Skip to main content

Research Biases — Previous Year Questions

📊 Topic Overview​

MetricValue
Papers Appeared5 of 28 (17.9%)
Priority🟡 MEDIUM
Section DistributionSec A: 2 | Sec B: 1 | Sec C: 2
Most RepeatedBiases in psychological research and ways to minimise them (2 times)
ProbabilityMedium–Low

🎯 Must-Know Concepts​

  1. Types of researcher bias and participant bias
  2. Demand characteristics and how to control them
  3. Experimenter expectancy effects (Rosenthal effect)
  4. Sampling bias and selection bias
  5. Ways to minimise bias and maximise objectivity

🔥 Key Question: Biases in Psychological Research and How to Minimise Them (Section A/B)​

Answer (~400 words):​

Definition: Research bias refers to systematic errors introduced into a study that distort results and lead to invalid conclusions. Unlike random error (which increases variability), bias produces consistent deviations in one direction. Bias can originate from the researcher, the participants, the instruments, or the research context.

Types of Bias:

A. Researcher/Experimenter Biases:

1. Experimenter Expectancy Effect (Rosenthal Effect) The researcher unconsciously influences participants' behaviour or their own data recording in the direction of their hypothesis. Rosenthal and Fode (1963) demonstrated this with rats — experimenters told their rats were "maze-bright" obtained better results, even though rats were randomly assigned. Control: Double-blind procedure — both researcher and participant are blind to condition assignment.

2. Observer Bias When recording observational data, researchers tend to see what they expect to see. Ambiguous behaviours are coded consistent with the hypothesis. Control: Blind coding, operational definitions, multiple observers, inter-rater reliability checks.

3. Confirmation Bias Tendency to seek, interpret, and recall information that confirms one's pre-existing beliefs. Control: Pre-registration of hypotheses; peer review; disconfirmatory research strategies.

B. Participant Biases:

1. Demand Characteristics (Orne, 1962) Participants try to figure out the purpose of the study and modify their behaviour accordingly — either to confirm the hypothesis (social desirability) or to sabotage it (negativistic subject role). Control: Deception (followed by debriefing), cover stories, within-subjects designs with distractor tasks, naturalistic observation.

2. Social Desirability Bias Participants respond in ways they believe are socially acceptable or normative rather than truthfully. Control: Anonymity, use of indirect questions, behavioural measures rather than self-report.

3. Volunteer Bias (Rosenthal & Rosnow) Volunteers systematically differ from non-volunteers (more educated, motivated, less authoritarian). Findings may not generalise to the general population. Control: Incentivise participation; compare respondents vs non-respondents on known variables.

C. Measurement Biases:

  • Social desirability scales (e.g., Marlowe-Crowne) can detect and statistically control response bias.
  • Ambiguous item wording introduces systematic error.

Ways to Maximise Objectivity:

  1. Standardise procedures: Identical instructions and conditions for all participants.
  2. Blind/double-blind procedures: Prevent expectancy effects.
  3. Random assignment: Controls for selection bias.
  4. Use validated instruments: Reduces measurement bias.
  5. Inter-rater reliability: Multiple coders for qualitative data.
  6. Reflexivity (in qualitative research): Researcher explicitly acknowledges and documents their own position and potential biases.

📝 Short Note Version (Section C — ~120 words):​

Research biases are systematic errors that distort findings. Key types: (1) Experimenter expectancy effect — researcher unconsciously influences results toward hypothesis; controlled by double-blind procedure; (2) Demand characteristics — participants alter behaviour based on perceived study purpose; controlled by deception and naturalistic settings; (3) Social desirability bias — participants respond normatively; controlled by anonymity; (4) Volunteer bias — volunteers differ systematically from non-volunteers; (5) Confirmation bias — researcher seeks confirming evidence. Objectivity is maximised through standardisation, randomisation, blinding, validated instruments, and (in qualitative research) reflexivity. Kerlinger (1986) noted that science aims to be "self-correcting" — peer review and replication are the ultimate checks on bias.

🧠 Memory Aids​

Researcher biases (EO-C): Expectancy, Observer bias, Confirmation bias

Participant biases (D-S-V): Demand characteristics, Social desirability, Volunteer bias

Controls (SD-B-R-V-I-R): Standardise → Double-blind → Randomise → Validate instruments → Inter-rater reliability → Reflexivity

📚 Related Content​

✅ Self-Assessment​

  1. What is the Rosenthal effect and how is it controlled?
  2. How do demand characteristics affect research validity?
  3. What is the difference between observer bias and confirmation bias?
  4. Why are volunteers problematic for external validity?
  5. How does a double-blind procedure control for experimenter expectancy?

💡 Exam Tips​

  • Short note (Section C) is the most common format — know the concise version.
  • For Section A, organise by source of bias: researcher, participant, measurement.
  • Rosenthal's experimenter expectancy effect with rats is a memorable, high-scoring example — always include it.