Study tips12 min read

Close-ended questions in research: types, examples and when psychologists use them

Explore dichotomous, multiple choice, Likert and rating scale items, compare them with open-ended questions and practice critiquing survey wording.

Medly TeamNovember 18, 2026
On this page

Close-ended questions in research give participants a fixed set of answers to choose from, so every response can be scored the same way and analyzed statistically. This guide explains what they are, the main types with psychology examples, and when psychologists choose them over open-ended questions. It also covers the wording biases examiners ask about, with practice questions at the end.

What are close-ended questions in research?

A close-ended question (also called a closed-ended or fixed-response question) is a survey item that "asks a question and provides a set of response options for respondents to choose from," in the words of an open research methods textbook used in many college psychology courses. An open-ended question lets people answer however they want.

The same textbook gives a simple rule for choosing. Closed items suit studies where researchers "have a good idea of the different responses that participants might make." Open items suit early, exploratory work where they don't.

Close-ended questions are a form of self-report, so they share its limits. As OpenStax Psychology puts it, people "may lie, misremember, or answer questions in a way that they think makes them look good."

Types of close-ended questions, with psychology examples

Most close-ended items fall into five types. Knowing which level of data each produces helps you answer the "which statistic would you use?" part of a methods question.

TypeFormatPsychology exampleData
DichotomousTwo options"Have you ever seen a therapist?"Nominal
Multiple choiceA list of categories"Main source of stress?"Nominal
Likert itemAgree to disagree, often 5 points"I feel nervous before exams." Strongly disagree to strongly agreeOrdinal
Rating scaleNumbered points"How well did you sleep?" 1 (very badly) to 7 (very well)Ordinal
RankingPut options in order"Rank these study methods from most to least useful."Ordinal

Source: Research Methods in Psychology, "Constructing Survey Questionnaires". Levels of data follow standard measurement terms. Last checked: November 17, 2026.

On a questionnaire, the first two examples look like this:

"Have you ever seen a therapist?"

A. Yes

B. No

"What is your main source of stress?"

A. Coursework

B. Money

C. Relationships

D. Other

A few details that come up in exams:

  • Likert scales come from Rensis Likert's 1932 method for measuring attitudes. Strictly, a Likert scale is several agreement statements added together into one score. A single statement is a Likert-type item.
  • Rating scales in practice run from 3 to 11 points. The textbook says five or seven options "generally allow about as much precision as respondents are capable of."
  • Many researchers treat rating scale totals as interval data so they can calculate means. If your course is strict about levels of measurement, call single items ordinal.
  • Multiple choice options should be mutually exclusive (no overlap) and exhaustive (every possible answer has a home, often through an "Other" option).

Close-ended vs open-ended questions: which gives better data?

Neither is better in general. Each one fits a different research question, and exam answers earn marks for linking the choice to reliability and validity.

FeatureClose-endedOpen-ended
DataQuantitativeQualitative
SpeedQuick to answer and scoreSlow; answers need coding
ReliabilityHigh: scored the same way each timeLower: coders may disagree
ValidityOptions may not fit the personCaptures their own words
Best forTesting a known ideaExploring a new topic

Source: Research Methods in Psychology and Pew Research Center. Last checked: November 17, 2026.

Reliability is the strong point. Every participant sees the same options, and a "4" means the same thing whoever scores it, so there's no need to check agreement between coders.

Validity is the weak point, because the options themselves shape the answers. After the 2008 US presidential election, Pew Research Center asked voters which issue mattered most. When the economy was on the list, 58% chose it. When the question was open-ended, only 35% named the economy unprompted. Only 8% of the closed group gave an answer outside the five options listed, compared with 43% of the open group.

That's why many researchers use both. Pew describes running open-ended pilot questions first, then turning the most common answers into the options for a closed version.

How question wording biases closed-ended answers

A closed item can be perfectly reliable and still measure the wrong thing. These are the biases examiners expect you to name.

Social desirability bias is the pull toward answers that look good. Pew notes that people tend to under-report things like drug use and over-report things like voting and charitable giving. A short yes or no item about a sensitive behavior makes the "good" answer very easy to tick. Anonymity helps, and so does wording that makes the less flattering answer feel normal.

Acquiescence bias is the habit of agreeing with statements whatever they say. Pew found it is stronger with agree or disagree items and when an interviewer is present. Two fixes: include reverse-worded items in a Likert scale, or offer a choice between two opposing statements.

Scale labels change answers too. In a study described in Norbert Schwarz's 1999 review, people rated their success in life. On a 0 to 10 scale, 34% picked a value from 0 to 5. On a -5 to +5 scale, only 13% picked the matching values from -5 to 0, because a negative number reads like failure.

Question order matters as well. In an October 2003 Pew poll, 45% favored legal agreements for gay and lesbian couples when the question came after one on same-sex marriage, compared with 37% when it didn't.

How to critique a survey question in a research methods exam

When an exam gives you a survey item and asks you to evaluate it, work through it in this order:

  1. Name the type of item and the level of data it produces.
  2. Check it is brief, relevant, unambiguous, specific and objective (the textbook's BRUSO checklist).
  3. Look for a double-barreled item: two ideas with one answer.
  4. Look for leading or loaded wording that hints at the "right" answer.
  5. Check the options are mutually exclusive and exhaustive.
  6. Ask which bias it invites: social desirability, acquiescence or an order effect.
  7. Say what this does to validity, then rewrite the item.

Example survey item

Question (4 marks)

Evaluate this survey item and suggest a better version: "Don't you agree that social media is harmful and makes teens anxious?"

A. Yes

B. No

The opening "Don't you agree" is leading and invites acquiescence. "Harmful" and "anxious" are two separate ideas, so the item is double-barreled. A yes or no format forces people who hold a middle view to pick a side. Together these lower validity: a high "yes" rate could reflect the wording, not real opinion. A better version uses two neutral Likert items: "Social media has a negative effect on my mood" and "Using social media makes me feel anxious," each rated from strongly disagree to strongly agree.

You can download a one-page survey question checklist to keep beside your notes. To get feedback on a written evaluation like this one, you can paste your answer into Medly and get it marked, with notes on what would earn the next mark.

If you're preparing for the MCAT, research design questions also appear in the Psychological, Social, and Biological Foundations of Behavior section, so this checklist fits into an MCAT study schedule too.

Worked examples

AP Psychology tests this topic through article analysis: you get a short study and a concept, and you explain whether the findings support or refute it. Medly's AP Psychology lesson on nonexperimental measurement teaches it in that order: classify the measure (a Likert rating coded 1 to 5 is quantitative), check whether the wording signals a preferred answer, then ask how peer review and replication might change the conclusion. The two Medly practice questions below are worked through step by step, following Medly's mark scheme.

Worked example 1: survey wording and social desirability bias

A youth sports club used a nonexperimental survey asking players to report how many of their last five practices they attended. One version asked, "As a committed teammate, how many of your last five practices did you attend?" A neutral version asked, "How many of your last five scheduled practices did you attend?" Coaches' attendance registers were checked after the survey.

Wording groupMean reported (of 5)Mean recorded (of 5)
"As a committed teammate..."4.73.4
Neutral wording3.53.4

Source: a Medly practice question (Source A, fictional data).

Question (2 marks)

Explain how at least one of the research findings supports or refutes the concept that survey wording that signals a socially approved answer can steer reports and produce social desirability bias and inaccurate self-reports.

Answer
  1. Compare the gap between reported and recorded attendance in each group. "Committed teammate" wording: 4.7 reported minus 3.4 recorded = a gap of 1.3 practices. Neutral wording: 3.5 minus 3.4 = a gap of 0.1 practices.
  2. The gap is much larger for the approving wording, so the finding supports the concept (1 mark for using Source A to compare the two gaps).
  3. Explain why: "committed teammate" signals that attending is the approved answer, which steers players to report more practices than they attended. That inflated report is social desirability bias, and because it is an inaccurate report about oneself, it is also self-report bias (1 mark for applying the mechanism to the finding).

Most lost marks come from quoting the numbers without saying what the difference shows. Name the gap, then name the bias.

Worked example 2: peer review and replication

An initial nonexperimental survey of 210 residents found that 73% agreed with the statement, "Neighborhood messaging apps improve my sense of connection." The authors concluded that using such apps increases neighborhood connection.

During peer review, reviewers required the authors to replace the causal claim "increases" with "is associated with" because the study only asked residents to report their app use and connection; it did not manipulate app use. An independent registered replication used the same questionnaire with a new group of residents and found that 46% agreed with the statement.

StudyAgreeing (%)
Original survey73%
Independent replication46%

Source: a Medly practice question (Sources B and C, fictional data).

Question (2 marks)

Explain how at least one of the research findings supports or refutes the concept that conclusions from nonexperimental research can be strengthened, qualified, or changed through peer review and replication.

Answer
  1. State your position: the evidence supports the concept. Peer review qualified the original causal conclusion, and the replication did not reproduce the same level of agreement (1 mark).
  2. Use a specific result: agreement fell from 73% in the original survey to 46% in the independent replication, using the same questionnaire. That is a drop of 27 percentage points, so the first conclusion should be qualified or changed, not treated as final (1 mark).
  3. You could instead use the peer-review change: "increases" became "is associated with" because app use was only reported, not manipulated, so the study can show an association but not cause and effect.

Notice that the agreement statement is a close-ended item. Its fixed response options make the two percentages directly comparable, which is what lets the replication challenge the original finding.

Practice questions

Now it's your turn. Try each question yourself first, then open the answer to check your working.

Question 1: (2 marks)

A researcher asks 200 college students: "Have you used a campus counseling service this year?"

A. Yes

B. No

Name the type of close-ended question and the level of data it produces.

Answer
  1. There are only two possible answers, so it is a dichotomous question (1 mark).
  2. Yes and No are categories with no order, so the data are nominal (1 mark).

Marks are often lost by calling yes or no answers ordinal. There is no "more" or "less" between the two.

Question 2: (3 marks)

A survey item reads: "How many hours did you sleep last night?"

A. 0 to 4

B. 4 to 6

C. 6 to 8

D. 8 to 10

Identify two problems with the response options and rewrite them.

Answer
  1. The options overlap. Someone who slept exactly 4 hours could pick the first or second option, so they are not mutually exclusive (1 mark).
  2. Someone who slept more than 10 hours has no option, so the list is not exhaustive (1 mark).
  3. Rewrite (1 mark):

A. Less than 4 hours
B. 4 to less than 6
C. 6 to less than 8
D. 8 to less than 10
E. 10 hours or more

Question 3: (3 marks)

"How often do you feel stressed and lonely at college?"

A. Never

B. Sometimes

C. Often

Explain one problem with this item and rewrite it as two separate items.

Answer
  1. It is double-barreled: stress and loneliness are different, so a student who is stressed but not lonely can't answer accurately (1 mark).
  2. "Sometimes" and "Often" are vague, so two students could mean different things by the same option (credit this as an alternative problem).
  3. Rewrite: "In the past two weeks, on how many days did you feel stressed?" and "In the past two weeks, on how many days did you feel lonely?", each with these options (2 marks):

A. 0 days
B. 1 to 4 days
C. 5 to 9 days
D. 10 to 14 days

Question 4: (4 marks)

A psychologist measures how often students cheat on assignments using a questionnaire of close-ended items. Students write their names at the top, and their instructor collects the sheets. Evaluate this method, referring to reliability and validity.

Answer
  1. Strength: fixed options give quantitative data that are scored the same way for every student, which supports reliability and makes groups easy to compare (1 mark).
  2. Weakness: cheating is a sensitive behavior, and names plus the instructor collecting the sheets make social desirability bias likely (1 mark).
  3. So students will probably under-report cheating, which lowers the validity of the results (1 mark).
  4. Improvement: make the survey anonymous and collected by someone outside the course, and word items so that admitting cheating feels less judged (1 mark).

Question 5: (2 marks)

A team studying why first-year students leave college has no earlier research to draw on. Should their first survey use close-ended or open-ended questions? Explain.

Answer
  1. Open-ended, because the team doesn't yet know the likely reasons, and a fixed list could miss the real ones (1 mark).
  2. They can then turn the most common answers into options for a close-ended survey of a larger sample (1 mark).

Medly can give you more research methods questions like these, explain any step you're unsure of and mark your written evaluations, so you can see where each mark comes from.

Frequently asked questions

Are close-ended questions qualitative or quantitative?

Quantitative. Fixed options can be counted, scored and compared, which is why they suit statistical analysis.

Is a Likert scale a close-ended question?

Yes. Each Likert item offers a fixed set of agreement options, usually five, so it is a close-ended format.

Are close-ended questions more reliable than open-ended questions?

Usually, yes. Every answer is scored the same way, so there's no disagreement between coders, although the options can still reduce validity.

What is an example of a close-ended question in psychology research?

"I feel nervous before exams," rated from strongly disagree to strongly agree, is a typical one. A yes or no question such as "Have you ever seen a therapist?" is another.

Should you write "close-ended" or "closed-ended"?

Both are used and mean the same thing. Textbooks more often say "closed-ended" or "closed questions."

Why do researchers combine open and closed questions?

Open questions show what people think in their own words, and closed questions measure how common those views are. Many surveys use an open pilot to write the closed options.

Try Medly for free