Designing an investigation | AQA A-Level Psychology Revision
- Revision Notes
- Aug 5
- 33 min read
Updated: 6 days ago
For 7182 specification, first teach in September 2025
AQA A-Level Psychology | Free Revision Notes
Estimated study time: 70 minutes
This Designing an investigation A-Level Psychology revision page explains how to turn a research question into a practical, ethical and scientifically useful study. You will learn how to develop an aim and hypothesis, select a research method, sample and experimental design, and operationalise the variables. You will also consider control, reliability, validity, ethics and risk. A successful investigation must fit together as one coherent plan, with every decision linked clearly to the behaviour being studied.
Learning Objectives 🎯
By the end of this revision page, you should be able to:
Develop a clear and appropriate research aim.
Write directional, non-directional and null hypotheses.
Select a suitable research method.
Define an appropriate target population and sampling method.
Select an experimental design where appropriate.
Identify and operationalise variables clearly.
Plan suitable control procedures.
Explain how reliability and validity could be improved.
Identify and address ethical issues.
Complete a basic risk assessment for a psychological investigation.
Produce a coherent investigation plan from an unfamiliar scenario.
Revision Notes 📚
Designing psychological research
Psychologists do not begin an investigation by collecting data immediately.
A research study must first be designed carefully so that:
the aim is clear;
the hypothesis is testable;
the method addresses the research question;
the sample is appropriate;
the variables are measurable;
the procedure is controlled and replicable;
the data will be reliable and valid;
participants are treated ethically;
foreseeable risks are managed.
Poor decisions at the design stage can weaken the entire investigation, even if the data are later analysed accurately.
The main stages of investigation design
A useful planning sequence is:
Identify the topic or research question.
Write the aim.
Write the hypothesis.
Select the research method.
Define the target population.
Select the sample and sampling method.
Select the experimental design, where relevant.
Identify and operationalise the variables.
Plan the materials and procedure.
Control relevant extraneous variables.
Address demand characteristics and investigator effects.
Consider reliability.
Consider validity.
Address ethical issues.
Assess and manage risk.
Plan how the data will be recorded and analysed.
These decisions should be connected. A design is not a collection of isolated choices.
Developing an appropriate aim
What is a research aim?
A research aim is a clear statement of the general purpose of an investigation.
It describes what the researcher intends to investigate but does not normally predict the result.
For example:
To investigate whether background speech affects the number of words recalled by sixth-form students.
This aim identifies:
the factor being studied, background speech;
the measured outcome, number of words recalled;
the intended population, sixth-form students.
Features of a strong aim
A strong aim should:
identify what is being investigated;
name the relevant variables or behaviours;
identify the population where appropriate;
be concise;
match the chosen method;
avoid predicting a result.
Aim and hypothesis compared
Aim | Hypothesis |
States the purpose of the study | Predicts the result |
Usually begins with “To investigate…” | Usually begins with “There will be…” |
Does not normally state a direction | May be directional or non-directional |
Identifies the variables or behaviour | States how the variables are expected to be related |
Broad research intention | Testable prediction |
The distinction is covered in more detail in aims and hypotheses.
Weak and improved aims
Weak aim
To study memory.
This is too vague because it does not identify:
what aspect of memory will be studied;
what will be changed or compared;
how memory will be measured;
who will be investigated.
Improved aim
To investigate whether revising in silence or with background speech affects the number of words recalled by sixth-form students.
This is clearer because it identifies:
the two conditions;
the measured outcome;
the population.
Aim for an experiment
An experimental aim usually identifies:
the independent variable;
the dependent variable.
For example:
To investigate the effect of sleep deprivation on participants’ reaction times.
The independent variable is sleep condition.
The dependent variable is reaction time.
Aim for a correlation
A correlational aim identifies two co-variables.
For example:
To investigate whether there is a relationship between hours of sleep and memory-test scores in sixth-form students.
Neither variable is manipulated.
Both are measured.
Aim for an observation
An observational aim identifies the behaviour and context.
For example:
To investigate the frequency of cooperative and competitive behaviours shown by students during a group problem-solving activity.
Aim for a self-report study
A self-report aim may identify:
the psychological construct being investigated;
the group being surveyed;
any comparison or relationship being examined.
For example:
To investigate sixth-form students’ reported levels of examination stress.
Developing an appropriate hypothesis
What is a hypothesis?
A hypothesis is a clear, testable prediction about the outcome of an investigation.
A hypothesis should be written before the results are collected.
The hypothesis must fit:
the aim;
the method;
the variables;
the research design.
Operationalised hypotheses
A hypothesis should explain exactly how the variables will be measured or manipulated.
This is called operationalisation.
A vague hypothesis might state:
Background speech will make memory worse.
This does not explain:
what background speech means;
how memory will be measured;
who will take part.
A more precise hypothesis is:
Sixth-form students will recall fewer words from a list of \(20\) words when a recorded conversation is played than when the task is completed in silence.
This identifies:
the population;
the two conditions;
the measurement of recall;
the predicted direction.
Directional hypothesis
A directional hypothesis predicts the direction of a difference or relationship.
It may predict that one condition will produce:
higher scores;
lower scores;
more behaviour;
fewer responses.
For example:
Sixth-form students will recall significantly fewer words in the background-speech condition than in the silent condition.
For a correlation:
There will be a significant negative correlation between students’ examination-stress ratings and their reported hours of sleep.
This predicts that higher stress will be associated with fewer hours of sleep.
When a directional hypothesis may be suitable
A directional hypothesis is appropriate when the researcher has a justified reason to predict the direction of the outcome.
The direction must be established before the data are inspected.
It should not be chosen after the researcher sees which condition produced the higher mean.
Non-directional hypothesis
A non-directional hypothesis predicts that a difference or relationship will exist but does not state its direction.
For example:
There will be a significant difference in the number of words recalled by sixth-form students in the silent and background-speech conditions.
For a correlation:
There will be a significant correlation between students’ examination-stress ratings and their reported hours of sleep.
When a non-directional hypothesis may be suitable
A non-directional hypothesis may be used when:
there is no clear justification for predicting a direction;
previous findings are inconsistent;
either possible direction would be meaningful.
Null hypothesis
A null hypothesis predicts that there will be no genuine difference, correlation or association.
Any apparent result is attributed to chance.
For example:
There will be no significant difference in the number of words recalled by sixth-form students in the silent and background-speech conditions. Any difference will be due to chance.
For a correlation:
There will be no significant correlation between examination-stress ratings and reported hours of sleep in sixth-form students. Any relationship will be due to chance.
Hypothesis checklist
Before finalising a hypothesis, ask:
Does it match the aim?
Does it identify the population?
Are both variables or conditions named?
Are the variables operationalised?
Is the predicted difference, correlation or association clear?
Is the direction stated if the hypothesis is directional?
Could the hypothesis be tested using the proposed data?
Selecting a research method
Method and design are not the same
Students often use the terms method and design as though they mean the same thing.
They refer to different decisions.
Decision | Meaning | Examples |
Research method | The overall way evidence is collected | Experiment, observation, questionnaire, interview, correlation |
Type of experiment | The setting or basis of the manipulation | Laboratory, field, natural or quasi-experiment |
Experimental design | How participants are allocated to conditions | Repeated measures, independent groups or matched pairs |
Sampling method | How participants are selected | Random, systematic, stratified, opportunity or volunteer |
Procedure | The exact sequence followed by researchers and participants | Instructions, task, timing and recording process |
A study could therefore use:
a laboratory experiment;
a repeated measures design;
an opportunity sample.
Each phrase describes a different feature.
Research methods in the specification
AQA requires knowledge of:
experimental methods;
observational techniques;
self-report techniques;
correlations;
content analysis;
case studies.
The selected method must answer the research question.
Experimental methods
When is an experiment appropriate?
An experiment is appropriate when the researcher wants to investigate the effect of an independent variable on a dependent variable.
The researcher manipulates or compares the independent variable and measures the dependent variable.
For example:
Does background speech affect memory-test performance?
The researcher could manipulate whether speech is present and measure the number of words recalled.
Laboratory experiment
A laboratory experiment takes place in a controlled environment.
It allows the researcher to:
manipulate the independent variable;
control relevant extraneous variables;
standardise the procedure;
measure the dependent variable precisely.
It may be suitable when control and replicability are priorities.
However, the artificial setting or task may reduce ecological validity.
Field experiment
A field experiment takes place in a natural setting, but the researcher manipulates the independent variable.
It may produce behaviour that is more natural than behaviour recorded in a laboratory.
However, controlling extraneous variables may be more difficult.
Natural experiment
In a natural experiment, the change in the independent variable occurs without direct manipulation by the researcher.
The researcher investigates the effect of a naturally occurring difference or event.
Natural experiments may allow investigation of variables that could not be manipulated practically or ethically.
However, participants cannot usually be randomly allocated to naturally occurring conditions.
Quasi-experiment
A quasi-experiment uses an independent variable based on an existing difference between people.
For example, a researcher might compare people from two pre-existing groups.
The independent variable is not manipulated by the researcher.
Selecting an experimental type
Ask:
Can the independent variable be manipulated?
Is manipulation ethical?
Is control important?
Is natural behaviour important?
Does a naturally occurring difference already exist?
Can participants be allocated to conditions?
Could the study be replicated?
Observational methods
When is an observation appropriate?
An observation is suitable when the researcher wants to record behaviour as it occurs.
It may be preferable when:
participants cannot describe the behaviour accurately;
self-report answers may be affected by social desirability;
actual behaviour is more important than reported attitudes;
the behaviour can be defined and recorded clearly.
Naturalistic observation
A naturalistic observation takes place in the behaviour’s usual setting.
It may produce more natural behaviour.
However, researchers may have less control over events and extraneous variables.
Controlled observation
A controlled observation takes place in a situation arranged or controlled by the researcher.
The researcher may be able to:
ensure that each participant experiences similar conditions;
make the target behaviour more likely to occur;
use standardised recording procedures.
Overt observation
In an overt observation, participants know that they are being observed.
This supports informed consent but may increase participant reactivity or demand characteristics.
Covert observation
In a covert observation, participants are unaware that they are being observed.
This may reduce changes in behaviour caused by awareness of the observer.
However, it creates ethical concerns involving consent and privacy.
Participant observation
In a participant observation, the researcher becomes part of the group being studied.
This may provide detailed insight but can make objectivity and consistent recording more difficult.
Non-participant observation
In a non-participant observation, the researcher observes without joining the group.
This may make systematic recording easier, although the researcher may gain less detailed understanding of the group’s experience.
Behavioural categories
An observational study requires clear behavioural categories.
A behavioural category should be:
observable;
measurable;
specific;
unambiguous;
distinct from other categories.
Weak category:
Friendly behaviour.
This is open to interpretation.
Improved category:
The participant offers an object to another group member without being asked.
Event and time sampling
An observational design may use:
event sampling;
time sampling.
Event sampling records every occurrence of a specified behaviour.
Time sampling records behaviour at predetermined time intervals.
The selected technique should match:
the frequency of the behaviour;
the duration of the observation;
the number of behaviours being recorded;
the practical demands on the observer.
Self-report methods
Questionnaires
A questionnaire contains written questions answered by participants.
It may include:
open questions;
closed questions.
Questionnaires may allow researchers to collect data from many participants efficiently.
However, responses may be affected by:
unclear questions;
response bias;
social desirability;
limited response options.
Interviews
An interview involves questions asked directly by a researcher.
A structured interview uses predetermined questions in a standardised order.
An unstructured interview allows greater flexibility and follow-up questions.
Structured interviews may be easier to replicate.
Unstructured interviews may produce greater detail but may be less reliable across participants.
Selecting question types
Closed questions may be suitable when the researcher needs:
quantitative data;
straightforward comparison;
responses that are easy to code.
Open questions may be suitable when the researcher needs:
qualitative data;
detailed explanations;
unexpected responses.
Questions should avoid:
leading wording;
ambiguity;
double questions;
unnecessary technical language;
assumptions about participants.
Correlations
When is a correlation appropriate?
A correlation is appropriate when the researcher wants to investigate whether two co-variables are related.
For example:
hours of sleep and memory score;
stress rating and wellbeing rating;
revision time and examination performance.
Both co-variables are measured.
Neither is manipulated.
Correlational design decisions
The researcher must:
operationalise both co-variables;
collect one paired score for each participant;
use a sample that contains a suitable range of scores;
select an appropriate measurement method;
avoid claiming causation from the result.
A correlation may show:
a positive relationship;
a negative relationship;
zero correlation.
It cannot establish that one co-variable caused the other.
Content analysis
When is content analysis appropriate?
A content analysis is suitable when the researcher wants to analyse communication or recorded material systematically.
The researcher develops coding categories and records the frequency or presence of particular content.
Material might include:
written responses;
interview transcripts;
media content;
recorded communication;
documents.
The coding categories must be clear enough for the content to be classified consistently.
Case studies
When is a case study appropriate?
A case study is an in-depth investigation of an individual, group, institution or event.
It may combine several sources of evidence.
A case study may be appropriate when:
the case is rare;
detailed information is required;
the behaviour cannot be reproduced experimentally;
the researcher wants to investigate development or change over time.
However, findings from one unusual case may not generalise to a wider population.
Selecting a sample
Population and sample
The target population is the complete group about whom the researcher intends to draw conclusions.
The sample is the smaller group that actually takes part.
For example:
target population: UK sixth-form students;
sample: \(40\) sixth-form students recruited from one college.
The sample should be described precisely.
Saying only:
Students will take part.
does not explain:
their age;
educational stage;
location;
relevant characteristics.
Sampling and generalisation are covered in populations and samples.
Questions to ask when defining the sample
Who is the target population?
How many participants are needed?
Is the group accessible?
Does the sample reflect the target population?
Could the sampling method introduce bias?
Can meaningful consent be obtained?
Are any participants especially vulnerable?
Will the sample provide an appropriate range of scores?
Sampling methods
Random sampling
In random sampling, every member of the target population has an equal chance of selection.
A researcher needs:
a complete list of the target population;
a random method of selection.
Random sampling can reduce researcher selection bias, although obtaining a complete sampling frame may be difficult.
Systematic sampling
In systematic sampling, participants are selected using a fixed interval from an ordered list.
For example, the researcher might select every tenth person after choosing a starting point.
The method is more systematic than opportunity sampling but still requires access to a suitable list.
Stratified sampling
In stratified sampling, the sample reflects relevant subgroups within the target population in the correct proportions.
The researcher must:
Identify the relevant strata.
Calculate the required number from each group.
Select participants from each stratum.
This may improve representativeness for the selected characteristics.
Opportunity sampling
In opportunity sampling, participants are selected because they are available and willing at the time.
It is often quick and practical.
However, the sample may be biased because accessible participants may differ from the wider target population.
Volunteer sampling
In volunteer sampling, participants respond to an advertisement, invitation or request.
It allows people to choose whether to take part.
However, volunteers may differ systematically from people who do not volunteer.
For example, they may be:
more interested in psychology;
more confident;
more motivated;
more willing to disclose information.
Selecting a sampling method
The best method depends on:
access to the population;
time and resources;
the need for representativeness;
the sensitivity of the research;
the availability of a sampling frame;
the required sample size.
A strong justification should explain why the method is suitable for the particular investigation.
Worked sampling decision
A researcher wants a sample reflecting the proportions of Year 12 and Year 13 students in a college.
The college contains:
\(60\%\) Year 12 students;
\(40\%\) Year 13 students.
The researcher wants:
$$N=50$$
participants.
The number of Year 12 students required is:
$$0.60\times50=30$$
The number of Year 13 students required is:
$$0.40\times50=20$$
A stratified sample would contain:
\(30\) Year 12 students;
\(20\) Year 13 students.
Selecting an experimental design
The three experimental designs
Where the study is an experiment, the researcher must decide how participants will take part in the conditions.
The three designs are:
repeated measures;
independent groups;
matched pairs.
These are covered fully in experimental designs.
Repeated measures design
In a repeated measures design, the same participants take part in every condition.
Advantages for the investigation
Participant variables are controlled because the same people complete each condition.
Fewer participants may be needed.
Scores can be compared within each participant.
Potential problem
Participants may experience order effects, including:
practice;
fatigue;
boredom;
awareness of the aim.
Control procedure
Counterbalancing may be used.
For two conditions, half the participants complete:
$$A\rightarrow B$$
and the other half complete:
$$B\rightarrow A$$
Independent groups design
In an independent groups design, different participants take part in each condition.
Advantages for the investigation
Each participant completes only one condition.
Order effects are avoided.
Participants are less able to compare conditions directly.
Potential problem
Participant variables may differ between the groups.
For example, one group may contain participants with stronger memories before the manipulation begins.
Control procedure
Random allocation may be used to assign participants to conditions.
This does not guarantee identical groups, but it reduces systematic researcher bias in allocation.
Matched pairs design
In a matched pairs design, participants are paired according to a relevant characteristic.
One member of each pair is placed in each condition.
Possible matching variables
Participants might be matched on:
age;
an initial memory score;
reading ability;
previous experience;
another variable relevant to the investigation.
Advantages for the investigation
Some participant variables are controlled.
Order effects are avoided.
Potential problems
Matching can take considerable time.
Participants cannot be matched on every relevant characteristic.
If one participant withdraws, the matched data may be affected.
Selecting an experimental design
Ask:
Could completing one condition influence performance in another?
Are participant variables likely to affect the dependent variable?
Is a sufficiently large sample available?
Can participants be matched accurately?
Could participants guess the aim if they complete both conditions?
Is counterbalancing practical?
Selecting and operationalising variables
Independent variable
The independent variable, or IV, is the variable manipulated or used to form conditions in an experiment.
For example:
Presence or absence of background speech.
This should be operationalised more precisely:
silence;
a recorded conversation played at the same volume for all participants.
Dependent variable
The dependent variable, or DV, is the outcome measured by the researcher.
Vague DV:
Memory performance.
Operationalised DV:
The number of words correctly recalled from a list of \(20\) words within two minutes.
This makes the measurement:
observable;
numerical;
repeatable.
Extraneous variables
An extraneous variable is any variable other than the IV that could affect the DV.
Possible extraneous variables in a memory investigation include:
room noise;
length of exposure to the word list;
difficulty of the words;
time allowed for recall;
instructions;
participant tiredness;
previous experience of memory tasks.
The researcher should identify the most relevant variables rather than stating that “everything will be controlled”.
Operationalisation
Operationalisation means defining a variable so that it can be manipulated or measured.
Operationalisation should be:
precise;
objective;
observable;
replicable;
relevant to the psychological construct.
This is developed in variables and operationalisation.
Operationalising an IV
General variable:
Noise.
Operationalised IV:
Participants complete the memory task either in silence or while a recording of two people speaking is played at the same volume.
Operationalising a DV
General variable:
Concentration.
Operationalised DV:
The number of targets correctly identified in a two-minute visual-search task.
Operationalising co-variables
For a correlation, both co-variables must be measured clearly.
General co-variable:
Sleep.
Operationalised co-variable:
The participant’s self-reported number of hours slept during the previous night.
General co-variable:
Memory ability.
Operationalised co-variable:
The number of words correctly recalled from a list of \(20\) words.
Operationalising observational behaviour
General behaviour:
Aggression.
Possible behavioural categories:
pushes another person;
hits another person;
shouts an insult;
takes an object from another person without permission.
The categories should not overlap unnecessarily.
Planning control procedures
Why control is needed
Control procedures reduce the influence of alternative explanations.
A well-designed experiment aims to ensure that the main systematic difference between conditions is the independent variable.
Control does not mean that every possible variable is eliminated.
It means that relevant variables are:
held constant;
randomised;
balanced;
measured;
accounted for in the procedure.
The main techniques are covered in control procedures.
Standardisation
A standardised procedure is kept consistent across participants and conditions.
Standardisation may include:
identical instructions;
the same task materials;
equal exposure times;
consistent room conditions;
the same equipment;
the same scoring system;
the same time allowed;
a scripted debrief.
Standardisation supports:
replicability;
reliability;
consistent treatment of participants.
Example of standardisation
All participants:
study the same \(20\) words;
view the list for two minutes;
complete a one-minute distractor task;
have two minutes to write recalled words;
receive the same written instructions.
Random allocation
Random allocation assigns participants to experimental conditions using a random process.
It is relevant to an independent groups design.
Possible methods include:
a random-number generator;
drawing condition labels;
another impartial random procedure.
Random allocation reduces systematic bias in group assignment.
It is different from random sampling.
Random sampling | Random allocation |
Selects people from the population | Places sampled participants into conditions |
A sampling decision | A design and control decision |
Aims to reduce selection bias | Aims to reduce allocation bias |
Counterbalancing
Counterbalancing controls order effects in repeated measures designs.
For two conditions:
half complete Condition A followed by B;
half complete Condition B followed by A.
The order should be allocated systematically or randomly rather than allowing participants to choose.
Randomisation
Randomisation may be used within the procedure.
For example, the researcher could:
randomise the order of questionnaire items;
randomise word-list order;
randomise stimulus presentation;
randomise trial order.
This reduces the likelihood that a fixed sequence systematically affects results.
Control groups
A control group provides a baseline against which another condition can be compared.
For example:
treatment group receives a relaxation programme;
control group does not receive the programme.
The groups should be treated similarly apart from the factor being investigated.
Controlling situational variables
Situational variables may include:
room temperature;
noise;
lighting;
equipment;
time of day;
interruptions;
instructions.
Possible controls include:
using the same room;
testing participants at similar times;
displaying identical instructions;
checking equipment before each session;
using the same researcher.
Controlling participant variables
Participant variables may include:
age;
prior knowledge;
ability;
motivation;
tiredness;
hearing or vision differences.
Possible design responses include:
repeated measures;
matching;
random allocation;
selecting participants using relevant criteria;
measuring a relevant characteristic.
Demand characteristics
Demand characteristics are cues that allow participants to guess the aim or expected behaviour.
Participants may then:
deliberately support the hypothesis;
deliberately oppose the hypothesis;
monitor their behaviour unusually closely.
Possible ways of reducing demand characteristics include:
using a standardised procedure;
avoiding suggestive instructions;
withholding the full aim where ethically justified;
including an appropriate alternative explanation of the task;
using an independent groups design where completing both conditions would reveal the aim;
debriefing participants afterwards.
Investigator effects
Investigator effects occur when a researcher’s behaviour or expectations influence:
participants;
data recording;
interpretation.
Possible controls include:
standardised instructions;
objective scoring criteria;
training observers;
using more than one observer;
keeping the researcher unaware of condition where possible;
using automated recording;
avoiding leading questions.
Designing for reliability
What is reliability?
Reliability refers to consistency.
A reliable method should produce similar results when:
repeated under comparable conditions;
used by different trained researchers;
applied to the same material consistently.
Reliability is covered fully in reliability.
Reliability in experiments
Reliability may be improved through:
standardised instructions;
standardised timing;
consistent materials;
objective measurement;
detailed procedural records;
a pilot study.
Reliability in questionnaires and interviews
Reliability may be improved by:
using clear questions;
avoiding ambiguous wording;
using the same question order;
using a structured interview;
providing consistent scoring criteria;
conducting test-retest checks where appropriate.
Test-retest reliability
Test-retest reliability involves administering the same measure to the same people on two occasions and comparing the results.
Similar results suggest consistency over time.
When planning test-retest reliability, the researcher should consider:
whether the measured characteristic is expected to remain stable;
whether the interval is long enough to reduce memory of previous answers;
whether the interval is short enough to avoid major genuine change.
Reliability in observations
Observational reliability may be improved through:
clear behavioural categories;
observer training;
an agreed recording system;
practice observations;
inter-observer reliability checks.
Inter-observer reliability
Inter-observer reliability concerns the extent to which observers record behaviour consistently.
Two observers may independently record the same behaviour.
Their records are then compared.
A high level of agreement suggests that the categories and recording process are reliable.
Reliability in content analysis
Content analysis may be made more reliable by:
defining coding categories precisely;
training coders;
coding the same material independently;
comparing coders’ classifications;
revising unclear categories.
Designing for validity
What is validity?
Validity concerns whether the investigation or measure captures what it is intended to investigate.
A study may produce consistent results but still fail to measure the intended psychological construct.
Reliability alone does not guarantee validity.
Validity is covered fully in validity.
Face validity
Face validity concerns whether a measure appears to assess what it is intended to assess.
For example, a questionnaire intended to measure examination stress should contain items that appear relevant to examination stress.
A researcher may ask experts or members of the target population to inspect the measure.
Concurrent validity
Concurrent validity can be assessed by comparing the new measure with an established measure of the same construct.
If the results are similar, this supports the validity of the new measure.
Ecological validity
Ecological validity concerns whether findings can be applied to everyday settings or behaviour.
It may be improved by:
using realistic tasks;
studying behaviour in an appropriate natural setting;
avoiding unnecessary artificiality.
However, increasing naturalism may reduce control.
Temporal validity
Temporal validity concerns whether findings remain applicable across different times.
A study may have limited temporal validity if:
behaviour is strongly influenced by a particular historical period;
technology or social conditions change;
the sample reflects attitudes specific to one time.
Operational validity
A key design question is whether the operational measure represents the intended construct.
For example:
Concentration is measured as the number of visual targets found in two minutes.
The researcher should consider whether this task captures concentration or is also strongly affected by:
visual ability;
task familiarity;
reading speed;
motivation.
Improving validity
Validity may be improved through:
careful operationalisation;
realistic tasks;
appropriate comparison measures;
control of relevant extraneous variables;
suitable sampling;
piloting materials;
avoiding leading questions;
clear behavioural categories;
selecting a method suited to the research question.
Reliability and validity compared
Reliability | Validity |
Concerns consistency | Concerns accuracy or appropriateness |
Would the method produce similar results again? | Does the method measure what it claims to measure? |
Improved through standardisation | Improved through suitable operationalisation and design |
Includes test-retest and inter-observer reliability | Includes face, concurrent, ecological and temporal validity |
A measure can be reliable but invalid | A useful measure normally needs adequate reliability |
Using a pilot study
A pilot study is a small-scale trial of the planned investigation.
It is completed before the main study.
A pilot may reveal:
unclear instructions;
ambiguous questions;
unsuitable behavioural categories;
tasks that are too difficult or too easy;
procedures that take too long;
ineffective manipulations;
equipment problems;
unforeseen ethical issues;
practical risks;
floor or ceiling effects.
The researcher can then improve the design before collecting the main data.
Addressing ethics
Ethical research
Psychological investigations should be designed and conducted in accordance with appropriate ethical principles and the British Psychological Society’s code of ethics.
Ethical issues should be considered before recruitment begins.
The main issues are covered in ethics in psychological research.
Informed consent
Participants should normally receive enough information to make an informed decision about whether to take part.
Information may include:
the general nature of the task;
the time involved;
foreseeable discomfort;
how data will be used;
the right to withdraw;
contact details for questions.
The information should be understandable to the participant group.
Deception
Deception occurs when:
participants are misled;
information is withheld;
the true aim is not disclosed.
Deception should not be used merely because it makes the procedure easier.
Where withholding information is necessary and ethically justified, researchers should consider:
whether the study could be completed without deception;
whether the deception may cause distress;
how participants will be debriefed;
whether participants can withdraw their data afterwards.
Right to withdraw
Participants should be able to leave the study without penalty.
They should be told:
that participation is voluntary;
how to stop the procedure;
whether and how they can request removal of their data.
Researchers should avoid language or incentives that create inappropriate pressure to continue.
Protection from harm
Participants should be protected from physical and psychological harm.
The researcher should consider whether the investigation could cause:
stress;
embarrassment;
anxiety;
fatigue;
pain;
conflict;
damage to self-esteem;
disclosure of sensitive information.
Harm should not exceed what participants would ordinarily expect in everyday life.
Privacy
Privacy concerns participants’ reasonable expectations about being observed or having information collected.
An observation in a setting accessible to the public does not automatically remove all privacy considerations.
The researcher should consider:
the location;
the sensitivity of the behaviour;
whether individuals could expect not to be recorded;
whether consent is required.
Confidentiality
Participants’ information should be handled securely.
Possible safeguards include:
using participant numbers rather than names;
removing identifying details;
restricting access to data;
reporting group results;
storing data securely;
explaining how long data will be retained.
Debriefing
A debrief is provided after participation.
It may explain:
the full aim;
any deception;
why information was withheld;
how data will be used;
how to ask questions;
how to withdraw data;
where to seek support if necessary.
A debrief should be clear and appropriate for the participants.
Ethical design checklist
Before conducting a study, ask:
Have participants received appropriate information?
Is consent valid and voluntary?
Can participants withdraw easily?
Is any deception necessary and justified?
Could the procedure cause harm or distress?
Is privacy respected?
Will data remain confidential?
Is a suitable debrief prepared?
Are extra safeguards needed for the participant group?
Managing risk
What is a research risk?
A risk is the possibility that the investigation may cause harm, injury, distress, loss or another unwanted outcome.
Risk assessment should protect:
participants;
researchers;
members of the public;
property and equipment;
confidential information.
Risk and ethical harm
Risk assessment and ethical review overlap, but they are not identical.
Ethical review asks whether the treatment of participants is acceptable.
Risk assessment asks what could go wrong and how its likelihood or impact can be reduced.
Basic risk-assessment process
A practical risk assessment may involve:
Identifying potential hazards.
Identifying who could be affected.
Considering the likelihood and seriousness of harm.
Introducing control measures.
Planning what to do if a problem occurs.
Reviewing the controls after piloting.
Physical risks
Possible physical risks include:
trips caused by equipment or cables;
unsuitable lighting;
prolonged screen use;
fatigue;
loud sounds;
unsafe locations;
tasks involving movement.
Possible controls include:
checking the room;
securing equipment;
limiting task duration;
allowing breaks;
keeping sound at a safe and consistent level;
stopping the task if a participant reports discomfort.
Psychological risks
Possible psychological risks include:
stress;
embarrassment;
upsetting memories;
frustration;
perceived failure;
uncomfortable personal questions.
Possible controls include:
avoiding unnecessarily intrusive questions;
warning participants about sensitive content;
allowing questions to be skipped;
reminding participants of withdrawal rights;
stopping the study if significant distress occurs;
providing an appropriate debrief.
Researcher risks
Researchers may also face risks, particularly in:
field settings;
lone working;
participant observation;
research involving sensitive topics;
unfamiliar environments.
Controls may include:
working with another researcher;
using an agreed location;
informing a supervisor of the schedule;
avoiding unsafe environments;
establishing clear boundaries with participants.
Data risks
Research data may create risks if personal information is:
lost;
shared accidentally;
stored insecurely;
linked to identifiable participants.
Controls include:
anonymising data;
limiting access;
using secure storage;
separating consent details from research data;
avoiding unnecessary collection of identifying information.
Stop criteria
A study plan should identify circumstances in which the procedure will be stopped.
Examples include:
a participant asks to stop;
the participant shows significant distress;
equipment becomes unsafe;
the task no longer follows the approved procedure;
an unexpected risk emerges.
Creating a standardised procedure
What should a procedure include?
A clear procedure should allow another researcher to replicate the investigation.
It should explain:
where the research takes place;
how participants are recruited;
what participants are told;
how consent is obtained;
how participants are allocated;
what materials are used;
what participants do;
how long each stage lasts;
how variables are controlled;
how responses are recorded;
how participants are debriefed.
Writing participant instructions
Instructions should:
be clear;
use accessible language;
avoid revealing the hypothesis unnecessarily;
explain the task;
explain how to respond;
remind participants of withdrawal rights;
be identical across relevant participants.
Recording results
The design should specify:
what data will be recorded;
who will record it;
the unit of measurement;
how responses will be scored;
how missing responses will be handled;
whether data are quantitative or qualitative;
how participant identities will be protected.
Planning data analysis
Before the study begins, researchers should consider:
the likely type of data;
the level of measurement;
suitable descriptive statistics;
an appropriate graph or table;
whether an inferential test will be needed.
For example, an experiment producing related ordinal ratings may require:
a median;
a measure of dispersion;
a suitable graph;
a Wilcoxon test.
The analysis must match the original design.
Complete investigation plan
A clear plan may use the following headings.
Title
A concise description of the investigation.
Aim
A clear statement of what will be investigated.
Hypotheses
Include:
directional or non-directional hypothesis;
null hypothesis.
Research method
State and justify:
experiment;
observation;
questionnaire;
interview;
correlation;
content analysis;
case study.
Type of experiment or observation
Where relevant, state:
laboratory, field, natural or quasi-experiment;
naturalistic or controlled observation;
overt or covert observation;
participant or non-participant observation.
Target population
Define the population to which the researcher hopes to generalise.
Sample
State:
sample size;
relevant participant characteristics;
inclusion criteria where necessary.
Sampling method
State and justify:
random;
systematic;
stratified;
opportunity;
volunteer.
Experimental design
Where relevant, state and justify:
repeated measures;
independent groups;
matched pairs.
Variables
Identify and operationalise:
IV and DV;
co-variables;
extraneous variables;
behavioural categories.
Materials
List the materials needed, such as:
instructions;
word lists;
questionnaires;
recording sheets;
timers;
standardised stimuli;
consent and debrief documents.
Procedure
Describe the sequence in enough detail for replication.
Controls
Explain the use of:
random allocation;
counterbalancing;
randomisation;
standardisation;
control groups;
consistent testing conditions.
Reliability
Explain how consistency will be promoted and assessed.
Validity
Explain how relevant forms of validity will be considered.
Ethics
Identify ethical issues and explain how they will be addressed.
Risk
Identify potential risks and control measures.
Data recording and analysis
Explain:
what data will be produced;
how results will be recorded;
which descriptive display may be suitable;
which inferential test may be appropriate.
Complete worked example
Research question
Does background speech affect immediate word recall?
Aim
To investigate whether the presence of background speech affects the number of words recalled by sixth-form students.
Directional hypothesis
Sixth-form students will recall significantly fewer words from a list of \(20\) words when a recorded conversation is played than when the task is completed in silence.
Null hypothesis
There will be no significant difference in the number of words recalled by sixth-form students in the silent and background-speech conditions. Any difference will be due to chance.
Research method
A laboratory experiment would be suitable because:
the researcher will manipulate background sound;
the number of words recalled will be measured;
sound volume, exposure time and materials can be controlled.
Sample
The target population is sixth-form students.
An opportunity sample of \(30\) students could be recruited from a sixth-form college.
This would be practical, although it may limit generalisation because students from one college may not represent all sixth-form students.
Experimental design
A repeated measures design could be used so that every participant completes:
the silent condition;
the background-speech condition.
This controls individual differences in memory ability.
Counterbalancing should be used to reduce order effects.
Independent variable
The IV is background-sound condition, operationalised as:
silence;
a recording of two people speaking played at a standard volume.
Dependent variable
The DV is immediate memory performance, operationalised as:
The number of words correctly recalled from a list of \(20\) words within two minutes.
Materials
The researcher would need:
two matched word lists;
the speech recording;
headphones or standard audio equipment;
written instructions;
a timer;
response sheets;
consent information;
a debrief sheet.
Standardised procedure
Each participant would:
Read the information sheet.
Provide informed consent.
Receive standardised written instructions.
Study one word list for two minutes.
Complete a brief standardised distractor task.
Recall as many words as possible within two minutes.
Complete the second condition after an appropriate interval.
Receive a debrief.
Controls
The researcher should control:
word-list difficulty;
exposure time;
recall time;
room conditions;
recording volume;
instructions;
equipment;
scoring.
The order of conditions should be counterbalanced.
The order of the word lists could be matched to the sound order so one list is not always used in the same condition.
Reliability
Reliability could be improved using:
standardised instructions;
identical timing;
objective scoring;
consistent equipment;
a detailed procedure;
a pilot study.
Validity
The researcher should ensure that:
the lists are of comparable difficulty;
the memory task measures immediate recall;
the background speech is clearly different from silence;
uncontrolled sound does not affect participants.
The researcher should also consider whether recalling isolated words represents memory use outside the research setting.
Ethics
Participants should:
provide informed consent;
be told that they may withdraw;
be protected from excessive sound or distress;
have their scores recorded confidentially;
receive a full debrief.
Risk management
The researcher should:
keep sound at a safe volume;
inspect cables and equipment;
allow participants to stop if they experience discomfort;
keep personal data secure;
ensure that participation does not interfere unreasonably with lessons.
Planned data
The investigation will produce:
two numerical recall scores per participant;
related interval data.
An appropriate inferential test would therefore be a related t-test.
This conclusion depends on the data meeting the required features of the test.
Worked observation design
Research question
How frequently do students display cooperative behaviour during a group task?
Aim
To investigate the frequency of cooperative behaviours shown by students during a ten-minute group problem-solving activity.
Method
A controlled, overt, non-participant observation could be used.
Behavioural categories
Possible categories include:
offers an object to another participant;
provides task information without being asked;
asks another participant for their opinion;
praises another participant’s contribution.
Sampling behaviour
Event sampling could be used to record each occurrence of the categories.
Reliability
Two trained observers could record the same session independently.
Their records could be compared to assess inter-observer reliability.
Validity
The group task should provide a realistic opportunity for cooperation without directing participants to perform the target behaviours.
Ethics and risk
Participants should:
know they are being observed;
provide consent;
understand how recordings or notes will be used;
have identifying information removed.
The task should avoid unnecessary conflict or embarrassment.
Worked questionnaire design
Research question
What are students’ attitudes towards independent revision?
Aim
To investigate sixth-form students’ attitudes towards independent revision.
Method
A questionnaire could be used because it allows several students to report their attitudes efficiently.
Question design
The questionnaire could include:
closed rating questions for quantitative data;
one or two open questions for more detailed responses.
Example closed question:
How useful do you find independent revision?
Possible responses:
very useful;
useful;
neither useful nor unhelpful;
unhelpful;
very unhelpful.
Reliability
Reliability could be improved by:
using clear instructions;
avoiding ambiguous wording;
keeping the question order consistent;
piloting the questionnaire;
using test-retest reliability where appropriate.
Validity
Face validity could be checked by asking suitable people whether the questions appear to assess attitudes towards independent revision.
Concurrent validity could be considered if an established measure of the same construct were available.
Ethics
Participants should:
receive information about the study;
consent voluntarily;
be able to skip questions;
have responses kept confidential;
be debriefed where appropriate.
Evaluating whether a design is coherent
A strong investigation plan should be internally consistent.
For example:
the hypothesis should match the aim;
the method should generate data capable of testing the hypothesis;
the design should fit the method;
the variables should match the planned measures;
the sample should belong to the target population;
the control procedures should address realistic extraneous variables;
the reliability procedures should fit the method;
the validity discussion should concern the actual measure;
the ethics and risks should be specific to the proposed procedure;
the statistical test should match the final data.
Example of an incoherent plan
A researcher states:
aim: investigate a correlation between sleep and memory;
method: manipulate sleep duration;
design: compare two groups;
hypothesis: participants sleeping longer will have higher memory scores.
This is not a correlational design because sleep has been manipulated and conditions are being compared.
It is an experiment testing a difference.
The aim and hypothesis should be rewritten to match the procedure, or the method should be changed so both variables are measured rather than manipulated.
Writing a strong exam answer
Use the scenario
Do not write a generic research-methods description.
Apply every decision to the investigation.
Weak answer:
Use a repeated measures design because it controls participant variables.
Stronger answer:
Use a repeated measures design so each student completes both the silent and background-speech conditions. This controls individual differences in memory ability because each participant is compared with themselves.
Give procedural detail
Weak answer:
Standardise the procedure.
Stronger answer:
Give all participants the same written instructions, display each word list for two minutes and allow two minutes for recall in both sound conditions.
Explain controls
Weak answer:
Control the room.
Stronger answer:
Test all participants in the same quiet room so that uncontrolled background noise does not affect the number of words recalled.
Make ethical solutions practical
Weak answer:
Protect participants from harm.
Stronger answer:
Play the recording at a safe volume and tell participants that they may stop the task immediately if they experience discomfort.
Operationalise every variable
Weak answer:
Measure stress.
Stronger answer:
Measure stress using each participant’s score on a standardised questionnaire completed immediately before the task.
Key Words 🔑
Key word | Student-friendly definition | How it may be used in an exam |
Aim | A statement describing the purpose of an investigation. | You may write an aim from a research scenario. |
Hypothesis | A testable prediction about the outcome of a study. | You may write a directional, non-directional or null hypothesis. |
Directional hypothesis | A prediction stating the expected direction of a difference or relationship. | It should specify which condition will produce more or less, or the direction of a correlation. |
Non-directional hypothesis | A prediction that a difference or relationship will occur without stating its direction. | It may be used where no direction is justified. |
Null hypothesis | A prediction that there will be no genuine difference, correlation or association. | It should state that any apparent result is due to chance. |
Research method | The overall technique used to collect evidence. | Examples include experiment, observation, questionnaire and correlation. |
Target population | The wider group to which the researcher hopes to generalise. | It should be defined clearly when planning a sample. |
Sample | The participants who actually take part. | You may describe its size and relevant characteristics. |
Sampling method | The procedure used to recruit or select participants. | Examples include random, systematic, stratified, opportunity and volunteer sampling. |
Experimental design | The way participants are allocated to experimental conditions. | Select repeated measures, independent groups or matched pairs. |
Independent variable | The variable manipulated or used to create experimental conditions. | It should be operationalised precisely. |
Dependent variable | The measured outcome of an experiment. | It should be observable and replicable. |
Co-variable | One of two measured variables in a correlation. | Both co-variables must be operationalised. |
Extraneous variable | A factor other than the IV that could affect the DV. | Relevant extraneous variables should be controlled. |
Operationalisation | Defining a variable so it can be measured or manipulated. | Operational definitions should appear in hypotheses and procedures. |
Standardisation | Keeping relevant features of the procedure consistent. | It supports reliability and replicability. |
Random allocation | Assigning participants to conditions using a random process. | It may reduce systematic allocation bias. |
Counterbalancing | Varying the order of conditions across participants. | It helps control order effects in repeated measures designs. |
Demand characteristics | Cues that allow participants to guess the aim and alter their behaviour. | The design should reduce unnecessary clues. |
Investigator effects | The influence of a researcher’s behaviour or expectations on the study. | Standardisation and objective recording can reduce these effects. |
Reliability | The consistency of a measure or procedure. | You may suggest test-retest or inter-observer reliability. |
Validity | The extent to which a measure or investigation captures what it intends to study. | You may discuss face, concurrent, ecological or temporal validity. |
Pilot study | A small-scale trial completed before the main investigation. | It can identify problems with materials, procedures or ethics. |
Informed consent | Agreement to participate based on appropriate information. | The consent process should be included in the design. |
Right to withdraw | The participant’s right to stop participation or request removal of data. | Participants should know how to exercise this right. |
Confidentiality | Protecting participants’ personal information and data. | Data may be anonymised and stored securely. |
Debrief | Information provided after participation explaining the study. | It is especially important where information was withheld. |
Risk assessment | A process for identifying hazards and reducing possible harm. | It should cover participants, researchers, data and equipment. |
Replicability | The ability of another researcher to repeat the procedure. | A detailed standardised method supports replication. |
Hints from the Examiner Reports 💡
No lesson-specific examiner guidance was identified in the provided reports.
Common Mistakes ⚠️
Mistake: Writing an aim that predicts the outcome.
Why this is incorrect:
The aim describes the purpose of the study. The hypothesis makes the prediction.
How to improve:
Begin the aim with:
To investigate whether...
Mistake: Writing a hypothesis without operationalised variables.
Why this is incorrect:
Terms such as memory, stress or aggression may be measured in many different ways.
How to improve:
State exactly what participants will do and how the outcome will be recorded.
Mistake: Writing a directional hypothesis without stating a direction.
Why this is incorrect:
A directional hypothesis must predict which condition will produce the higher or lower result, or whether a correlation will be positive or negative.
How to improve:
Use wording such as:
higher than;
lower than;
more than;
fewer than;
positive correlation;
negative correlation.
Mistake: Confusing the research method with the experimental design.
Why this is incorrect:
A laboratory experiment is a type of experiment. Repeated measures is a way of allocating participants to conditions.
How to improve:
State each decision separately.
Mistake: Confusing random sampling with random allocation.
Why this is incorrect:
Random sampling selects people from the population. Random allocation places sampled participants into conditions.
How to improve:
Identify whether the decision concerns recruitment or condition assignment.
Mistake: Selecting a method because it is convenient rather than suitable.
Why this is incorrect:
The method must produce data capable of testing the hypothesis.
How to improve:
Explain how the method addresses the research question.
Mistake: Describing the sample only as “students”.
Why this is incorrect:
This does not define the target population or sample precisely.
How to improve:
Include relevant information such as age group, educational stage or setting.
Mistake: Choosing repeated measures without addressing order effects.
Why this is incorrect:
Completing one condition may influence performance in the next.
How to improve:
Explain how counterbalancing will be used.
Mistake: Choosing independent groups without addressing participant variables.
Why this is incorrect:
Pre-existing differences between groups may influence the dependent variable.
How to improve:
Use random allocation and keep recruitment criteria consistent.
Mistake: Treating matched pairs as unrelated data.
Why this is incorrect:
Each participant is deliberately paired with another participant, so the resulting scores are related.
How to improve:
Identify the matching variable and explain how pairs will be formed.
Mistake: Naming an extraneous variable without explaining how it will be controlled.
Why this is incorrect:
Identifying a possible problem does not show how the design will address it.
How to improve:
State a specific control linked to the variable.
Mistake: Claiming that every variable will be controlled.
Why this is incorrect:
It is rarely possible to eliminate every influence on behaviour.
How to improve:
Identify the most relevant variables and explain realistic controls.
Mistake: Using vague behavioural categories.
Why this is incorrect:
Different observers may interpret terms such as friendly, aggressive or helpful differently.
How to improve:
Define observable actions that can be recorded consistently.
Mistake: Stating that a method is reliable without explaining why.
Why this is incorrect:
Reliability must be supported through design features or checks.
How to improve:
Refer to standardisation, test-retest reliability or inter-observer reliability as appropriate.
Mistake: Treating reliability and validity as the same concept.
Why this is incorrect:
Reliability concerns consistency. Validity concerns whether the intended construct is measured appropriately.
How to improve:
Address each separately in the investigation plan.
Mistake: Giving generic ethical statements.
Why this is incorrect:
An answer such as “follow ethical guidelines” does not address the actual risks of the study.
How to improve:
Identify a specific issue and a practical way to deal with it.
Mistake: Assuming informed consent resolves every ethical issue.
Why this is incorrect:
Researchers must also consider withdrawal, harm, privacy, confidentiality, deception and debriefing.
How to improve:
Review the complete procedure for several possible ethical concerns.
Mistake: Ignoring risks to researchers or data.
Why this is incorrect:
Risk assessment is not limited to physical harm experienced by participants.
How to improve:
Consider participants, researchers, equipment, locations and confidential information.
Mistake: Describing a procedure that could not be replicated.
Why this is incorrect:
Statements such as “give participants a memory test” omit essential detail.
How to improve:
State the materials, timing, instructions, scoring and sequence.
Mistake: Choosing an inferential test before identifying the final data.
Why this is incorrect:
Test selection depends on the research purpose, design and level of measurement.
How to improve:
Work out what data the procedure will produce before selecting the test.
Exam-Style Questions ✍️
Question 1
A psychologist wants to investigate whether background speech affects immediate memory.
Write an appropriate research aim.[2 marks]
Question 2
Using the investigation in Question 1, write:
a) an operationalised directional hypothesis;[3 marks]
b) an operationalised null hypothesis.[3 marks]
Question 3
A researcher wants to investigate whether students’ examination-stress ratings are related to the number of hours they sleep.
a) Identify an appropriate research method.[1 mark]
b) Explain why this method is appropriate.[2 marks]
c) Identify and operationalise the two co-variables.[4 marks]
Question 4
A psychologist plans to compare memory performance before and after a revision workshop. The same participants will complete both tests.
a) Identify the experimental design.[1 mark]
b) Explain one strength of using this design.[2 marks]
c) Explain one problem that could arise and how it could be controlled.[3 marks]
Question 5
A researcher observes cooperative behaviour during a group task.
Design three suitable behavioural categories.[3 marks]
Explain how the researcher could assess the reliability of the observational data.[3 marks]
Question 6
A psychologist wants the proportions of Year 12 and Year 13 students in a sample to reflect their proportions in the college population.
a) Identify an appropriate sampling method.[1 mark]
b) Explain how the sample should be selected.[3 marks]
c) Explain one reason why this method may produce a more representative sample than opportunity sampling.[2 marks]
Question 7
A laboratory experiment investigates whether background noise affects concentration.
Suggest:
a) one operationalised independent variable;[2 marks]
b) one operationalised dependent variable;[2 marks]
c) two relevant extraneous variables;[2 marks]
d) one control for each extraneous variable.[2 marks]
Question 8
A researcher develops a questionnaire to measure examination anxiety.
Explain how the researcher could address:
a) reliability;[3 marks]
b) face validity;[2 marks]
c) concurrent validity.[2 marks]
Question 9
A psychologist asks students to complete a stressful timed task while being recorded.
Explain four ethical or risk issues that should be considered when designing the investigation. For each issue, suggest an appropriate way of dealing with it.[8 marks]
Question 10
Design an investigation into whether working with background music affects students’ performance on a concentration task.
Your answer should include:
an appropriate aim;
an operationalised hypothesis;
the research method;
the sample and sampling method;
the experimental design;
the operationalised IV and DV;
relevant control procedures;
reliability and validity;
ethics and risk;
how the results would be recorded.
[16 marks]
Comments