How to Support Claims and Use Qualification in Research Writing


A research paper is going to be built around demonstrating a major claim through the use of reasons and evidence. Therefore, it’s important to know how to support claims and use qualification in research writing. 

Claims, Unsupported Claims, and Opinions

In casual language, we make unsupported claims all the time. Someone I know posted on Facebook today about their friend, “Your poems are amazing.” Another wrote in Instagram, “Running with friends is just good for the soul sometimes!” And another Facebook post read, “I love you Burger King, but your new tacos SUCK!”

What all three of these posts have in common is that they are unsubstantiated claims. That is, they are unsupported by reasons or evidence that can allow us to evaluate them. Another term for unsupported claims are “opinions.” Therefore, these are three statements of opinion. Without supporting evidence, they can’t rise to the level of verifiable and debatable facts. They wouldn’t work in research writing. 

However, these claims have been stated in a context where we don’t demand or expect them to be supported, so they can pass through our mental truth watchdogs (except perhaps the claim about Burger King’s tacos, since that’s a negative claim, and if it isn’t understood as just one person’s opinion, could sway our beliefs and actions).

Claims in research writing, however, have to be supported by reasons and evidence. Most especially major claims, but that goes for secondary claims as well. There’s no other option. If you are writing, and you find that a claim is made without evidence:

  • You must find reliable evidence to back it.
  • If there isn’t reliable evidence, it must be dropped or modified so it can be supported. Otherwise, it’s an opinion, and isn’t valid in a research setting. 

Example: A research paper that makes a claim that a certain form of writing instruction is superior than another form needs to contain reliable evidence that the form was investigated and found to be superior in some way.

I have students of great creativity who often invent wonderful ideas for research claims. While I champion creativity, in the end, it has to have backing, or it becomes an example of another great theory that fails due to lack of evidence. If so, it has to go!

Qualifying Claims, Rejecting or Qualifying Evidence

Some of the evidence you will use to back up a claim will come in the form of data, which are the facts and figures given as the results of experiments, surveys, reports, and other studies. For example, in 2017 suicides were the tenth largest cause of death in the United States, with 47,173 deaths, or 7.3 per 100,000 people (CDC National Center for Health Statistics). Some of the evidence you will use will come in the form of claims made by sources. For example, “Grammar is no longer being taught in junior high and high school, and students are moving into academic discourse communities (i.e., college) not knowing how to write a complete sentence” (Martinson 122).*

Either way, evidence in research papers needs to be used responsibly. And one of the keys to using evidence responsibly is not to overstate it. 

Let’s begin with the second example, Martinson claiming that grammar “is no longer being taught in junior high and high school”, and that students come to college “not knowing how to write a complete sentence.”

Before repeating that claim in your research paper, before trusting in its authority because it is published, you want to ask yourself whether it’s a valid claim or an opinion. Ask: Has Martinson provided evidence to back this claim? If she has, is the evidence valid?

Looking at the source in more detail, we see Martinson is telling an anecdote, a personal story. She recalls that when teaching a college lesson on sentence rhythm, and discussing adjectives, “no one” knew what she was talking about, which “confirmed my belief in the rumors I had heard: Grammar is no longer being taught in junior high and high school” (122). 

Is Martinson’s statement a valid claim or is it an opinion? (take a moment to think about it).

The answer is that it is an unsupported claim, otherwise known as an opinion. Common sense and experience tells us it’s not likely that “no one” in her class knew what an adjective was. This is hyperbole (exaggeration) for emotional impact, perhaps even humor. And, if you think about it, the fact that students in her class didn’t appear to know what an adjective is doesn’t necessarily mean that they didn’t, or that they were never taught what it is. In other words, if Martinson doesn’t provide any harder, more valid evidence to back up her claim, then we have to take it as just her opinion. We can’t use it in our research paper as a fact. We also don’t want to take it out of context. 

Martinson, it turns out, is setting up her paper with a personal story in order to explore four research questions, one of which is the following: “Does grammar have a place in the writing classroom of the new century? If so, what is it?”

Verdict: I wouldn’t use that claim at all. It is not evidence to support the major claim of any paper, and is not intended as such. Do not repeat unsubstantiated claims as supporting evidence for your research paper. Read your source carefully.

Back to the first example, “in 2017 suicides were the tenth largest cause of death in the United States, with 47,173 deaths, or 7.3 per 100,000 people (CDC National Center for Health Statistics).” The CDC is the primary source for this data, so assuming it is printed correctly, it appears to be trustworthy.

Sidebar on accuracy: Last month I visited the Gus Grissom Air Museum in Peru, Indiana, and was reading a display card about one of the helicopters exhibited, a Bell UH-1 or “Huey” helicopter that was used in Vietnam. The display card stated that the service ceiling, or highest altitude, this helicopter could fly was 14,300 meters. I immediately knew that figure was a mistake from personal experience. In April 2000 I hiked for 15 days in Nepal, up to Everest Base Camp, with my maximum trekking climb up a peak called Kala Patar, roughly 18,300 feet. It’s common on high altitude hikes and climbs for people to suddenly develop altitude sickness, which causes swelling of the brain and fluid on the lungs and can be fatal. When they do, there is often a helicopter rescue. However, we had been informed that we had hiked above the altitude that helicopters could fly, due to the air density not being enough to support the rotors. In other words, if we got ill, we would have to walk down or be carried. Now, there are approximately 3.3 feet in a meter. The card was saying that the service ceiling of the helicopter was approximately 47,200 feet, an impossible altitude for that type of aircraft. The museum had mixed up feet and meters! Can you imagine the number of people learning the wrong fact because of a simple mistake? What if you were a pilot and your aircraft manual mixed up gallons and liters, pounds and kilograms, or feet and meters? Can you imagine the possible results? Accuracy is important! Tedious, boring, but reliable attention to detail.

Back to the main lesson: Assuming the Web page had correctly presented the CDC facts, we have to ask, how accurate is the evidence? Clicking on the original source, “Deaths: Final Data for 2017” you will read the “Methods” section (my bolding):

Methods—Information reported on death certificates is presented in descriptive tabulations. The original records are filed in state registration offices. Statistical information is compiled in a national database through the Vital Statistics Cooperative Program of the National Center for Health Statistics. Causes of death are processed in accordance with the International Classification of Diseases, 10th Revision.

What can we learn from reading about the methods used to get these figures? Here’s what I learn about this massive data collection undertaking:

  • Some original records (death certificates) could be wrong.
  • Some records could be missing, not obtained, or not counted.
  • Copying the original records to the database could have introduced errors.
  • Classifying the causes of death could have introduced errors or subjective judgment where a cause of death was unclear.
  • In other words, what I see is that there are at least four to six ways to make mistakes for each record entering the final report.

Clearly, then, though the data seems highly specific: 47,173 deaths due to suicide in 2017, errors are likely. Therefore, it’s best to qualify your claims and your use of evidence.

Qualification in this sense means to express a necessary level of caution, uncertainty or margin of error about claims and evidence [definition is mine]. Another way to think of qualification is as a “limiter” – it adds limitations to the claims and evidence.

To add reasonable and necessary qualification to research writing, use words known as qualifiers, such as:

  • could, might, may, seems to, suggests
  • perhaps
  • sometimes, often
  • at least, about, around
  • or explain the qualification

So to qualify the CDC report, I might rewrite my original sentence as:

  • “The CDC reports there were 47,173 deaths due to suicide in 2017, though it’s possible the method of compiling data could introduce some errors.” (explaining)
  • “The CDC reported 47,173 deaths due to suicide in 2017.” (the word “reported” introduces the idea that the actual number could vary and is a responsible way to handle this kind of qualification.)
  • Even more accurate: “The CDC reported 47,173 deaths of US residents due to suicide in 2017, with a margin of error of about 1 percent.” (the word “reported” introduces the idea that the actual number could vary and is a responsible way to handle this kind of qualification. For the reasons why I added “US residents” and the “margin of error” see the next paragraph.)

Explanation: if you were to drill down further into the Methods, from a later discussion in the report, you’d find this statement: 

Data in this report are based on information from all resident death certificates filed in the 50 states and the District of Columbia. More than 99% of deaths occurring in this country are believed to be registered. (2)

From this information we learn that there was a category of non-resident deaths that weren’t counted, and that it’s estimated that 1% of deaths are not registered on death certificates. 1% sounds like a small number,  but it does represent about 472 deaths possibly unaccounted for. Further, the percentage of unregistered deaths is a “belief,” an estimate, a reasonable guess, but not a hard fact. So the number of unregistered deaths of residents of the US could be greater or smaller to some degree.

In conclusion, provided that the data were faithfully compiled, that no mistakes were made printing, that the percentage of unreported deaths is a reasonable one, and that we understand that we are looking at residents of the U.S. and not citizens who died outside the country, we can accept the figures as reasonable. But not exact. Therefore we use qualification.

CAUTION: You may read style advice that states the following: “An over-reliance on qualifiers is a sign of amateurish writing. To improve your writing, go through your text and find all the qualifiers. Take them out wherever you can” ( I would say this is good advice in general. But remember that here we are only talking about the necessary use of qualification in research writing. Clearly, without qualification, there could be misunderstanding.

Do not “de-qualify” sources!

Your final bit of advice, “do not ‘de-qualify’ sources.” Do not remove qualifiers to make sources read as if they are more definite than they are. This is a distortion and removes all necessary caution from research findings. So in other words, if a study concludes:

This study suggests that IOPs [intensive outpatient procedures] show great promise in delivering full doses of evidence-based treatment and producing rapid and clinically meaningful symptom reduction for different types of veterans including men and women as well as combat and MST trauma survivors. (10)**

Do not remove qualifiers (this study has two, which I’ve bolded, and do not write something like this:

The study found that intensive outpatient procedures produced “rapid and clinically meaningful symptom reduction for different types of veterans including men and women as well as combat and MST trauma survivors” (10). 

If you do, you will have removed important reasonable expressions of doubt by the original authors of the experiment, and will have distorted the findings.

I hope this lesson on the necessity of supporting claims with evidence, on looking carefully at the evidence and methods presented in sources, and on including necessary research qualifiers, helps you to write more professional, ethical, and accurate research papers. 

If you enjoyed this lesson, you might also like, “Tricks Research Writers Use to Find Primary Sources.”


*Martinsen, Amy. “The Tower of Babel and the Teaching of Grammar: Writing Instruction for a New Century.” The English Journal, vol. 90, no. 1, 2000, pp. 122–126. JSTOR,

**Zalta, Alyson K., et al. “Evaluating Patterns and Predictors of Symptom Change during a Three-Week Intensive Outpatient Treatment for Veterans with PTSD.” BMC Psychiatry, vol. 18, no. 1, July 2018, pp. 1–15. EBSCOhost, doi:10.1186/s12888-018-1816-6.

For several decades I have conducted writing workshops of all kinds, and for 14 years I have taught writing on the faculty of Indiana University Southeast. Now I have decided to give back for these opportunities by making my lessons available online. I hope you enjoy this lesson, and the other lessons here on my writing Web site, You may download and use any lesson here free of charge, provided you give credit as: © Copyright Michael Jackman. All Rights Reserved.

Although the lessons are free of charge, you can help support maintaining lessons on this site through the paypal link on the top right of this post. Thank you for your support! – Michael

I appreciate comments. Thank you!

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: