Research methods — AQA GCSE Psychology
Research methods explained
Research methods covers 29 specification points, including Formulation of testable hypotheses Null… and Types of variable Independent variable,….
Read the full explanation
Open a point below for its explanation and revision guidance.
Formulation of testable hypotheses Null hypothesis and alternative hypothesis.
A testable hypothesis is a precise, operationalised prediction about the relationship between variables that a study could actually check. To make it testable, state the independent variable and dependent variable clearly and say how each is measured, for example: 'Participants who drink 200 mg of caffeine will score higher on a 20-item reaction-time test than participants who drink decaffeinated coffee.' The alternative hypothesis predicts a difference or relationship will occur. The null hypothesis predicts no difference or relationship, and any apparent pattern is due to chance. Both are stated before data collection so the analysis can compare them.
Types of variable Independent variable, dependent variable, extraneous variables.
Variables are the factors that can change in a study. The independent variable is the factor the researcher manipulates or selects to create conditions. The dependent variable is the factor measured to see the effect, and it should be operationalised, for example reaction time in milliseconds. Extraneous variables are other factors that could affect the dependent variable if not controlled, such as noise, time of day or participant age. Researchers control extraneous variables through standardised procedures, random allocation or counterbalancing so that any difference in the dependent variable can be attributed to the independent variable rather than to these other influences.
Sampling methods Target populations, samples and sampling methods and how to select samples using these methods:
In psychological research the target population is the whole group the researcher wants to draw conclusions about, for example all Year 11 students in a school. Because testing everyone is usually impossible, the researcher selects a sample, a smaller group drawn from that population, and the sampling method is the procedure used to choose them. The method matters because a biased sample can make findings unrepresentative, so conclusions may not generalise back to the target population. Learners should be able to identify the target population in a scenario, distinguish it from the sample actually tested, and recognise how a named sampling method is carried out step by step. This underpins later evaluation of generalisability.
random
Random sampling is a method of selecting participants in which every member of the target population has an equal chance of being chosen. A common procedure is to list all members of the target population, assign each a number, and then use a random technique such as a random number generator or drawing numbers from a hat to pick the required number of participants. Because selection is not influenced by the researcher, random sampling reduces selection bias and tends to produce a more representative sample than an opportunistic approach. However, it requires a complete list of the target population, which may be difficult or impossible to obtain, and participants who are selected may decline to take part, which can reintroduce bias.
opportunity
In AQA GCSE Psychology research methods, 'opportunity' names one way of selecting participants for a sample. An opportunity sample is formed by asking whoever is available and willing at the time and place the researcher is working — for example, a psychologist testing passers-by in a shopping centre on a Saturday morning. The researcher does not use a random or systematic procedure; the people present simply get the chance to take part. This makes it quick, cheap and convenient, but the sample may be unrepresentative because it reflects only those in that location at that moment, so findings may not generalise to the wider target population. In an MCQ you may be asked to identify which sampling method a described scenario illustrates, or to pick the main weakness of an opportunity sample.
systematic
In AQA GCSE Psychology research methods, 'systematic' refers to systematic sampling, a technique for selecting participants. The researcher takes the target population (often arranged in a list, the sampling frame) and selects every nth person — for example, every 5th name on a register. The starting point is usually chosen at random, then a fixed interval is applied. This gives an unbiased, objective rule for selection and is quick to carry out, but it can be biased if the list has a repeating pattern that matches the interval. In an MCQ you may be asked to identify systematic sampling from a description, to spot the fixed interval, or to choose its main strength or weakness.
stratified.
Stratified sampling is a probability method in which the researcher first identifies subgroups (strata) within the target population, such as year groups in a school or age bands in a town, then draws a random sample from within each stratum in proportion to its size. For example, if a population is 60% female and 40% male, a stratified sample of 50 would aim for about 30 females and 20 males, each chosen randomly within their group. This makes the sample more representative of the population's structure than simple random sampling alone, though it requires knowing the population's composition in advance and dividing it accurately. In an MCQ, identify stratified sampling by its defining feature: deliberate proportional representation of identified subgroups, each sampled randomly.
Strengths and weaknesses of each sampling method.
Each sampling method trades representativeness against practicality. Random sampling gives every member an equal chance and reduces bias, but needs a full list and can still produce unrepresentative samples by chance. Systematic sampling (every nth person) is quick and unbiased if the list is random, yet can coincide with a pattern in the list. Stratified sampling represents subgroups proportionally but requires knowing the population's structure. Opportunity sampling is fast and cheap but biased by who is available. Volunteer sampling is convenient and may suit sensitive topics but attracts a motivated, unrepresentative group. A strong answer names the method, states a strength and a weakness, and explains each with reference to bias, representativeness, cost or practicality.
Understanding principles of sampling as applied to scientific data.
Sampling means selecting the participants or data a study will actually use from a larger group. The target population is everyone the researcher wants to draw conclusions about; the sample is the smaller set actually studied. A representative sample reflects the target population's key features, such as age or gender mix, so findings can be generalised. Random sampling gives every member an equal chance of selection, reducing researcher bias. Opportunity sampling uses whoever is available, which is quick but may be unrepresentative. Volunteer sampling relies on people offering themselves, which can produce a biased, motivated group. Larger samples tend to be more representative, though a biased large sample is still biased.
Designing research Quantitative and qualitative methods:
When designing research, psychologists choose between quantitative and qualitative methods. Quantitative methods produce numerical data that can be counted, averaged and analysed statistically, for example scores on a rating scale or the number of seconds a task takes. Qualitative methods produce non-numerical, descriptive data such as words, themes or detailed accounts, for example from open-ended interview answers. Quantitative data is easy to compare and analyse but may oversimplify experience. Qualitative data gives rich detail and insight but is harder to summarise and compare. Many designs use both, giving mixed methods. The choice depends on the research aim: measuring a difference suits quantitative data, while exploring feelings suits qualitative data.
the experimental method (experimental designs, independent groups, repeated measures, matched pairs, including strengths and weaknesses of each experimental design)
The experimental method tests a causal hypothesis by manipulating an independent variable (IV) and measuring a dependent variable (DV) while controlling extraneous variables. Experimental design describes how participants are allocated to conditions. In independent groups, different participants do each condition, so there is no order effect but participant variables differ between groups and more people are needed. In repeated measures, the same participants do every condition, so participant variables are controlled and fewer people are needed, but order effects (practice, fatigue, boredom) arise and must be counterbalanced. In matched pairs, separate participants are paired on a relevant variable such as age or ability, reducing participant variables while avoiding order effects, but matching is time-consuming and never perfect. Each design trades control against practicality.
laboratory experiments
A laboratory experiment is one conducted in a controlled setting, not necessarily a literal laboratory, where the researcher manipulates the IV and controls extraneous variables. High control allows cause and effect to be established and permits replication, because procedures and standardised instructions can be repeated. However, the artificial setting and tasks may reduce ecological validity, and participants who know they are being studied may show demand characteristics. Laboratory experiments therefore trade control and replicability against realism. A learner should be able to recognise a laboratory experiment from a description and evaluate its strengths and weaknesses, distinguishing it from field and natural experiments.
field and natural experiments
A field experiment manipulates an independent variable in a real-world setting, such as testing whether background music changes how much shoppers spend in an actual shop. The researcher still controls the IV and measures the DV, but participants usually do not know they are being studied. A natural experiment does not involve the researcher manipulating the IV: the change happens naturally, for example comparing stress levels in pupils before and after a school moves to a new building. The researcher records the effect of that naturally occurring change on a measured variable. Learn to tell them apart by asking who caused the change in the IV: the researcher (field) or events outside the researcher's control (natural). Both raise ethical and control issues you should be able to identify.
interviews
An interview is a research method in which a researcher asks a participant questions to gather data, usually face to face but also by phone or online. In a structured interview the same fixed questions are asked in the same order, giving consistent, comparable answers that are easy to analyse but limit follow-up. In an unstructured interview the questions are open and flexible, so the interviewer can probe answers and gain rich detail, but responses are harder to compare and analyse. A semi-structured interview combines set questions with the freedom to follow up. Interviews can collect detailed qualitative data and allow clarification, but they can be affected by interviewer effects, social desirability and the time needed to transcribe and analyse responses.
questionnaires
A questionnaire is a self-report method in which participants answer a written set of questions, usually about their thoughts, feelings or behaviour. Questions may be closed, giving fixed options that produce quantitative data, or open, inviting free responses that produce qualitative data. Questionnaires can be administered on paper, by post or online, which lets researchers reach large samples quickly and cheaply, and standardised wording supports replication. However, respondents may misunderstand items, give socially desirable answers, or fail to return the questionnaire, lowering response rates and validity. Learners should be able to identify a questionnaire, distinguish open from closed items, and weigh its strengths and weaknesses against other methods such as interviews.
case studies
A case study is an in-depth investigation of a single individual, small group or event, often using several methods together such as interviews, observations and tests. Because the focus is detailed, case studies can produce rich qualitative and quantitative data and are useful where a phenomenon is rare or unusual, for example studying one patient with a distinctive memory difficulty. The depth allows researchers to explore context and development over time. However, findings from one case cannot easily be generalised to the wider population, and the researcher's close involvement may introduce bias. Learners should recognise a case study, note its use of multiple methods, and weigh its strengths against its limited generalisability.
observation studies (including categories of behaviour and interobserver reliability).
Observation studies record behaviour as it happens rather than asking people to report it. A researcher first decides what to observe, then builds a behaviour checklist or coding scheme: a set of clearly defined categories of behaviour, each with an operational definition so two observers would classify the same act the same way. For example, a study of playground aggression might code 'hits', 'shouts at' and 'takes property' separately, with examples for each. Categories must be mutually exclusive and cover the behaviours of interest, or data are lost or double-counted. Interobserver reliability is the extent to which two or more observers watching the same event produce consistent records. It is checked by having observers code the same session independently, then comparing their records; high agreement shows the categories are clear and the coding is reliable.
Strengths and weaknesses of each research method and types of research for which they are suitable.
Each research method has strengths and weaknesses, and some methods suit particular research questions better than others. Experiments allow control of variables and causal conclusions but can be artificial and may raise ethical issues. Observations record natural behaviour with good ecological validity but risk observer bias and cannot easily show cause. Questionnaires and interviews gather self-report data efficiently but depend on honesty and memory, and questions can be leading. Correlational studies show relationships between variables but cannot establish causation. Case studies give rich detail about one individual or group but generalise poorly. When evaluating, match the method to the aim: if the researcher wants to establish cause and effect, an experiment is suitable; if they want detailed natural behaviour, observation may fit.
Correlation An understanding of association between two variables and the use of scatter diagrams to show possible correlational relationships.
A correlation describes an association between two co-variables: as one changes, the other tends to change too. It does not show that one variable causes the other. A positive correlation means both rise together (for example, taller students tending to weigh more); a negative correlation means one rises as the other falls (for example, more hours of revision tending to link with fewer errors). A scatter diagram plots each participant as one point, with one co-variable on the horizontal axis and the other on the vertical axis. The pattern of points shows the direction and strength of the possible relationship: points tightly clustered along a rising line suggest a strong positive correlation, a falling line a strong negative one, and a shapeless cloud little or no correlation. Correlation is not causation, so a scatter diagram can only suggest a possible relationship.
The strengths and weaknesses of correlations.
Correlations examine whether two co-variables are associated. A strength is that they allow researchers to study relationships between variables that would be unethical or impractical to manipulate directly, such as the link between hours of sleep and exam performance, and they can guide further research by suggesting hypotheses. They also use existing data efficiently. A weakness is that correlation does not establish causation: a third variable may explain the association, so a correlation between ice-cream sales and drowning does not mean one causes the other. Correlations also cannot show which variable influences which, and they may miss non-linear relationships. A scatter diagram displays the pattern of points, but interpretation of strength and direction remains a judgement.
Computation of formulae is not required.
This guidance sets the boundary for the mathematical demand of Research methods. You need to understand what each descriptive and inferential measure is for, when it is appropriate and how to interpret a reported value, but you are not required to calculate formulae yourself. So you should be able to recognise that a mean summarises a set of scores, that a range describes spread, and that a correlation coefficient expresses the strength and direction of a relationship, and to say what a given figure suggests about the data. You should also be able to choose a suitable measure for a described study and justify the choice in words. What you will not be asked to do is work through the arithmetic of a formula, such as computing a standard deviation by hand. Treat any numerical work as interpretation and selection rather than computation, and always explain your reasoning in prose.
Research procedures The use of standardised procedures, instructions to participants, randomisation, allocation to conditions, counterbalancing and extraneous variables (including explaining the effect of extraneous variables and how to control for them).
Research procedures are the practical steps that make a study consistent and fair. Standardised procedures mean every participant experiences the same task, materials and setting, so results are comparable. Instructions to participants should be identical for everyone, reducing variation caused by how the study is explained. Randomisation means assigning participants or conditions in an unbiased way, for example drawing names, so researcher choice does not skew groups. Allocation to conditions is how participants are placed into experimental groups, and random allocation helps make groups equivalent. Counterbalancing controls order effects in repeated measures designs by varying the order of conditions, for example half the participants do task A then B while the other half do B then A.
Planning and conducting research How research should be planned, taking into consideration the reliability and/or validity of:
This guidance row frames how you plan and conduct research while protecting reliability and validity. Read it as a planning checklist, not a fact to memorise. Reliability means consistency: if the procedure were repeated, would it give the same results? Validity means measuring what you intend to measure, in a way that reflects real behaviour. When planning, decide your aim, hypothesis, variables, sampling, materials, standardised instructions, controls and ethics, then ask how each choice affects consistency and accuracy. For example, using the same scripted instructions and timing for every participant supports reliability, while using a task that genuinely taps the behaviour you claim to study supports validity. Consider piloting, inter-rater checks and controlling extraneous variables.
sampling methods
Sampling methods describe how researchers choose participants from a target population. A random sample gives every member an equal chance of selection, reducing bias. Opportunity sampling uses whoever is available at the time, which is quick but may be unrepresentative. Volunteer sampling relies on people offering themselves, which can attract a particular type of person. Systematic sampling selects every nth person from a list. Stratified sampling builds a sample matching the proportions of subgroups in the population. Each method trades practicality against representativeness and bias, which affects how far findings generalise.
experimental designs
Experimental designs describe how participants are allocated to conditions in an experiment. In independent groups, different participants take part in each condition, so no one does both; this avoids order effects but needs more participants and risks participant-variable differences between groups. In repeated measures, the same participants do every condition, controlling participant variables and needing fewer people, but order effects such as practice or fatigue can occur and must be countered by counterbalancing. In matched pairs, participants are paired on a relevant variable and one of each pair goes to each condition, reducing participant variables without repeated testing, though matching is time-consuming and never perfect. Recognise each design from a description and weigh its strengths and weaknesses.
quantitative and qualitative methods.
Quantitative methods collect numerical data that can be counted, summarised and analysed statistically, for example scores on a rating scale, number of errors recalled, or reaction times in milliseconds. Qualitative methods collect non-numerical data that capture meaning, experience and description, for example verbatim responses in an unstructured interview, diary entries or open-ended questionnaire answers. The two approaches answer different questions: quantitative data allow comparison and patterns across participants, while qualitative data give richer insight into why people behave as they do. Learners should classify a described method or data set as quantitative or qualitative, and evaluate the trade-off between easy statistical analysis and depth of understanding. A method can yield both types, so judge the data actually produced rather than the label of the technique alone.
Ethical considerations Students should demonstrate knowledge and understanding of:
This heading introduces the ethical considerations strand of Research methods. You must show knowledge and understanding of the ethical principles that govern psychological research, and be able to recognise them in described studies. In a multiple-choice question you will typically be given a short scenario or a definition and asked to select the correct ethical principle, the correct safeguard, or the correct reason a procedure is unacceptable. Learn each principle by what it protects and by the practical safeguard that satisfies it: informed consent through a briefing and consent form, the right to withdraw stated clearly, confidentiality through anonymised data, protection from harm through risk assessment and debriefing, and deception justified only when necessary and followed by full debrief. Practise matching scenarios to principles so you can reject plausible but wrong options.
ethical issues in psychological research as outlined in the British Psychological Society guidelines
This statement requires knowledge and understanding of the ethical issues in psychological research as set out in the British Psychological Society (BPS) guidelines. The BPS code sets out principles researchers must follow, including respect for participants, informed consent, confidentiality, protection from harm, and the responsible handling of deception. In a multiple-choice question you may be asked which guideline applies to a scenario, or which action a researcher should take. Learn each issue with its safeguard: obtain informed consent and record the right to withdraw; keep data confidential and anonymised; assess and minimise risk of harm and debrief participants; use deception only when justified and always debrief. Match the scenario to the specific BPS principle rather than a loosely related one.
ways of dealing with each of these issues.
This criterion asks you to know practical ways of dealing with ethical issues in psychological research. The main issues are informed consent, deception, protection from harm, confidentiality, privacy and the right to withdraw. For each issue, learn a concrete method. Consent can be handled with a consent form or presumptive consent. Deception can be dealt with by a full debrief afterwards. Protection from harm is managed by risk assessment, a suitable withdrawal procedure and follow-up care. Confidentiality is protected by anonymising data, for example using numbers instead of names. Privacy is respected by not observing people where they would not expect it. The right to withdraw is dealt with by telling participants clearly, at the start and throughout, that they may leave. In an MCQ, match the issue to the correct method rather than a vague statement that it was 'handled'.
Your focus
- State what makes a hypothesis testable by operationalising the independent and dependent variables.
- Distinguish an alternative hypothesis from a null hypothesis.
- Explain why both hypotheses are formulated before data collection.
Show all 87 objectives
- Identify the independent and dependent variables in a described study.
- Explain what an extraneous variable is and give an example.
- Describe how extraneous variables can be controlled to protect the validity of a study.
- Define target population and sample accurately.
- Identify the target population and sample in a described study.
- Explain how the sampling method used affects representativeness of findings.
- Define random sampling accurately.
- Describe a step-by-step procedure for carrying out random sampling.
- Evaluate random sampling by stating at least one strength and one limitation.
- Define an opportunity sample in terms of available and willing participants.
- Identify opportunity sampling from a described research scenario.
- Explain one strength and one weakness of using an opportunity sample.
- Define systematic sampling in terms of selecting every nth participant.
- Identify systematic sampling from a described research scenario.
- Explain one strength and one weakness of using a systematic sample.
- Define stratified sampling and identify its key features.
- Distinguish stratified sampling from other sampling methods in a multiple-choice item.
- Explain why proportional subgroup selection can improve representativeness.
- Identify a strength and a weakness of each named sampling method.
- Explain how each strength or weakness affects the quality of research.
- Compare sampling methods using research concepts such as bias and generalisability.
- Define target population and sample accurately.
- Explain what makes a sample representative and why this matters for generalisation.
- Evaluate at least two sampling methods in terms of bias and practicality.
- Distinguish quantitative from qualitative data with accurate examples.
- Explain one strength and one limitation of each data type.
- Justify a method choice by reference to the research aim.
- Identify the three experimental designs and how participants are allocated in each.
- Distinguish the strengths and weaknesses of independent groups, repeated measures and matched pairs.
- Select the correct design feature or evaluation when given a described study.
- Define a laboratory experiment and recognise it from a description.
- Explain strengths such as control, causal conclusions and replication.
- Explain weaknesses such as reduced ecological validity and demand characteristics.
- Distinguish a field experiment from a natural experiment by identifying who manipulates the independent variable.
- Describe a realistic example of each design and explain what the researcher controls and measures.
- Evaluate how reduced control over extraneous variables in these designs can affect the validity of conclusions.
- Describe structured, unstructured and semi-structured interviews and how their question formats differ.
- Explain the strengths and weaknesses of interviews, including the type and quality of data they produce.
- Evaluate how interviewer effects and social desirability can affect the validity of interview data.
- Identify a questionnaire from a description of a research method.
- Distinguish open from closed questionnaire items and the data each produces.
- Evaluate one strength and one weakness of using questionnaires.
- Identify a case study from a description of a research method.
- Explain why case studies often use more than one method.
- Evaluate one strength and one weakness of case studies, including generalisability.
- Identify what is meant by categories of behaviour in an observation study.
- Explain how interobserver reliability is checked by comparing independent observers' records.
- Apply the idea of operational definitions to a given observation scenario.
- State a strength and a weakness of a named research method.
- Explain which type of research a given method is suitable for and why.
- Evaluate the suitability of a method for a specific research aim.
- Define correlation as an association between two co-variables.
- Distinguish positive from negative correlations using examples.
- Interpret the direction and strength of points on a scatter diagram and state that correlation does not prove causation.
- Explain at least one strength and one weakness of using correlations.
- Apply strengths and weaknesses to a concrete research example.
- Reach a reasoned judgement about the suitability of a correlation for a given research aim.
- Identify an appropriate measure for a described research situation.
- Explain in words what a reported value indicates about the data.
- Justify a choice of measure without performing formula computation.
- Describe standardised procedures and identical instructions to participants.
- Explain how randomisation, allocation to conditions and counterbalancing reduce bias and order effects.
- Explain the effect of an extraneous variable and how to control for it.
- Explain how planning choices affect the reliability of a study
- Explain how planning choices affect the validity of a study
- Apply reliability and validity considerations to a planned piece of research
- Identify different sampling methods and describe how each selects participants
- Evaluate the strengths and weaknesses of each sampling method
- Explain how sampling affects the generalisability of research findings
- Describe independent groups, repeated measures and matched pairs designs.
- Explain one strength and one weakness of each experimental design.
- Select and justify an appropriate design for a described experiment.
- Distinguish quantitative from qualitative data using examples.
- Classify a described method or data set as producing quantitative or qualitative data.
- Evaluate one strength and one limitation of each type of data for a given research aim.
- Identify the ethical principle illustrated in a described research scenario.
- Select the appropriate safeguard that would address a given ethical issue.
- Distinguish between closely related ethical principles such as confidentiality and anonymity.
- Identify the BPS ethical issue illustrated in a described study.
- Select the guideline-consistent action a researcher should take in a given scenario.
- Distinguish between ethical issues that are easily confused, such as deception and debriefing.
- State a practical way of dealing with informed consent and deception.
- Match each ethical issue to an appropriate method of dealing with it.
- Explain how confidentiality, privacy and the right to withdraw can be protected.
Research methods exam tips
Marking Points
- Defines a testable hypothesis as an operationalised prediction linking the independent variable to the dependent variable with stated measurement.
- Identifies the alternative hypothesis as predicting that a difference or relationship will occur between conditions.
- Identifies the null hypothesis as predicting no difference or relationship, with any variation attributed to chance.
- Explains that both hypotheses are formulated before data collection so the result can be tested against them.
- Recognises that operationalising variables makes a hypothesis testable rather than vague.
- Defines the independent variable as the factor manipulated or selected by the researcher to create conditions.
- Defines the dependent variable as the factor measured to assess the effect, ideally operationalised with a stated measure.
- Defines extraneous variables as other factors that could influence the dependent variable if not controlled.
- Explains how extraneous variables are controlled, for example through standardised procedures, random allocation or counterbalancing.
- Distinguishes extraneous variables from the independent and dependent variables in a described study.
- Defines the target population as the whole group about whom the researcher wishes to draw conclusions, and the sample as the smaller subgroup actually studied.
- States that the sampling method is the procedure used to select participants from the target population.
- Identifies the target population and the sample correctly when given a short research scenario.
- Explains that the choice of sampling method affects how representative the sample is and therefore how far findings can be generalised.
- Recognises that a biased or unrepresentative sample weakens conclusions drawn about the target population.
- Defines random sampling as a method where every member of the target population has an equal chance of being selected.
- Describes a valid procedure, such as numbering all members of the target population and using a random number generator or drawing lots to select participants.
- Explains that random selection reduces researcher bias and can improve representativeness of the sample.
- Identifies a limitation, such as the need for a complete list of the target population or the risk that selected participants refuse to take part.
- Recognises that an opportunity sample uses participants who are available and willing at the time and place of the research.
- Distinguishes opportunity sampling from random, systematic, stratified and volunteer sampling by its reliance on whoever is present.
- Identifies a strength: it is quick, easy and inexpensive to obtain participants.
- Identifies a weakness: the sample may be biased or unrepresentative, limiting generalisation to the target population.
- Recognises that systematic sampling selects every nth participant from a sampling frame using a fixed interval.
- Distinguishes systematic sampling from random sampling (equal chance) and opportunity sampling (whoever is available).
- Identifies a strength: it uses an objective, unbiased rule that is quick and easy to apply.
- Identifies a weakness: bias can occur if the list has a repeating pattern that coincides with the interval.
- Identifies that the population is first divided into subgroups or strata based on a shared characteristic.
- Recognises that participants are then selected randomly from within each stratum.
- Notes that the number drawn from each stratum is proportional to that stratum's size in the population.
- Distinguishes stratified sampling from opportunity, volunteer and systematic sampling by its deliberate proportional subgroup structure.
- Acknowledges that it requires prior knowledge of the population's composition.
- States a strength of a named sampling method, such as reduced bias or improved representativeness, and explains why.
- States a weakness of the same or another named method, such as time, cost, or unrepresentativeness, and explains why.
- Covers more than one sampling method, comparing their relative strengths and weaknesses.
- Links strengths and weaknesses to research concepts such as bias, generalisability, validity or practicality.
- Uses a concrete example, such as opportunity sampling of passers-by producing a biased sample, to illustrate the point.
- Defines the target population as the whole group the researcher wants to draw conclusions about, and the sample as the subset actually studied.
- Explains representativeness: the sample should reflect key characteristics of the target population so results can be generalised.
- Describes random sampling as giving every member of the population an equal chance of being selected, reducing bias.
- Identifies a weakness of opportunity sampling: it uses whoever is available and may be unrepresentative.
- Identifies a weakness of volunteer sampling: self-selecting participants may be motivated or atypical, creating bias.
- Recognises that a larger sample is generally more representative, but size alone does not remove bias.
- Defines quantitative data as numerical data that can be counted, measured and analysed statistically.
- Defines qualitative data as non-numerical, descriptive data such as words or themes.
- Gives a valid example of a quantitative method, such as a closed rating scale or a timed task producing scores.
- Gives a valid example of a qualitative method, such as open-ended interview answers producing detailed accounts.
- States a strength of quantitative data: it is easy to compare, summarise and analyse.
- States a strength of qualitative data: it provides rich, detailed insight into experience.
- Explains that the choice of method should match the research aim, and that mixed methods can combine both.
- Identifies that the experimental method manipulates an IV and measures a DV to test a causal hypothesis while controlling extraneous variables.
- States that independent groups uses different participants in each condition, avoiding order effects but risking participant variables and needing more participants.
- States that repeated measures uses the same participants in all conditions, controlling participant variables and needing fewer participants but producing order effects that require counterbalancing.
- States that matched pairs uses different but paired participants, reducing participant variables and avoiding order effects, but matching is time-consuming and imperfect.
- Correctly matches a strength or weakness to the correct design rather than confusing designs.
- Defines a laboratory experiment as one carried out in a controlled setting where the IV is manipulated and extraneous variables are controlled.
- States that high control allows cause and effect to be established.
- States that standardised procedures allow replication.
- Identifies reduced ecological validity from the artificial setting or tasks.
- Identifies demand characteristics or participant awareness as a possible weakness.
- A field experiment involves the researcher deliberately manipulating the independent variable in a natural, everyday setting rather than a laboratory.
- A natural experiment involves a naturally occurring change in the independent variable that the researcher does not manipulate or control.
- In a field experiment the researcher still controls and measures variables, but the setting is realistic and participants may be unaware they are being studied.
- In a natural experiment the researcher measures the effect of the naturally occurring change on a dependent variable, often comparing conditions that already exist.
- The key distinguishing question is who caused the change in the independent variable: the researcher (field) or events outside the researcher's control (natural).
- Both designs can have reduced control over extraneous variables compared with laboratory experiments, which can affect the validity of conclusions.
- An interview is a method where a researcher asks a participant questions to collect data, typically face to face but also by telephone or online.
- Structured interviews use the same fixed questions in the same order, producing consistent, comparable data that are easier to analyse but allow little probing.
- Unstructured interviews use open, flexible questions, allowing the interviewer to follow up and gain rich detail, but responses are harder to compare and analyse.
- Semi-structured interviews combine set questions with the freedom to probe further, balancing consistency with detail.
- Interviews can gather detailed qualitative data and allow clarification of answers, but may be affected by interviewer effects and social desirability bias.
- Analysing interview data, especially from unstructured interviews, can be time-consuming because responses must be transcribed and interpreted.
- Identifies a questionnaire as a self-report method using a written set of questions answered by participants.
- Distinguishes closed questions, which give fixed response options and yield quantitative data, from open questions, which allow free responses and yield qualitative data.
- States a strength such as reaching a large sample quickly and cheaply, or standardised questions allowing replication.
- States a weakness such as social desirability bias, misunderstanding of items, or low response rates reducing validity.
- Recognises that questionnaires can be administered on paper, by post or online.
- Identifies a case study as an in-depth investigation of a single individual, small group or event.
- Recognises that case studies often combine several methods, such as interviews, observations and tests, to gather detailed data.
- States a strength such as producing rich, detailed data or being useful for studying rare or unusual cases.
- States a weakness such as limited generalisability because the sample is one case, or possible researcher bias.
- Recognises that case studies can explore context and change over time.
- Defines categories of behaviour as a checklist or coding scheme of clearly defined, operationalised behaviours to record during observation.
- Explains that categories should be mutually exclusive and comprehensive so each behaviour is recorded once and relevant acts are not missed.
- Defines interobserver reliability as the consistency of records produced by two or more observers coding the same behaviour independently.
- Describes how interobserver reliability is checked: observers code the same session separately and their records are compared for agreement.
- Explains that low interobserver agreement indicates unclear categories or inconsistent coding, prompting revision of the scheme.
- States a strength and a weakness of a named research method, for example that experiments allow control and causal inference but may lack ecological validity.
- Explains that observations record natural behaviour with good ecological validity but risk observer bias and cannot easily demonstrate cause and effect.
- Explains that self-report methods such as questionnaires and interviews are efficient but rely on honesty and accurate memory and may use leading questions.
- Explains that correlations show relationships between variables but cannot establish causation.
- Explains that case studies provide rich, detailed data about an individual or group but generalise poorly.
- Matches a method to a type of research for which it is suitable, for example using an experiment when the aim is to test cause and effect.
- Defines correlation as an association between two co-variables rather than a cause-and-effect relationship.
- Identifies a positive correlation as both co-variables increasing together and a negative correlation as one increasing while the other decreases.
- Explains that a scatter diagram plots one point per participant, with one co-variable on each axis.
- Describes how the pattern, direction and tightness of plotted points indicate the strength and direction of a possible relationship.
- States that a correlation does not demonstrate causation.
- Explains a strength: correlations allow study of relationships between variables that cannot ethically or practically be manipulated.
- Explains a strength: correlations can suggest hypotheses and directions for further research, and use existing data efficiently.
- Explains a weakness: correlation does not demonstrate causation, because a third variable may account for the association.
- Explains a weakness: correlations cannot establish which co-variable influences the other, and may overlook non-linear relationships.
- Links strengths and weaknesses to a concrete example, such as the association between sleep and exam performance.
- Reaches a reasoned judgement about when a correlation is an appropriate method.
- Standardised procedures keep the task, materials and setting the same for all participants so results are comparable.
- Instructions to participants should be identical for everyone to prevent variation from how the study is explained.
- Randomisation assigns participants or conditions without researcher bias, helping to make groups equivalent.
- Allocation to conditions places participants into experimental groups, and random allocation reduces bias.
- Counterbalancing varies the order of conditions in repeated measures designs to control order effects.
- Extraneous variables are unwanted factors that could affect the dependent variable; their effect must be explained and controlled, for example by standardising procedures or randomising allocation.
- Random sampling: every member of the target population has an equal chance of selection, which reduces selection bias.
- Opportunity sampling: uses participants who are readily available, so it is convenient but may produce an unrepresentative sample.
- Volunteer sampling: participants put themselves forward, which is easy to arrange but may attract a biased type of participant.
- Systematic sampling: selects every nth member of a list, giving a structured but potentially biased selection if the list is ordered.
- Stratified sampling: the sample reflects the proportions of subgroups in the target population, improving representativeness.
- Identifies independent groups as different participants in each condition, with the strength of no order effects and the weakness of participant variables and larger samples.
- Identifies repeated measures as the same participants in all conditions, with the strength of controlled participant variables and the weakness of order effects requiring counterbalancing.
- Identifies matched pairs as participants paired on a relevant variable with one of each pair per condition, reducing participant variables but being time-consuming and imperfect.
- Selects the design that matches a described study and justifies the choice by reference to order effects, participant variables or sample demands.
- Defines quantitative data as numerical information that can be counted or measured.
- Defines qualitative data as non-numerical information describing meaning, experience or description.
- Classifies a described method or data set correctly as quantitative or qualitative.
- Explains an advantage of quantitative data, such as ease of comparison or statistical analysis.
- Explains an advantage of qualitative data, such as richer detail about participants' experiences.
- Recognises that a single method can generate both quantitative and qualitative data.
- Identifies informed consent as giving participants enough information to decide whether to take part, with the right to withdraw respected.
- Recognises confidentiality as keeping participants' data anonymous and not identifiable in reports.
- Recognises protection from harm as avoiding physical or psychological harm, including stress, and using debriefing where needed.
- Identifies deception as misleading participants, acceptable only when necessary and followed by a full debrief.
- Selects the safeguard that matches the scenario rather than a merely related principle.
- Links informed consent to the BPS requirement that participants understand what they are agreeing to and may withdraw.
- Links confidentiality to the BPS requirement that participants' data are kept private and not identifiable.
- Links protection from harm to the BPS requirement to avoid physical and psychological harm and to debrief.
- Links deception to the BPS position that it must be justified and followed by full debriefing.
- Selects the guideline that specifically applies to the described scenario rather than a related but different principle.
- Identifies informed consent and a way of dealing with it, such as a consent form or presumptive consent.
- Identifies deception and a way of dealing with it, such as a full debrief after the study.
- Identifies protection from harm and a way of dealing with it, such as risk assessment, a withdrawal procedure or follow-up care.
- Identifies confidentiality and a way of dealing with it, such as anonymising data with codes or numbers.
- Identifies privacy and a way of dealing with it, such as not observing people in places where they would not expect to be observed.
- Identifies the right to withdraw and a way of dealing with it, such as telling participants clearly at the start and reminding them throughout.
Examiner Tips
- 💡Underline the independent and dependent variables in any hypothesis before deciding whether it is testable.
- 💡Check that the null hypothesis explicitly says 'no difference' or 'no relationship' and mentions chance.
- 💡Practise rewriting a vague aim into an operationalised alternative hypothesis and its matching null.
- 💡For any study, ask 'what did the researcher change?' for the independent variable and 'what did they measure?' for the dependent variable.
- 💡List at least one extraneous variable and state a practical control for it.
- 💡Operationalise the dependent variable with a concrete measure to show it is testable.
- 💡Underline the group named in the scenario to pin down the target population before choosing an answer.
- 💡Check whether an option describes who is selected rather than how data are gathered.
- 💡Eliminate options that confuse sample size with representativeness.
- 💡Look for the phrase equal chance of being selected as the defining feature of random sampling.
- 💡When describing the procedure, mention the full list of the target population and a random selection device.
- 💡Separate the strength (reduced bias) from the limitation (needs a complete list, possible refusal) in your answer.
- 💡Read the scenario carefully and look for words like 'available', 'willing' or 'whoever was there' as clues for opportunity sampling.
- 💡When asked for a weakness, link it to generalisation rather than just saying 'it is bad'.
- 💡Eliminate options that describe random or systematic procedures before choosing your answer.
- 💡Look for a stated interval such as 'every 10th person' as the clue for systematic sampling.
- 💡When asked for a strength, mention the objective rule or the fixed interval rather than just 'it is fair'.
- 💡Rule out options describing 'whoever is available' (opportunity) or 'equal chance' (random) before answering.
- 💡Look for the key phrase about dividing the population into groups and sampling proportionally from each.
- 💡Eliminate options describing non-probability methods such as opportunity or volunteer sampling first.
- 💡Check whether the option mentions random selection within subgroups, which is the defining feature.
- 💡Structure your answer method by method, giving a strength and a weakness for each.
- 💡Use research terminology such as bias, representativeness and generalisability to gain credit.
- 💡Support each point with a brief example to show understanding rather than just naming the method.
- 💡Learn a one-line definition of population, sample and representativeness before answering.
- 💡When a question names a sampling method, link it to one strength or weakness of bias.
- 💡Read all options before choosing; eliminate options that confuse sample size with representativeness.
- 💡Link each method to the type of data it produces before judging its strength.
- 💡Use the research aim to justify a choice between quantitative and qualitative methods.
- 💡Give a concrete example of the data, such as scores or written answers, to show understanding.
- 💡Read the stem carefully to identify which design is being described before choosing an answer.
- 💡Link each strength or weakness explicitly to its design, for example order effects belong to repeated measures.
- 💡Eliminate options that swap the features of two designs, such as attributing counterbalancing to independent groups.
- 💡Look for clues about control and setting in the stem before selecting an answer.
- 💡Match each strength or weakness to the feature that causes it, such as control leading to causal conclusions.
- 💡Reject options that describe field or natural experiments when the stem describes a controlled setting.
- 💡Ask yourself who changed the independent variable; if the researcher did, it is a field experiment, and if events did, it is a natural experiment.
- 💡Use a concrete example for each design to show you understand the difference, such as music in a real shop for field and a school relocation for natural.
- 💡When evaluating, link reduced control of extraneous variables to a specific effect on validity rather than just saying it is less controlled.
- 💡Name the type of interview (structured, unstructured or semi-structured) and link its features to strengths and weaknesses of the data it produces.
- 💡Use a concrete example, such as asking fixed questions about sleep habits for structured interviews versus open questions about feelings for unstructured ones.
- 💡When evaluating, connect interviewer effects or social desirability to a specific effect on the validity of the data rather than stating it vaguely.
- 💡Read the stem carefully to check whether it describes written self-report or a face-to-face exchange before choosing.
- 💡Link each strength or weakness to a named concept such as validity, reliability or response rate rather than vague praise or criticism.
- 💡When asked to distinguish question types, check whether fixed options (closed) or free responses (open) are described.
- 💡Check whether the stem describes depth of detail about one case rather than breadth across many participants.
- 💡Tie strengths and weaknesses to named concepts such as generalisability, validity and bias.
- 💡When a stem mentions several data-collection techniques used together, consider whether a case study is being described.
- 💡Learn the term 'operational definition' and be ready to apply it to a named behaviour in a scenario.
- 💡When asked about reliability, always mention two observers coding independently and comparing records.
- 💡Link unclear categories directly to low agreement, showing cause and effect in your answer.
- 💡Name the method explicitly before giving its strength or weakness so the examiner can credit the link.
- 💡Use the phrase 'suitable for' when explaining which type of research a method fits, and justify it with the aim.
- 💡Balance each strength with a corresponding weakness to show evaluation rather than description.
- 💡Read the question carefully to check whether it asks about direction (positive or negative) or strength (strong or weak) of the correlation.
- 💡When describing a scatter diagram, name both axes and say what the overall pattern of points shows.
- 💡Always add that correlation does not show causation when asked to interpret a relationship.
- 💡Balance your answer by giving at least one strength and one weakness, each with an explanation.
- 💡Use a concrete example to show how the strength or weakness applies in practice.
- 💡Finish with a short judgement about when a correlation is the most suitable method.
- 💡When a question gives you a value, spend your answer explaining what it means for the study rather than reworking the arithmetic.
- 💡Justify your choice of measure by linking it to the type of data and the aim of the research, using words such as 'because the data are...'.
- 💡Practise reading short study descriptions and naming an appropriate measure, then writing one sentence of justification.
- 💡Check the command word: if it asks you to explain or justify, give reasoning rather than a bare label.
- 💡Name the specific procedure and then state what it controls, for example 'counterbalancing controls order effects'.
- 💡When asked about extraneous variables, always give both the effect on the dependent variable and a control method.
- 💡Distinguish clearly between randomisation, random allocation and random sampling in your answer.
- 💡Use a concrete example, such as varying the order of two conditions across participants, to show understanding.
- 💡When planning, write one sentence for each design choice explaining how it protects reliability or validity.
- 💡Use a concrete example, such as standardised instructions for consistency or a pilot to check a task measures the intended behaviour.
- 💡Separate your reliability points from your validity points so each is clearly justified.
- 💡Match each sampling method to its key strength and weakness before answering.
- 💡Link the method to representativeness and generalisability of findings.
- 💡Read the question carefully to identify which specific sampling method is being described.
- 💡Read the scenario and name the design explicitly before weighing one strength and one weakness.
- 💡Link each weakness to a concrete consequence, such as fatigue lowering scores in the second condition.
- 💡Use precise terms: participant variables, order effects, counterbalancing and matched on a relevant variable.
- 💡Look at the actual data described in the scenario, not just the name of the method, before classifying it.
- 💡When justifying a choice, link the type of data to the research aim rather than stating a generic preference.
- 💡Use precise terms such as numerical, non-numerical, statistical analysis and rich description to show understanding.
- 💡Read the scenario and name the specific principle being tested before looking at the options.
- 💡Eliminate options that describe a different principle from the one the question asks about.
- 💡Check whether the question asks for the principle breached or the safeguard that would prevent it, as these differ.
- 💡Name the specific BPS ethical issue in the scenario before evaluating the options.
- 💡Reject options that describe a safeguard belonging to a different ethical issue.
- 💡Where two options seem ethical, choose the one that directly addresses the issue raised in the scenario.
- 💡Learn one clear method for each named issue so you can match them quickly in an MCQ.
- 💡Watch for options that name an issue but give a method that belongs to a different issue.
- 💡Read all options before choosing, because a plausible-sounding method may address the wrong issue.
Common Mistakes
- Writing a vague prediction such as 'caffeine affects people' and calling it testable; correction: specify the independent variable, dependent variable and how each is measured.
- Confusing the null hypothesis with a failed study; correction: the null hypothesis is a prediction of no difference or relationship, stated before data are collected.
- Assuming the alternative hypothesis must be proven correct; correction: data are analysed to decide whether to accept the alternative or retain the null.
- Swapping the independent and dependent variables; correction: the independent variable is manipulated or selected, the dependent variable is measured.
- Treating extraneous variables as the same as the independent variable; correction: extraneous variables are unwanted influences that should be controlled.
- Leaving the dependent variable vague, such as 'memory'; correction: operationalise it, for example number of words correctly recalled.
- Confusing the sample with the target population: the correction is that the target population is the whole group of interest and the sample is the subset selected from it.
- Assuming any sample automatically represents the population: the correction is that representativeness depends on how the sample was selected.
- Treating the sampling method as the same thing as the research method: the correction is that sampling concerns who is chosen, whereas the research method concerns how data are collected.
- Describing random sampling as simply picking people at random from whoever is nearby: the correction is that every member of the target population must have an equal chance of selection, which requires a full list.
- Confusing random sampling with random allocation of participants to conditions: the correction is that sampling concerns who is chosen, whereas allocation concerns which condition chosen participants are placed in.
- Claiming random sampling guarantees a perfectly representative sample: the correction is that it reduces bias but does not guarantee representativeness.
- Confusing opportunity sampling with random sampling: opportunity uses whoever is available, whereas random sampling gives everyone in the target population an equal chance of selection.
- Assuming an opportunity sample is automatically representative: it reflects only those present at that time and place, so bias is likely.
- Thinking opportunity sampling requires a sampling frame or list of the population: it does not, which is part of why it is convenient.
- Confusing systematic sampling with random sampling: systematic uses a fixed interval, whereas random gives everyone an equal chance.
- Forgetting that a sampling frame (a list of the population) is normally needed for systematic sampling.
- Assuming systematic sampling is always unbiased: a repeating pattern in the list can create bias.
- Confusing stratified sampling with opportunity sampling: correction is that stratified sampling is a probability method using random selection within defined strata, not whoever is available.
- Thinking strata are chosen randomly rather than defined by the researcher: correction is that the researcher identifies the subgroups, then selects randomly within them.
- Assuming equal numbers are taken from each stratum: correction is that numbers are proportional to each stratum's size in the population.
- Listing methods without stating strengths or weaknesses: correction is to pair each method with at least one strength and one weakness and explain each.
- Giving a strength without explaining why it is a strength: correction is to link it to a research concept such as reduced bias or practicality.
- Treating a weakness as a strength, for example calling opportunity sampling representative: correction is to recognise that convenience samples are typically biased.
- Confusing the sample with the target population: the correction is that the population is the whole group and the sample is the subset studied.
- Assuming a large sample is automatically representative: the correction is that a large but biased sample still gives biased results.
- Thinking random sampling means choosing people at random by preference: the correction is that every member must have an equal chance of selection.
- Assuming qualitative data is always better because it is detailed: the correction is that it is harder to summarise and compare, so the aim decides the best choice.
- Thinking any number makes data quantitative: the correction is that data must be numerical and analysable as quantities, not just labelled with a number.
- Believing quantitative data cannot be biased: the correction is that how numbers are collected and interpreted can still introduce bias.
- Confusing repeated measures with independent groups: the correction is that repeated measures uses the same participants in every condition, whereas independent groups uses different participants.
- Claiming repeated measures has no weaknesses: the correction is that it produces order effects such as practice and fatigue, which counterbalancing only partly reduces.
- Saying matched pairs eliminates all participant variables: the correction is that matching reduces but cannot remove individual differences because matching is never perfect.
- Assuming a laboratory experiment must occur in a room called a laboratory: the correction is that it refers to any highly controlled setting.
- Confusing laboratory experiments with field experiments: the correction is that field experiments occur in a natural setting where the researcher still manipulates the IV.
- Claiming laboratory experiments always lack validity: the correction is that control can raise internal validity even though ecological validity may be reduced.
- Thinking a natural experiment means any experiment done outdoors; correction: the defining feature is that the researcher does not manipulate the IV, not the location.
- Confusing field and natural experiments; correction: in a field experiment the researcher manipulates the IV, whereas in a natural experiment the change occurs without the researcher's control.
- Assuming a field experiment has no control at all; correction: the researcher still manipulates the IV and measures the DV, but control over extraneous variables is weaker than in a lab.
- Believing participants in a natural experiment are always unaware; correction: awareness depends on the situation, and the defining feature is the lack of researcher manipulation of the IV.
- Thinking all interviews are the same; correction: structured, unstructured and semi-structured interviews differ in how fixed the questions are and how much probing is allowed.
- Assuming interviews only produce quantitative data; correction: interviews often produce detailed qualitative data, though structured formats can yield more comparable answers.
- Ignoring interviewer effects; correction: the interviewer's presence, tone or expectations can influence how participants respond, affecting validity.
- Believing unstructured interviews are easy to analyse; correction: flexible, open responses are harder to compare and analyse than fixed structured answers.
- Confusing a questionnaire with an interview: correction — a questionnaire is completed in writing by the respondent, whereas an interview involves a researcher asking questions directly.
- Assuming all questionnaire data are qualitative: correction — closed questions produce quantitative data, while open questions produce qualitative data.
- Claiming questionnaires guarantee honest answers: correction — respondents may give socially desirable answers, so validity can be reduced.
- Treating a case study as a large-scale survey: correction — a case study focuses in depth on one individual, small group or event rather than many participants.
- Assuming case study findings generalise to the whole population: correction — because only one case is studied, generalisability is limited.
- Believing a case study uses only one method: correction — case studies typically combine several methods to build a detailed picture.
- Confusing interobserver reliability with interobserver validity: reliability concerns consistency between observers, not whether the observation measures what it claims.
- Assuming one observer is enough to establish reliable categories: reliability requires comparison of at least two independent observers' records.
- Writing vague categories such as 'being naughty' without operational definitions, which makes consistent coding impossible; categories need observable, defined behaviours.
- Listing generic strengths and weaknesses without linking them to a named method or research aim; each point should be tied to the specific method and purpose.
- Claiming a correlation proves causation; correlations show relationships only, and causation requires controlled manipulation.
- Treating one method as universally best; suitability depends on the research question, so match method to aim.
- Saying a correlation proves that one variable causes the other; correct this by stating it shows only an association, so a third factor could explain both.
- Confusing negative correlation with 'no correlation'; correct this by explaining that a negative correlation is a real relationship in which one co-variable falls as the other rises.
- Plotting a bar chart or line graph instead of a scatter diagram; correct this by using one point per participant with one co-variable on each axis.
- Assuming a scatter diagram shows exact values for each individual's cause and effect; correct this by treating the pattern of points as evidence of a possible association only.
- Claiming a correlation proves cause and effect; correct this by stating it shows only an association and a third variable may be responsible.
- Listing strengths and weaknesses without explaining them; correct this by saying why each point matters, for example why studying an unmanipulable variable is useful.
- Ignoring the direction of the relationship; correct this by noting that a correlation cannot show which co-variable influences the other.
- Treating a scatter diagram as proof of a strong relationship; correct this by explaining that the pattern of points only suggests a possible association.
- Believing you must memorise and apply formulae to earn marks: the specification states computation is not required, so focus on selecting and interpreting measures and justifying choices in words.
- Confusing interpretation with calculation, for example trying to derive a value by hand instead of explaining what a supplied value indicates about the data.
- Assuming 'no computation' means the topic can be ignored: you still need to know what each measure shows, when it is used and how to read it.
- Treating a reported correlation coefficient as proof of cause: interpret it as strength and direction of relationship only.
- Confusing randomisation with random sampling: randomisation concerns assigning participants or conditions, whereas random sampling concerns how participants are selected; correct by stating which process is being described.
- Thinking counterbalancing removes all extraneous variables: it specifically controls order effects in repeated measures designs; correct by naming the specific control for each threat.
- Describing an extraneous variable without explaining its effect or control: correct by stating how it could change the dependent variable and one method to reduce it.
- Treating standardised procedures as identical to instructions to participants: instructions are one part of standardisation, not the whole of it; correct by distinguishing the broader procedure from the specific instructions.
- Treating reliability and validity as the same thing. Correction: reliability is about consistency across repeats; validity is about whether you measure what you intend to measure.
- Assuming a reliable study must be valid. Correction: a procedure can give consistent results while measuring the wrong thing, so check both separately.
- Thinking validity only concerns ethics. Correction: ethical approval matters, but validity concerns the accuracy and appropriateness of what is measured and how.
- Confusing random sampling with opportunity sampling. Correction: random gives everyone an equal chance; opportunity uses whoever is available.
- Believing a larger sample automatically removes bias. Correction: size does not fix a biased selection method; representativeness depends on how participants are chosen.
- Thinking volunteer sampling is unbiased because people choose freely. Correction: volunteers may share characteristics that make them unrepresentative.
- Confusing independent groups with matched pairs: correct this by stressing that matched pairs deliberately pair people on a relevant variable before splitting them, whereas independent groups allocate without pairing.
- Claiming repeated measures removes all extraneous variables: correct this by saying it controls participant variables but introduces order effects, which counterbalancing only partly reduces.
- Thinking counterbalancing removes order effects completely: correct this by explaining it balances them across the sample rather than eliminating them for each individual.
- Assuming a questionnaire is always quantitative: correction is that open-ended questions produce qualitative responses, so judge the data produced.
- Treating qualitative data as inferior or unusable: correction is that it provides depth and insight that numerical data may miss.
- Confusing qualitative with a small sample: correction is that the distinction concerns the type of data, not the number of participants.
- Confusing confidentiality with anonymity: confidentiality is about not disclosing who the participant is, while anonymity means their identity is not collected or is removed; the correction is to match the option to the precise meaning.
- Treating deception as always unacceptable: the correction is that deception may be justified when necessary, provided participants are fully debriefed afterwards.
- Assuming informed consent means simply signing a form: the correction is that participants must be given sufficient information about the study to make a genuine choice.
- Choosing an option because it sounds ethical rather than because it fits the scenario: the correction is to identify which specific principle the scenario breaches.
- Thinking the BPS guidelines are optional advice: the correction is that they set out the ethical standards researchers are expected to follow.
- Confusing the role of the debrief with that of informed consent: the correction is that consent happens before participation, while debriefing happens afterwards to explain the study and address any harm.
- Believing any use of deception breaches the guidelines outright: the correction is that deception may be acceptable when necessary and fully debriefed.
- Matching a scenario to a principle that is merely associated with it rather than the one actually breached: the correction is to identify the specific issue first.
- Thinking a debrief removes the need for consent: correction, a debrief addresses deception after the event, but consent should still be sought beforehand where possible.
- Believing confidentiality means simply not telling anyone: correction, it means protecting identity, for example by anonymising data.
- Assuming the right to withdraw only applies at the start: correction, participants should be reminded they can withdraw at any point, including during the study.