IB Psychology SLTopic 5 — Research DesignPaper 1 & 2Core idea~11 min read
Bias in the Researcher and the Participant
There are two people in a study and both of them can bend the data. The researcher wants a result. The participant wants to look good, or to be helpful, or to work out what is going on. Neither is being dishonest. That is exactly why bias is so hard to remove.
📘 What you need to know
Researcher bias: the researcher’s presence, behaviour or expectations interfere with the process.
Investigator effects: the researcher’s characteristics and manner change how participants respond.
Confirmation bias: noticing evidence that fits the hypothesis and overlooking what does not.
Question order and leading question bias both shape answers through wording.
Participant bias: people change their behaviour because they know they are being studied.
Demand characteristics, social desirability, dominant respondent and acquiescence bias are the four to name.
All of these damage validity. Almost none of them damage reliability.
Two sides, one set of data
Use this as a checklist. Given any study, run down the eight names and ask which ones the design leaves the door open for.
Researcher bias
Investigator effects
The researcher is part of the situation. Their age, gender, ethnicity, tone of voice, body language and even what they are wearing can change how a participant behaves. A participant who is reminded of someone from their past may respond quite differently to the same instructions.
A harsh or over-excited tone adds emotion to what should be a neutral task.
Dramatic body language sits badly with scientific work and puts people in the wrong frame of mind.
Bright, patterned or slogan-bearing clothing is too personal and can cue a response.
Confirmation bias
Researchers have hypotheses, and hypotheses are easy to fall in love with. Confirmation bias is the tendency to notice and record what supports the prediction and to overlook or explain away what contradicts it. In an observation this is especially dangerous, because the researcher decides moment by moment what counts.
Bias built into the questions
Two named problems live here. Question order bias is where early questions colour later answers, so asking something loaded first pushes everything after it. The fix is to open neutrally and move to sensitive topics later. Leading question bias is where the wording tells the participant what the answer should be. Bias can also creep in through instructions and task framing, without a single question being asked, if the researcher unconsciously signals what they expect.
Sampling bias
If a researcher recruits people who are easy to approach or who look interested, the sample stops representing the population. Opportunity and self-selecting samples are the most exposed to this. The result is a study whose findings cannot be generalised, which is a hit to external validity.
Investigator effects and confirmation bias are worth separating in your head. One is about how the researcher affects the participant; the other is about how the researcher reads the data. Examiners notice when students use the right one.
Participant bias
Taking part in research is not a normal experience, and people respond to that. They look for cues, they guess the aim, and they adjust.
Some participants try to help the researcher and some deliberately do the opposite. Both are demand characteristics, and both wreck ecological validity.
Social desirability bias
People under-report the unflattering and over-report the flattering. Ask about exercise, alcohol, prejudice, revision or generosity and the answers drift towards what looks good. It is worst in self-report methods, and worse still in an interview than a questionnaire, because someone is sitting there listening.
Dominant respondent bias
In a focus group, one confident person can take over. They speak first, speak longest, and set the tone, and the others start agreeing rather than saying what they think. Unless the researcher manages this carefully and sensitively, the data records the loudest voice rather than the group.
Acquiescence bias
Some participants say yes to almost everything. It can come from wanting to please, from personality, or from not caring enough to think. It is why questionnaires full of “do you agree…” items are risky, and why open questions and reverse-worded items are used as a fix.
Bias
Where it comes from
What it does
Standard fix
Investigator effects
The researcher’s manner and appearance
Changes how participants respond
Standardised script, neutral dress and tone
Confirmation bias
The researcher’s expectations
Selective recording and interpretation
Blind procedures, second coder
Demand characteristics
The participant guessing the aim
Artificial behaviour
Single blind, covert methods, filler tasks
Social desirability
Wanting to look good
Distorted self-report
Anonymity, indirect questions
Acquiescence
Wanting to agree
Meaningless yes responses
Open questions, reverse-worded items
How researchers fight back
🧩 Design features that reduce bias
Standardised instructions, read from a script, so every participant meets the same words.
Single blind: the participant does not know which condition they are in.
Double blind: the researcher running the session does not know either, which removes both expectation effects at once.
Anonymity, so there is less reason to manage how you look.
Neutral wording and a sensible question order.
Independent coding: a second researcher analyses the data without knowing the hypothesis.
Worked examples
WORKED EXAMPLE
Name the bias and the fix
Students complete a questionnaire about how much they revise. Their teacher hands it out, collects it in, and their names are on the top. Identify the likely bias and suggest an improvement. [3]
Step 1: Who is watching?
A teacher who will see the answers with names attached.
Step 2: Name the biasSocial desirability bias — students will over-report revision to look good.
Step 3: State the effect
Reported revision hours will be inflated, so the data is not valid.
Make responses anonymous, collected by someone neutralnote that reliability is untouched — you would get the same inflated answers again
WORKED EXAMPLE
Explain why a double-blind design helps
A researcher predicts that a new revision technique improves recall. She runs the sessions herself and knows which group is which. Explain two problems and how a double-blind design would help. [4]
Problem 1: from the researcher
Knowing the condition, she may unconsciously be warmer or clearer with the technique group. Investigator effects.Problem 2: from the participant
Cues from her manner let participants guess the aim and try harder. Demand characteristics.The fix
Someone who does not know the hypothesis or the conditions runs the sessions.
Both expectation effects are removed at oncesay what it protects: internal validity, because the IV becomes the only difference again
💡 Exam tip
Name the specific bias. “Bias” on its own earns almost nothing.
Say who the bias comes from — researcher or participant. Questions often ask for one side only.
Every bias point should end with its effect on validity.
Pair each bias with a realistic fix. Naming the problem is half the mark; solving it is the other half.
Single blind and double blind are easy terms to use and are rarely used well. Explain who does not know what.
Remember bias usually leaves reliability alone. Saying this shows you understand both terms.
⚠️ Common mix-up
Demand characteristics vs social desirability. One is guessing the aim; the other is managing your image.
Investigator effects vs confirmation bias. Affecting the participant is not the same as misreading the data.
Assuming bias means deliberate. Almost all of it is unconscious, which is what makes it hard to remove.
Saying bias makes a study unreliable. It makes it invalid.
Confusing sampling bias with participant bias. Sampling bias is about who was recruited, not how they behaved.
Thinking anonymity fixes everything. It helps with social desirability and does nothing for demand characteristics.
Up next: Reflexivity: The Researcher in the Research — what qualitative researchers do about bias when removing it is not an option.
Want this explained one-to-one?
Book a free session with an experienced IB Psychology tutor and get your trickiest topics made simple.
