency distance had a weight of 0.0388. Mean length of clause, calculated by dividing the number of words by the number of clauses, had a weight of 0.0374. Complex nominals per T-unit, calculated by dividing the number of complex nominals by the number of T-units, had a weight of 0.0379. Coordinate phrases rate, calculated by dividing the number of coordinate phrases by the number of clauses, had a weight of 0.0325. Grammatical error rate, representing the number of errors per essay, had a weight of 0.0322.

Criteria (output indicator)

The criteria used to evaluate the writing ability in this study were based on CEFR, which follows a six-point scale ranging from A1 to C2. To assess the quality of Japanese writing, the scoring criteria from Table 6 were utilized. These criteria were derived from the IELTS writing standards and served as assessment guidelines and prompts for the written output.

A prompt is a question or detailed instruction that is provided to the model to obtain a proper response. After several pilot experiments, we decided to provide the measures (Section Measures of writing proficiency for nonnative Japanese ) as the input prompt and use the criteria (Section Criteria (output indicator) ) as the output indicator. Regarding the prompt language, considering that the LLM was tasked with rating Japanese essays, would prompt in Japanese works better Footnote 5 ? We conducted experiments comparing the performance of GPT-4 using both English and Japanese prompts. Additionally, we utilized the Japanese local model OCLL with Japanese prompts. Multiple trials were conducted using the same sample. Regardless of the prompt language used, we consistently obtained the same grading results with GPT-4, which assigned a grade of B1 to the writing sample. This suggested that GPT-4 is reliable and capable of producing consistent ratings regardless of the prompt language. On the other hand, when we used Japanese prompts with the Japanese local model “OCLL”, we encountered inconsistent grading results. Out of 10 attempts with OCLL, only 6 yielded consistent grading results (B1), while the remaining 4 showed different outcomes, including A1 and B2 grades. These findings indicated that the language of the prompt was not the determining factor for reliable AES. Instead, the size of the training data and the model parameters played crucial roles in achieving consistent and reliable AES results for the language model.

The following is the utilized prompt, which details all measures and requires the LLM to score the essays using holistic and trait scores.

Please evaluate Japanese essays written by Japanese learners and assign a score to each essay on a six-point scale, ranging from A1, A2, B1, B2, C1 to C2. Additionally, please provide trait scores and display the calculation process for each trait score. The scoring should be based on the following criteria:

Moving average type-token ratio.

Number of lexical words (token) divided by the total number of words per essay.

Number of sophisticated word types divided by the total number of words per essay.

Mean length of clause.

Verb phrases per T-unit.

Clauses per T-unit.

Dependent clauses per T-unit.

Complex nominals per clause.

Adverbial clauses per clause.

Coordinate phrases per clause.

Mean dependency distance.

Synonym overlap paragraph (topic and keywords).

Word2vec cosine similarity.

Connectives per essay.

Conjunctions per essay.

Number of metadiscourse markers (types) divided by the total number of words.

Number of errors per essay.

Japanese essay text

出かける前に二人が地図を見ている間に、サンドイッチを入れたバスケットに犬が入ってしまいました。それに気づかずに二人は楽しそうに出かけて行きました。やがて突然犬がバスケットから飛び出し、二人は驚きました。バスケット の 中を見ると、食べ物はすべて犬に食べられていて、二人は困ってしまいました。(ID_JJJ01_SW1)

The score of the example above was B1. Figure 3 provides an example of holistic and trait scores provided by GPT-4 (with a prompt indicating all measures) via Bing Footnote 6 .

Example of GPT-4 AES and feedback (with a prompt indicating all measures).

Statistical analysis

The aim of this study is to investigate the potential use of LLM for nonnative Japanese AES. It seeks to compare the scoring outcomes obtained from feature-based AES tools, which rely on conventional machine learning technology (i.e. Jess, JWriter), with those generated by AI-driven AES tools utilizing deep learning technology (BERT, GPT, OCLL). To assess the reliability of a computer-assisted annotation tool, the study initially established human-human agreement as the benchmark measure. Subsequently, the performance of the LLM-based method was evaluated by comparing it to human-human agreement.

To assess annotation agreement, the study employed standard measures such as precision, recall, and F-score (Brants 2000 ; Lu 2010 ), along with the quadratically weighted kappa (QWK) to evaluate the consistency and agreement in the annotation process. Assume A and B represent human annotators. When comparing the annotations of the two annotators, the following results are obtained. The evaluation of precision, recall, and F-score metrics was illustrated in equations (13) to (15).

\({\rm{Recall}}(A,B)=\frac{{\rm{Number}}\,{\rm{of}}\,{\rm{identical}}\,{\rm{nodes}}\,{\rm{in}}\,A\,{\rm{and}}\,B}{{\rm{Number}}\,{\rm{of}}\,{\rm{nodes}}\,{\rm{in}}\,A}\)

\({\rm{Precision}}(A,\,B)=\frac{{\rm{Number}}\,{\rm{of}}\,{\rm{identical}}\,{\rm{nodes}}\,{\rm{in}}\,A\,{\rm{and}}\,B}{{\rm{Number}}\,{\rm{of}}\,{\rm{nodes}}\,{\rm{in}}\,B}\)

The F-score is the harmonic mean of recall and precision:

\({\rm{F}}-{\rm{score}}=\frac{2* ({\rm{Precision}}* {\rm{Recall}})}{{\rm{Precision}}+{\rm{Recall}}}\)

The highest possible value of an F-score is 1.0, indicating perfect precision and recall, and the lowest possible value is 0, if either precision or recall are zero.

In accordance with Taghipour and Ng ( 2016 ), the calculation of QWK involves two steps:

Step 1: Construct a weight matrix W as follows:

\({W}_{{ij}}=\frac{{(i-j)}^{2}}{{(N-1)}^{2}}\)

i represents the annotation made by the tool, while j represents the annotation made by a human rater. N denotes the total number of possible annotations. Matrix O is subsequently computed, where O_( i, j ) represents the count of data annotated by the tool ( i ) and the human annotator ( j ). On the other hand, E refers to the expected count matrix, which undergoes normalization to ensure that the sum of elements in E matches the sum of elements in O.

Step 2: With matrices O and E, the QWK is obtained as follows:

K = 1- \(\frac{\sum i,j{W}_{i,j}\,{O}_{i,j}}{\sum i,j{W}_{i,j}\,{E}_{i,j}}\)

The value of the quadratic weighted kappa increases as the level of agreement improves. Further, to assess the accuracy of LLM scoring, the proportional reductive mean square error (PRMSE) was employed. The PRMSE approach takes into account the variability observed in human ratings to estimate the rater error, which is then subtracted from the variance of the human labels. This calculation provides an overall measure of agreement between the automated scores and true scores (Haberman et al. 2015 ; Loukina et al. 2020 ; Taghipour and Ng, 2016 ). The computation of PRMSE involves the following steps:

Step 1: Calculate the mean squared errors (MSEs) for the scoring outcomes of the computer-assisted tool (MSE tool) and the human scoring outcomes (MSE human).

Step 2: Determine the PRMSE by comparing the MSE of the computer-assisted tool (MSE tool) with the MSE from human raters (MSE human), using the following formula:

\({\rm{PRMSE}}=1-\frac{({\rm{MSE}}\,{\rm{tool}})\,}{({\rm{MSE}}\,{\rm{human}})\,}=1-\,\frac{{\sum }_{i}^{n}=1{({{\rm{y}}}_{i}-{\hat{{\rm{y}}}}_{{\rm{i}}})}^{2}}{{\sum }_{i}^{n}=1{({{\rm{y}}}_{i}-\hat{{\rm{y}}})}^{2}}\)

In the numerator, ŷi represents the scoring outcome predicted by a specific LLM-driven AES system for a given sample. The term y i − ŷ i represents the difference between this predicted outcome and the mean value of all LLM-driven AES systems’ scoring outcomes. It quantifies the deviation of the specific LLM-driven AES system’s prediction from the average prediction of all LLM-driven AES systems. In the denominator, y i − ŷ represents the difference between the scoring outcome provided by a specific human rater for a given sample and the mean value of all human raters’ scoring outcomes. It measures the discrepancy between the specific human rater’s score and the average score given by all human raters. The PRMSE is then calculated by subtracting the ratio of the MSE tool to the MSE human from 1. PRMSE falls within the range of 0 to 1, with larger values indicating reduced errors in LLM’s scoring compared to those of human raters. In other words, a higher PRMSE implies that LLM’s scoring demonstrates greater accuracy in predicting the true scores (Loukina et al. 2020 ). The interpretation of kappa values, ranging from 0 to 1, is based on the work of Landis and Koch ( 1977 ). Specifically, the following categories are assigned to different ranges of kappa values: −1 indicates complete inconsistency, 0 indicates random agreement, 0.0 ~ 0.20 indicates extremely low level of agreement (slight), 0.21 ~ 0.40 indicates moderate level of agreement (fair), 0.41 ~ 0.60 indicates medium level of agreement (moderate), 0.61 ~ 0.80 indicates high level of agreement (substantial), 0.81 ~ 1 indicates almost perfect level of agreement. All statistical analyses were executed using Python script.

Results and discussion

Annotation reliability of the llm.

This section focuses on assessing the reliability of the LLM’s annotation and scoring capabilities. To evaluate the reliability, several tests were conducted simultaneously, aiming to achieve the following objectives:

Assess the LLM’s ability to differentiate between test takers with varying levels of oral proficiency.

Determine the level of agreement between the annotations and scoring performed by the LLM and those done by human raters.

The evaluation of the results encompassed several metrics, including: precision, recall, F-Score, quadratically-weighted kappa, proportional reduction of mean squared error, Pearson correlation, and multi-faceted Rasch measurement.

Inter-annotator agreement (human–human annotator agreement)

We started with an agreement test of the two human annotators. Two trained annotators were recruited to determine the writing task data measures. A total of 714 scripts, as the test data, was utilized. Each analysis lasted 300–360 min. Inter-annotator agreement was evaluated using the standard measures of precision, recall, and F-score and QWK. Table 7 presents the inter-annotator agreement for the various indicators. As shown, the inter-annotator agreement was fairly high, with F-scores ranging from 1.0 for sentence and word number to 0.666 for grammatical errors.

The findings from the QWK analysis provided further confirmation of the inter-annotator agreement. The QWK values covered a range from 0.950 ( p  = 0.000) for sentence and word number to 0.695 for synonym overlap number (keyword) and grammatical errors ( p  = 0.001).

Agreement of annotation outcomes between human and LLM

To evaluate the consistency between human annotators and LLM annotators (BERT, GPT, OCLL) across the indices, the same test was conducted. The results of the inter-annotator agreement (F-score) between LLM and human annotation are provided in Appendix B-D. The F-scores ranged from 0.706 for Grammatical error # for OCLL-human to a perfect 1.000 for GPT-human, for sentences, clauses, T-units, and words. These findings were further supported by the QWK analysis, which showed agreement levels ranging from 0.807 ( p  = 0.001) for metadiscourse markers for OCLL-human to 0.962 for words ( p  = 0.000) for GPT-human. The findings demonstrated that the LLM annotation achieved a significant level of accuracy in identifying measurement units and counts.

Reliability of LLM-driven AES’s scoring and discriminating proficiency levels

This section examines the reliability of the LLM-driven AES scoring through a comparison of the scoring outcomes produced by human raters and the LLM ( Reliability of LLM-driven AES scoring ). It also assesses the effectiveness of the LLM-based AES system in differentiating participants with varying proficiency levels ( Reliability of LLM-driven AES discriminating proficiency levels ).

Reliability of LLM-driven AES scoring

Table 8 summarizes the QWK coefficient analysis between the scores computed by the human raters and the GPT-4 for the individual essays from I-JAS Footnote 7 . As shown, the QWK of all measures ranged from k  = 0.819 for lexical density (number of lexical words (tokens)/number of words per essay) to k  = 0.644 for word2vec cosine similarity. Table 9 further presents the Pearson correlations between the 16 writing proficiency measures scored by human raters and GPT 4 for the individual essays. The correlations ranged from 0.672 for syntactic complexity to 0.734 for grammatical accuracy. The correlations between the writing proficiency scores assigned by human raters and the BERT-based AES system were found to range from 0.661 for syntactic complexity to 0.713 for grammatical accuracy. The correlations between the writing proficiency scores given by human raters and the OCLL-based AES system ranged from 0.654 for cohesion to 0.721 for grammatical accuracy. These findings indicated an alignment between the assessments made by human raters and both the BERT-based and OCLL-based AES systems in terms of various aspects of writing proficiency.

Reliability of LLM-driven AES discriminating proficiency levels

After validating the reliability of the LLM’s annotation and scoring, the subsequent objective was to evaluate its ability to distinguish between various proficiency levels. For this analysis, a dataset of 686 individual essays was utilized. Table 10 presents a sample of the results, summarizing the means, standard deviations, and the outcomes of the one-way ANOVAs based on the measures assessed by the GPT-4 model. A post hoc multiple comparison test, specifically the Bonferroni test, was conducted to identify any potential differences between pairs of levels.

As the results reveal, seven measures presented linear upward or downward progress across the three proficiency levels. These were marked in bold in Table 10 and comprise one measure of lexical richness, i.e. MATTR (lexical diversity); four measures of syntactic complexity, i.e. MDD (mean dependency distance), MLC (mean length of clause), CNT (complex nominals per T-unit), CPC (coordinate phrases rate); one cohesion measure, i.e. word2vec cosine similarity and GER (grammatical error rate). Regarding the ability of the sixteen measures to distinguish adjacent proficiency levels, the Bonferroni tests indicated that statistically significant differences exist between the primary level and the intermediate level for MLC and GER. One measure of lexical richness, namely LD, along with three measures of syntactic complexity (VPT, CT, DCT, ACC), two measures of cohesion (SOPT, SOPK), and one measure of content elaboration (IMM), exhibited statistically significant differences between proficiency levels. However, these differences did not demonstrate a linear progression between adjacent proficiency levels. No significant difference was observed in lexical sophistication between proficiency levels.

To summarize, our study aimed to evaluate the reliability and differentiation capabilities of the LLM-driven AES method. For the first objective, we assessed the LLM’s ability to differentiate between test takers with varying levels of oral proficiency using precision, recall, F-Score, and quadratically-weighted kappa. Regarding the second objective, we compared the scoring outcomes generated by human raters and the LLM to determine the level of agreement. We employed quadratically-weighted kappa and Pearson correlations to compare the 16 writing proficiency measures for the individual essays. The results confirmed the feasibility of using the LLM for annotation and scoring in AES for nonnative Japanese. As a result, Research Question 1 has been addressed.

Comparison of BERT-, GPT-, OCLL-based AES, and linguistic-feature-based computation methods

This section aims to compare the effectiveness of five AES methods for nonnative Japanese writing, i.e. LLM-driven approaches utilizing BERT, GPT, and OCLL, linguistic feature-based approaches using Jess and JWriter. The comparison was conducted by comparing the ratings obtained from each approach with human ratings. All ratings were derived from the dataset introduced in Dataset . To facilitate the comparison, the agreement between the automated methods and human ratings was assessed using QWK and PRMSE. The performance of each approach was summarized in Table 11 .

The QWK coefficient values indicate that LLMs (GPT, BERT, OCLL) and human rating outcomes demonstrated higher agreement compared to feature-based AES methods (Jess and JWriter) in assessing writing proficiency criteria, including lexical richness, syntactic complexity, content, and grammatical accuracy. Among the LLMs, the GPT-4 driven AES and human rating outcomes showed the highest agreement in all criteria, except for syntactic complexity. The PRMSE values suggest that the GPT-based method outperformed linguistic feature-based methods and other LLM-based approaches. Moreover, an interesting finding emerged during the study: the agreement coefficient between GPT-4 and human scoring was even higher than the agreement between different human raters themselves. This discovery highlights the advantage of GPT-based AES over human rating. Ratings involve a series of processes, including reading the learners’ writing, evaluating the content and language, and assigning scores. Within this chain of processes, various biases can be introduced, stemming from factors such as rater biases, test design, and rating scales. These biases can impact the consistency and objectivity of human ratings. GPT-based AES may benefit from its ability to apply consistent and objective evaluation criteria. By prompting the GPT model with detailed writing scoring rubrics and linguistic features, potential biases in human ratings can be mitigated. The model follows a predefined set of guidelines and does not possess the same subjective biases that human raters may exhibit. This standardization in the evaluation process contributes to the higher agreement observed between GPT-4 and human scoring. Section Prompt strategy of the study delves further into the role of prompts in the application of LLMs to AES. It explores how the choice and implementation of prompts can impact the performance and reliability of LLM-based AES methods. Furthermore, it is important to acknowledge the strengths of the local model, i.e. the Japanese local model OCLL, which excels in processing certain idiomatic expressions. Nevertheless, our analysis indicated that GPT-4 surpasses local models in AES. This superior performance can be attributed to the larger parameter size of GPT-4, estimated to be between 500 billion and 1 trillion, which exceeds the sizes of both BERT and the local model OCLL.

Prompt strategy

In the context of prompt strategy, Mizumoto and Eguchi ( 2023 ) conducted a study where they applied the GPT-3 model to automatically score English essays in the TOEFL test. They found that the accuracy of the GPT model alone was moderate to fair. However, when they incorporated linguistic measures such as cohesion, syntactic complexity, and lexical features alongside the GPT model, the accuracy significantly improved. This highlights the importance of prompt engineering and providing the model with specific instructions to enhance its performance. In this study, a similar approach was taken to optimize the performance of LLMs. GPT-4, which outperformed BERT and OCLL, was selected as the candidate model. Model 1 was used as the baseline, representing GPT-4 without any additional prompting. Model 2, on the other hand, involved GPT-4 prompted with 16 measures that included scoring criteria, efficient linguistic features for writing assessment, and detailed measurement units and calculation formulas. The remaining models (Models 3 to 18) utilized GPT-4 prompted with individual measures. The performance of these 18 different models was assessed using the output indicators described in Section Criteria (output indicator) . By comparing the performances of these models, the study aimed to understand the impact of prompt engineering on the accuracy and effectiveness of GPT-4 in AES tasks.

Based on the PRMSE scores presented in Fig. 4 , it was observed that Model 1, representing GPT-4 without any additional prompting, achieved a fair level of performance. However, Model 2, which utilized GPT-4 prompted with all measures, outperformed all other models in terms of PRMSE score, achieving a score of 0.681. These results indicate that the inclusion of specific measures and prompts significantly enhanced the performance of GPT-4 in AES. Among the measures, syntactic complexity was found to play a particularly significant role in improving the accuracy of GPT-4 in assessing writing quality. Following that, lexical diversity emerged as another important factor contributing to the model’s effectiveness. The study suggests that a well-prompted GPT-4 can serve as a valuable tool to support human assessors in evaluating writing quality. By utilizing GPT-4 as an automated scoring tool, the evaluation biases associated with human raters can be minimized. This has the potential to empower teachers by allowing them to focus on designing writing tasks and guiding writing strategies, while leveraging the capabilities of GPT-4 for efficient and reliable scoring.

PRMSE scores of the 18 AES models.

This study aimed to investigate two main research questions: the feasibility of utilizing LLMs for AES and the impact of prompt engineering on the application of LLMs in AES.

To address the first objective, the study compared the effectiveness of five different models: GPT, BERT, the Japanese local LLM (OCLL), and two conventional machine learning-based AES tools (Jess and JWriter). The PRMSE values indicated that the GPT-4-based method outperformed other LLMs (BERT, OCLL) and linguistic feature-based computational methods (Jess and JWriter) across various writing proficiency criteria. Furthermore, the agreement coefficient between GPT-4 and human scoring surpassed the agreement among human raters themselves, highlighting the potential of using the GPT-4 tool to enhance AES by reducing biases and subjectivity, saving time, labor, and cost, and providing valuable feedback for self-study. Regarding the second goal, the role of prompt design was investigated by comparing 18 models, including a baseline model, a model prompted with all measures, and 16 models prompted with one measure at a time. GPT-4, which outperformed BERT and OCLL, was selected as the candidate model. The PRMSE scores of the models showed that GPT-4 prompted with all measures achieved the best performance, surpassing the baseline and other models.

In conclusion, this study has demonstrated the potential of LLMs in supporting human rating in assessments. By incorporating automation, we can save time and resources while reducing biases and subjectivity inherent in human rating processes. Automated language assessments offer the advantage of accessibility, providing equal opportunities and economic feasibility for individuals who lack access to traditional assessment centers or necessary resources. LLM-based language assessments provide valuable feedback and support to learners, aiding in the enhancement of their language proficiency and the achievement of their goals. This personalized feedback can cater to individual learner needs, facilitating a more tailored and effective language-learning experience.

There are three important areas that merit further exploration. First, prompt engineering requires attention to ensure optimal performance of LLM-based AES across different language types. This study revealed that GPT-4, when prompted with all measures, outperformed models prompted with fewer measures. Therefore, investigating and refining prompt strategies can enhance the effectiveness of LLMs in automated language assessments. Second, it is crucial to explore the application of LLMs in second-language assessment and learning for oral proficiency, as well as their potential in under-resourced languages. Recent advancements in self-supervised machine learning techniques have significantly improved automatic speech recognition (ASR) systems, opening up new possibilities for creating reliable ASR systems, particularly for under-resourced languages with limited data. However, challenges persist in the field of ASR. First, ASR assumes correct word pronunciation for automatic pronunciation evaluation, which proves challenging for learners in the early stages of language acquisition due to diverse accents influenced by their native languages. Accurately segmenting short words becomes problematic in such cases. Second, developing precise audio-text transcriptions for languages with non-native accented speech poses a formidable task. Last, assessing oral proficiency levels involves capturing various linguistic features, including fluency, pronunciation, accuracy, and complexity, which are not easily captured by current NLP technology.

Data availability

The dataset utilized was obtained from the International Corpus of Japanese as a Second Language (I-JAS). The data URLs: [ https://www2.ninjal.ac.jp/jll/lsaj/ihome2.html ].

J-CAT and TTBJ are two computerized adaptive tests used to assess Japanese language proficiency.

SPOT is a specific component of the TTBJ test.

J-CAT: https://www.j-cat2.org/html/ja/pages/interpret.html

SPOT: https://ttbj.cegloc.tsukuba.ac.jp/p1.html#SPOT .

The study utilized a prompt-based GPT-4 model, developed by OpenAI, which has an impressive architecture with 1.8 trillion parameters across 120 layers. GPT-4 was trained on a vast dataset of 13 trillion tokens, using two stages: initial training on internet text datasets to predict the next token, and subsequent fine-tuning through reinforcement learning from human feedback.

https://www2.ninjal.ac.jp/jll/lsaj/ihome2-en.html .

http://jhlee.sakura.ne.jp/JEV/ by Japanese Learning Dictionary Support Group 2015.

We express our sincere gratitude to the reviewer for bringing this matter to our attention.

On February 7, 2023, Microsoft began rolling out a major overhaul to Bing that included a new chatbot feature based on OpenAI’s GPT-4 (Bing.com).

Appendix E-F present the analysis results of the QWK coefficient between the scores computed by the human raters and the BERT, OCLL models.

Attali Y, Burstein J (2006) Automated essay scoring with e-rater® V.2. J. Technol., Learn. Assess., 4

Barkaoui K, Hadidi A (2020) Assessing Change in English Second Language Writing Performance (1st ed.). Routledge, New York. https://doi.org/10.4324/9781003092346

Bentz C, Tatyana R, Koplenig A, Tanja S (2016) A comparison between morphological complexity. measures: Typological data vs. language corpora. In Proceedings of the workshop on computational linguistics for linguistic complexity (CL4LC), 142–153. Osaka, Japan: The COLING 2016 Organizing Committee

Bond TG, Yan Z, Heene M (2021) Applying the Rasch model: Fundamental measurement in the human sciences (4th ed). Routledge

Brants T (2000) Inter-annotator agreement for a German newspaper corpus. Proceedings of the Second International Conference on Language Resources and Evaluation (LREC’00), Athens, Greece, 31 May-2 June, European Language Resources Association

Brown TB, Mann B, Ryder N, et al. (2020) Language models are few-shot learners. Advances in Neural Information Processing Systems, Online, 6–12 December, Curran Associates, Inc., Red Hook, NY

Burstein J (2003) The E-rater scoring engine: Automated essay scoring with natural language processing. In Shermis MD and Burstein JC (ed) Automated Essay Scoring: A Cross-Disciplinary Perspective. Lawrence Erlbaum Associates, Mahwah, NJ

Čech R, Miroslav K (2018) Morphological richness of text. In Masako F, Václav C (ed) Taming the corpus: From inflection and lexis to interpretation, 63–77. Cham, Switzerland: Springer Nature

Çöltekin Ç, Taraka, R (2018) Exploiting Universal Dependencies treebanks for measuring morphosyntactic complexity. In Aleksandrs B, Christian B (ed), Proceedings of first workshop on measuring language complexity, 1–7. Torun, Poland

Crossley SA, Cobb T, McNamara DS (2013) Comparing count-based and band-based indices of word frequency: Implications for active vocabulary research and pedagogical applications. System 41:965–981. https://doi.org/10.1016/j.system.2013.08.002

Article   Google Scholar  

Crossley SA, McNamara DS (2016) Say more and be more coherent: How text elaboration and cohesion can increase writing quality. J. Writ. Res. 7:351–370

CyberAgent Inc (2023) Open-Calm series of Japanese language models. Retrieved from: https://www.cyberagent.co.jp/news/detail/id=28817

Devlin J, Chang MW, Lee K, Toutanova K (2019) BERT: Pre-training of deep bidirectional transformers for language understanding. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics, Minneapolis, Minnesota, 2–7 June, pp. 4171–4186. Association for Computational Linguistics

Diez-Ortega M, Kyle K (2023) Measuring the development of lexical richness of L2 Spanish: a longitudinal learner corpus study. Studies in Second Language Acquisition 1-31

Eckes T (2009) On common ground? How raters perceive scoring criteria in oral proficiency testing. In Brown A, Hill K (ed) Language testing and evaluation 13: Tasks and criteria in performance assessment (pp. 43–73). Peter Lang Publishing

Elliot S (2003) IntelliMetric: from here to validity. In: Shermis MD, Burstein JC (ed) Automated Essay Scoring: A Cross-Disciplinary Perspective. Lawrence Erlbaum Associates, Mahwah, NJ

Google Scholar  

Engber CA (1995) The relationship of lexical proficiency to the quality of ESL compositions. J. Second Lang. Writ. 4:139–155

Garner J, Crossley SA, Kyle K (2019) N-gram measures and L2 writing proficiency. System 80:176–187. https://doi.org/10.1016/j.system.2018.12.001

Haberman SJ (2008) When can subscores have value? J. Educat. Behav. Stat., 33:204–229

Haberman SJ, Yao L, Sinharay S (2015) Prediction of true test scores from observed item scores and ancillary data. Brit. J. Math. Stat. Psychol. 68:363–385

Halliday MAK (1985) Spoken and Written Language. Deakin University Press, Melbourne, Australia

Hirao R, Arai M, Shimanaka H et al. (2020) Automated essay scoring system for nonnative Japanese learners. Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pp. 1250–1257. European Language Resources Association

Hunt KW (1966) Recent Measures in Syntactic Development. Elementary English, 43(7), 732–739. http://www.jstor.org/stable/41386067

Ishioka T (2001) About e-rater, a computer-based automatic scoring system for essays [Konpyūta ni yoru essei no jidō saiten shisutemu e − rater ni tsuite]. University Entrance Examination. Forum [Daigaku nyūshi fōramu] 24:71–76

Hochreiter S, Schmidhuber J (1997) Long short- term memory. Neural Comput. 9(8):1735–1780

Article   CAS   PubMed   Google Scholar  

Ishioka T, Kameda M (2006) Automated Japanese essay scoring system based on articles written by experts. Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Sydney, Australia, 17–18 July 2006, pp. 233-240. Association for Computational Linguistics, USA

Japan Foundation (2021) Retrieved from: https://www.jpf.gp.jp/j/project/japanese/survey/result/dl/survey2021/all.pdf

Jarvis S (2013a) Defining and measuring lexical diversity. In Jarvis S, Daller M (ed) Vocabulary knowledge: Human ratings and automated measures (Vol. 47, pp. 13–44). John Benjamins. https://doi.org/10.1075/sibil.47.03ch1

Jarvis S (2013b) Capturing the diversity in lexical diversity. Lang. Learn. 63:87–106. https://doi.org/10.1111/j.1467-9922.2012.00739.x

Jiang J, Quyang J, Liu H (2019) Interlanguage: A perspective of quantitative linguistic typology. Lang. Sci. 74:85–97

Kim M, Crossley SA, Kyle K (2018) Lexical sophistication as a multidimensional phenomenon: Relations to second language lexical proficiency, development, and writing quality. Mod. Lang. J. 102(1):120–141. https://doi.org/10.1111/modl.12447

Kojima T, Gu S, Reid M et al. (2022) Large language models are zero-shot reasoners. Advances in Neural Information Processing Systems, New Orleans, LA, 29 November-1 December, Curran Associates, Inc., Red Hook, NY

Kyle K, Crossley SA (2015) Automatically assessing lexical sophistication: Indices, tools, findings, and application. TESOL Q 49:757–786

Kyle K, Crossley SA, Berger CM (2018) The tool for the automatic analysis of lexical sophistication (TAALES): Version 2.0. Behav. Res. Methods 50:1030–1046. https://doi.org/10.3758/s13428-017-0924-4

Article   PubMed   Google Scholar  

Kyle K, Crossley SA, Jarvis S (2021) Assessing the validity of lexical diversity using direct judgements. Lang. Assess. Q. 18:154–170. https://doi.org/10.1080/15434303.2020.1844205

Landauer TK, Laham D, Foltz PW (2003) Automated essay scoring and annotation of essays with the Intelligent Essay Assessor. In Shermis MD, Burstein JC (ed), Automated Essay Scoring: A Cross-Disciplinary Perspective. Lawrence Erlbaum Associates, Mahwah, NJ

Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 159–174

Laufer B, Nation P (1995) Vocabulary size and use: Lexical richness in L2 written production. Appl. Linguist. 16:307–322. https://doi.org/10.1093/applin/16.3.307

Lee J, Hasebe Y (2017) jWriter Learner Text Evaluator, URL: https://jreadability.net/jwriter/

Lee J, Kobayashi N, Sakai T, Sakota K (2015) A Comparison of SPOT and J-CAT Based on Test Analysis [Tesuto bunseki ni motozuku ‘SPOT’ to ‘J-CAT’ no hikaku]. Research on the Acquisition of Second Language Japanese [Dainigengo to shite no nihongo no shūtoku kenkyū] (18) 53–69

Li W, Yan J (2021) Probability distribution of dependency distance based on a Treebank of. Japanese EFL Learners’ Interlanguage. J. Quant. Linguist. 28(2):172–186. https://doi.org/10.1080/09296174.2020.1754611

Article   MathSciNet   Google Scholar  

Linacre JM (2002) Optimizing rating scale category effectiveness. J. Appl. Meas. 3(1):85–106

PubMed   Google Scholar  

Linacre JM (1994) Constructing measurement with a Many-Facet Rasch Model. In Wilson M (ed) Objective measurement: Theory into practice, Volume 2 (pp. 129–144). Norwood, NJ: Ablex

Liu H (2008) Dependency distance as a metric of language comprehension difficulty. J. Cognitive Sci. 9:159–191

Liu H, Xu C, Liang J (2017) Dependency distance: A new perspective on syntactic patterns in natural languages. Phys. Life Rev. 21. https://doi.org/10.1016/j.plrev.2017.03.002

Loukina A, Madnani N, Cahill A, et al. (2020) Using PRMSE to evaluate automated scoring systems in the presence of label noise. Proceedings of the Fifteenth Workshop on Innovative Use of NLP for Building Educational Applications, Seattle, WA, USA → Online, 10 July, pp. 18–29. Association for Computational Linguistics

Lu X (2010) Automatic analysis of syntactic complexity in second language writing. Int. J. Corpus Linguist. 15:474–496

Lu X (2012) The relationship of lexical richness to the quality of ESL learners’ oral narratives. Mod. Lang. J. 96:190–208

Lu X (2017) Automated measurement of syntactic complexity in corpus-based L2 writing research and implications for writing assessment. Lang. Test. 34:493–511

Lu X, Hu R (2022) Sense-aware lexical sophistication indices and their relationship to second language writing quality. Behav. Res. Method. 54:1444–1460. https://doi.org/10.3758/s13428-021-01675-6

Ministry of Health, Labor, and Welfare of Japan (2022) Retrieved from: https://www.mhlw.go.jp/stf/newpage_30367.html

Mizumoto A, Eguchi M (2023) Exploring the potential of using an AI language model for automated essay scoring. Res. Methods Appl. Linguist. 3:100050

Okgetheng B, Takeuchi K (2024) Estimating Japanese Essay Grading Scores with Large Language Models. Proceedings of 30th Annual Conference of the Language Processing Society in Japan, March 2024

Ortega L (2015) Second language learning explained? SLA across 10 contemporary theories. In VanPatten B, Williams J (ed) Theories in Second Language Acquisition: An Introduction

Rae JW, Borgeaud S, Cai T, et al. (2021) Scaling Language Models: Methods, Analysis & Insights from Training Gopher. ArXiv, abs/2112.11446

Read J (2000) Assessing vocabulary. Cambridge University Press. https://doi.org/10.1017/CBO9780511732942

Rudner LM, Liang T (2002) Automated Essay Scoring Using Bayes’ Theorem. J. Technol., Learning and Assessment, 1 (2)

Sakoda K, Hosoi Y (2020) Accuracy and complexity of Japanese Language usage by SLA learners in different learning environments based on the analysis of I-JAS, a learners’ corpus of Japanese as L2. Math. Linguist. 32(7):403–418. https://doi.org/10.24701/mathling.32.7_403

Suzuki N (1999) Summary of survey results regarding comprehensive essay questions. Final report of “Joint Research on Comprehensive Examinations for the Aim of Evaluating Applicability to Each Specialized Field of Universities” for 1996-2000 [shōronbun sōgō mondai ni kansuru chōsa kekka no gaiyō. Heisei 8 - Heisei 12-nendo daigaku no kaku senmon bun’ya e no tekisei no hyōka o mokuteki to suru sōgō shiken no arikata ni kansuru kyōdō kenkyū’ saishū hōkoku-sho]. University Entrance Examination Section Center Research and Development Department [Daigaku nyūshi sentā kenkyū kaihatsubu], 21–32

Taghipour K, Ng HT (2016) A neural approach to automated essay scoring. Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, Austin, Texas, 1–5 November, pp. 1882–1891. Association for Computational Linguistics

Takeuchi K, Ohno M, Motojin K, Taguchi M, Inada Y, Iizuka M, Abo T, Ueda H (2021) Development of essay scoring methods based on reference texts with construction of research-available Japanese essay data. In IPSJ J 62(9):1586–1604

Ure J (1971) Lexical density: A computational technique and some findings. In Coultard M (ed) Talking about Text. English Language Research, University of Birmingham, Birmingham, England

Vaswani A, Shazeer N, Parmar N, et al. (2017) Attention is all you need. In Advances in Neural Information Processing Systems, Long Beach, CA, 4–7 December, pp. 5998–6008, Curran Associates, Inc., Red Hook, NY

Watanabe H, Taira Y, Inoue Y (1988) Analysis of essay evaluation data [Shōronbun hyōka dēta no kaiseki]. Bulletin of the Faculty of Education, University of Tokyo [Tōkyōdaigaku kyōiku gakubu kiyō], Vol. 28, 143–164

Yao S, Yu D, Zhao J, et al. (2023) Tree of thoughts: Deliberate problem solving with large language models. Advances in Neural Information Processing Systems, 36

Zenker F, Kyle K (2021) Investigating minimum text lengths for lexical diversity indices. Assess. Writ. 47:100505. https://doi.org/10.1016/j.asw.2020.100505

Zhang Y, Warstadt A, Li X, et al. (2021) When do you need billions of words of pretraining data? Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, Online, pp. 1112-1125. Association for Computational Linguistics. https://doi.org/10.18653/v1/2021.acl-long.90

Download references

This research was funded by National Foundation of Social Sciences (22BYY186) to Wenchao Li.

Author information

Authors and affiliations.

Department of Japanese Studies, Zhejiang University, Hangzhou, China

Department of Linguistics and Applied Linguistics, Zhejiang University, Hangzhou, China

You can also search for this author in PubMed   Google Scholar

Contributions

Wenchao Li is in charge of conceptualization, validation, formal analysis, investigation, data curation, visualization and writing the draft. Haitao Liu is in charge of supervision.

Corresponding author

Correspondence to Wenchao Li .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Ethical approval

Ethical approval was not required as the study did not involve human participants.

Informed consent

This article does not contain any studies with human participants performed by any of the authors.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplemental material file #1, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Li, W., Liu, H. Applying large language models for automated essay scoring for non-native Japanese. Humanit Soc Sci Commun 11 , 723 (2024). https://doi.org/10.1057/s41599-024-03209-9

Download citation

Received : 02 February 2024

Accepted : 16 May 2024

Published : 03 June 2024

DOI : https://doi.org/10.1057/s41599-024-03209-9

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies