IB Psychology SLTopic 5 — Methods of ResearchPaper 1 & 2Core skill~10 min read
Questionnaires and Survey Design
A questionnaire is the cheapest way to reach a thousand people and the easiest way to get a thousand useless answers. The method is simple. The design is where all the skill and all the marks are.
📘 What you need to know
A questionnaire is a self-report method: participants tell you about their own thoughts, feelings and behaviour.
A survey is the wider study; the questionnaire is the tool it uses.
Closed questions give fixed options and produce quantitative data.
Open questions let people answer freely and produce qualitative data.
Standardised questions make a questionnaire easy to repeat, so reliability is usually good.
Social desirability bias is the biggest threat to validity: people under-report the bad and over-report the good.
Bad wording — leading, double-barrelled, or loaded — damages validity before anyone has answered.
Why ask instead of watch?
Some things simply cannot be observed. You cannot watch someone’s anxiety, their opinion on immigration, or how lonely they felt last winter. For those, the person is the only source. That is the whole case for self-report, and it is a strong one.
It also has an obvious weakness built into it. You are not measuring what people are like; you are measuring what people are willing to write down about what they are like. Keep that sentence in your head and most of the evaluation writes itself.
Open and closed questions
The tick boxes on the left can be counted in seconds. The three lines on the right may take a researcher an hour per participant to code into themes.
Feature
Closed questions
Open questions
Data produced
Quantitative: counts, scores, percentages
Qualitative: descriptions, themes
Analysis
Fast, statistical, easy to compare groups
Slow, interpretive, needs coding
Depth
Shallow: only the options you offered
Deep: reasons you never thought of
Risk
Forcing people into an answer that does not fit
Vague or off-topic replies you cannot use
A questionnaire made only of closed questions can tell you that people become more cautious as they get older. It can never tell you why. If a question asks about the limitations of questionnaires, that sentence is a ready-made answer.
Rating scales
A rating scale is a closed question with a range instead of categories. Participants pick a point between two ends, usually five options from “strongly disagree” to “strongly agree”. Each point is given a number, and the numbers across several items are added into a total score.
Why researchers like them: they turn something fuzzy, like how anxious someone feels, into a number that can be compared between people and across time. That is the whole idea behind psychometric testing.
They also invite two familiar problems. Some people avoid the extremes and cluster in the middle whatever they actually think. Others agree with everything put in front of them, which is acquiescence bias. Mixing in a few reverse-worded items forces people to actually read each one.
Writing questions that do not wreck your data
The second question already contains its own answer. Anything you collect from it tells you about the wording, not about the participant.
🧩 Five rules for writing questionnaire items
One idea per question. “Do you find school stressful and unfair?” is two questions wearing one coat, and a single answer cannot cover both.
No steering words. Drop “don’t you agree”, “surely”, “obviously”. They tell the participant what you want.
Plain language. If a 12-year-old could not read it, rewrite it. Jargon makes people guess.
Neutral first, sensitive later. Early questions shape later answers, which is question order bias. Warm up before you ask anything loaded.
Pilot it. Give it to a handful of people first and ask which questions confused them. Fix those before you collect real data.
Strengths and limitations
Strengths
Very large samples are possible, especially online, which helps generalisability.
Standardised questions mean it can be repeated exactly, so reliability can be tested.
Cheap and fast compared with interviewing everyone.
No researcher in the room, so investigator effects are reduced.
Limitations
Social desirability bias distorts anything sensitive.
Only the literate and the willing reply, so the sample skews.
You cannot check whether the question was understood.
Closed questions give the what and never the why.
Worked examples
WORKED EXAMPLE
Spot the problem and rewrite the question
A researcher studying study habits includes this item: “Most successful students revise every day. How often do you revise?” Identify one problem with this question and rewrite it. [3]
Step 1: Read the first sentence again
It tells the participant what a good answer looks like before they answer.
Step 2: Name the problemA leading question, which will push responses upward through social desirability bias.Step 3: Rewrite it neutrally
“On how many days in the last week did you revise?”
Remove the cue, ask for a countable factasking about last week rather than “usually” also removes guessing
WORKED EXAMPLE
Justify a choice of question type
A psychologist wants to know both how many students feel exam stress and what that stress feels like for them. Explain how the questionnaire should be designed. [4]
Split the aim in two
“How many” needs numbers. “What it feels like” needs words.
Match each to a question type
Closed rating-scale items for the level of stress = quantitative, comparable across studentsThen add depth
A small number of open questions afterwards = qualitative, explains the scoresA combined questionnaire, closed items firstput the closed items first so the open answers do not shape the ratings
💡 Exam tip
Always say what kind of data a question type produces. Closed to quantitative, open to qualitative — it is a free mark.
When evaluating validity, name the bias: social desirability, acquiescence, question order, or leading questions.
When evaluating reliability, mention test-retest and split-half. Questionnaires are one of the few methods where both apply neatly.
If asked to design one, include a pilot study. Very few students do, and it shows understanding of the method.
Anonymity is the standard fix for social desirability bias. Say it, and say why it works.
Large sample size is a strength of the method, not of the researcher. Explain it through generalisability.
⚠️ Common mix-up
Survey vs questionnaire. The survey is the study; the questionnaire is the set of questions it uses.
“Open question” does not mean “long question”. It means the answer is not fixed in advance.
Confusing closed questions with closed minds. Closed items can be excellent; they are just limited to what you thought to ask.
Saying questionnaires are unreliable. They are usually very reliable. It is validity that self-report puts at risk.
Treating a big sample as proof of generalisability. Ten thousand replies from one type of person still do not generalise.
Forgetting the non-responders. The people who ignore your survey may be exactly the ones you needed.
Up next: Interviews and What They Reveal — the same self-report problem, but with a person sitting opposite you.
Want this explained one-to-one?
Book a free session with an experienced IB Psychology tutor and get your trickiest topics made simple.
