Welcome to Research Methods in Psychology!
Hello and welcome to your study notes for Research methods (AQA AS Level Psychology 7181, Chapter 3.2.3). Research methods make up an entire 24-mark section (Section C) of Paper 2: Psychology in context, and research skills will also appear across Paper 1 and Paper 2. In fact, at least 10% of total marks across your AS assessment test mathematical skills in psychological contexts!
Don't worry if research methods or data analysis seem intimidating or technical at first. We will break down every single concept, formula, and study design into bite-sized, digestible steps with real-world examples, memory aids, and clear exam warnings.
1. Experimental Methods
In psychology, an experiment is a research method where a researcher manipulates an Independent Variable (IV) to measure its effect on a Dependent Variable (DV), while attempting to control all other extraneous variables. This is the only research method that allows us to establish a direct cause-and-effect relationship.
Types of Experiments
There are four types of experiments you need to know for your exam:
1. Laboratory Experiment:
• What it is: An experiment conducted in a highly controlled, artificial environment where the researcher directly manipulates the IV.
• Strength: High internal validity and excellent control over extraneous variables, meaning we can be confident that changes in the DV are caused solely by the IV. It is also easily replicated.
• Limitation: Low ecological validity (mundane realism) because the artificial setting and tasks may not reflect everyday real-world behaviour.
2. Field Experiment:
• What it is: An experiment conducted in a natural, real-world setting (like a school, street, or hospital) where the researcher directly manipulates the IV.
• Strength: Higher ecological validity and more natural behaviour from participants.
• Limitation: Less control over extraneous variables, making replication harder and ethical issues (like lack of informed consent) more likely.
3. Natural Experiment:
• What it is: The researcher takes advantage of a pre-existing, naturally occurring IV (a real-world event or environmental change that happens on its own, such as the introduction of television to a remote community). The researcher does not manipulate the IV.
• Strength: Allows psychologists to study real-life situations that would be unethical or impossible to create experimentally.
• Limitation: Participants cannot be randomly allocated to conditions, leading to potential participant variables and lower internal validity.
4. Quasi-Experiment:
• What it is: An experiment where the IV is based on an existing difference or innate characteristic between people (such as age, gender, or having a specific phobia vs not having a phobia). The IV is not manipulated because it already exists.
• Strength: Often carried out under controlled conditions, sharing the high control of lab studies.
• Limitation: Participants cannot be randomly assigned to conditions because their group membership is predetermined, meaning confounding participant variables cannot be fully ruled out.
The Role of Control Groups
A control group is a baseline group of participants who do not receive the experimental treatment or IV manipulation. By comparing the experimental group against the control group, researchers can be certain whether the IV actually caused the observed change in the DV.
Quick Review & Key Takeaway: Remember the crucial difference between natural and quasi-experiments! In natural experiments, the IV is an environmental event that happens naturally. In quasi-experiments, the IV is an existing human characteristic.
2. Observational Techniques
Observational methods involve watching and recording participants' behaviour without directly manipulating an IV.
Types of Observations
Observations are classified along three dimensions:
• Naturalistic vs Controlled: Naturalistic observation takes place in an unaltered, everyday setting where everything is left free to vary. Controlled observation takes place in a structured or laboratory environment where certain variables are regulated by the researcher.
• Covert vs Overt: In a covert observation, participants are unaware that they are being watched (e.g., observed via a two-way mirror). In an overt observation, participants know they are being observed and have given consent.
• Participant vs Non-participant: In a participant observation, the researcher joins the group being observed to gain first-hand insight. In a non-participant observation, the researcher remains outside the group as an objective onlooker.
Observational Design & Sampling
To make observations objective and systematic, researchers use specific tools:
• Behavioural Categories: Breaking down the target behaviour into clearly defined, observable, measurable, and non-overlapping components (e.g., operationalising "aggression" into kicking, pushing, and shouting).
• Event Sampling: Counting and recording every single time a specific target behaviour occurs during the observation period.
• Time Sampling: Recording target behaviour at predetermined, regular time intervals (e.g., recording what the participant is doing every 30 seconds).
Memory Trick: Think of Event sampling as counting Every event, and Time sampling as checking the Timer!
3. Self-Report Techniques
Self-report techniques ask participants to describe their own thoughts, feelings, attitudes, and behaviours.
Questionnaires
A questionnaire is a pre-set written list of questions.
• Open Questions: Do not have fixed responses; participants answer in their own words (e.g., "Explain how you feel when taking an exam."). They produce rich qualitative data but are harder to analyse statistically.
• Closed Questions: Offer fixed response options (e.g., Yes/No, multiple-choice, or rating scales). They produce quantitative data that is easy to graph and analyse, but lack descriptive depth.
• Types of Closed Scales:
- Likert scale: Participants indicate their level of agreement (e.g., from 1 = Strongly Disagree to 5 = Strongly Agree).
- Rating scale: Participants select a numeric value representing their feelings (e.g., rating anxiety from 1 to 10).
- Semantic differential scale: Participants place a mark on a continuum between two opposite adjectives (e.g., Calm ------------- Anxious).
Interviews
• Structured Interview: Involves predetermined, standardised questions delivered face-to-face or over the phone in a fixed order. Easy to replicate, but inflexible.
• Unstructured Interview: Works like a conversation with no fixed questions. The interviewer can follow up on interesting points, providing rich detail, but it is difficult to replicate and compare across interviewees.
4. Correlations
A correlation investigates the statistical relationship or association between two co-variables (e.g., hours of revision and exam score).
Crucial Distinction: Correlation vs Experiment
• In an experiment, the researcher deliberately manipulates an IV to measure the direct effect on a DV, allowing causality (cause-and-effect) to be established.
• In a correlation, there is NO IV manipulation and NO DV. We measure two co-variables as they naturally occur. A correlation demonstrates an association or link, but CANNOT prove cause and effect (because an unmeasured third variable might be causing the link!).
Examiner Warning: Never use the terms "IV" and "DV" or say "X causes Y" when discussing correlations. Always refer to co-variables and associations!
5. Scientific Processes
Aims and Hypotheses
• Aim: A general statement explaining the purpose of the research (e.g., "To investigate whether sleep deprivation affects memory recall.").
• Hypothesis: A precise, testable, operationalised statement predicting the outcome of the research.
• Null Hypothesis (\(H_0\)): Predicts no difference or no relationship (e.g., "There will be no significant difference in recall scores between sleep-deprived and non-sleep-deprived participants.").
• Directional Hypothesis (One-tailed): Predicts the specific direction of the outcome (e.g., "Participants who sleep 8 hours will recall significantly more words than participants who sleep 4 hours."). Used when previous research or theory suggests a particular direction.
• Non-directional Hypothesis (Two-tailed): Predicts a difference or relationship, but does not specify which direction it will go (e.g., "There will be a significant difference in word recall between the two groups."). Used when there is no previous research or findings are contradictory.
Sampling Techniques
Psychologists select a sample of participants from a broader target population.
• Random Sampling: Every member of the target population has an equal chance of being selected (e.g., putting all names in a hat or computer generator). Reduces bias, but is time-consuming.
• Systematic Sampling: Selecting every \(n^{\text{th}}\) person from an ordered list/sampling frame (e.g., every 5th name). Objective and straightforward.
• Stratified Sampling: Subgroups (strata) within the target population are identified, and participants are selected in exact proportion to their occurrence in the population. Highly representative, but very complex and time-consuming.
• Opportunity Sampling: Selecting anyone from the target population who is available and willing at the time of the study (e.g., approaching people in a library). Quick and convenient, but prone to severe sample bias.
• Volunteer Sampling: Participants self-select by responding to an advert or poster. Easy to recruit, but biased towards cooperative or motivated personality types.
Generalisation: The extent to which findings from a sample can be applied to the target population. If a sample suffers from sampling bias, generalisability is reduced.
Pilot Studies
A pilot study is a small-scale trial run of the entire research procedure conducted before the actual investigation.
• Purpose: To test the feasibility of the design, check that standardised instructions are clear, ensure timings are realistic, identify ambiguous questionnaire items, and modify any flawed procedures before investing time and money.
• Examiner Warning: A pilot study is not conducted to see if the hypothesis is supported or to get early results!
Experimental Designs
1. Repeated Measures Design:
• All participants take part in all conditions of the experiment.
• Strength: Participant variables are completely controlled; fewer participants needed.
• Limitation: Risk of order effects (practice, boredom, or fatigue).
• Control: Counterbalancing (the \(AB/BA\) technique), where half the participants do Condition A then B, while the other half do Condition B then A.
2. Independent Groups Design:
• Different participants are placed into each separate condition.
• Strength: No order effects; participants cannot guess the aim across conditions.
• Limitation: Differences between groups may be caused by participant variables (e.g., individual differences in ability) rather than the IV; requires double the participants.
• Control: Random allocation (e.g., drawing names out of a hat) to distribute individual differences evenly.
3. Matched Pairs Design:
• Participants are paired on key characteristics (e.g., IQ, age), and then one member of each pair is randomly allocated to Condition A and the other to Condition B.
• Strength: No order effects and participant variables are greatly reduced.
• Limitation: Matching is time-consuming, expensive, and can never match people perfectly.
Variables and Control
• Operationalisation: Defining variables in precise, measurable terms (e.g., measuring "memory" as "the number of words correctly recalled from a list of 20 words in 60 seconds").
• Extraneous Variables: Any nuisance variable, other than the IV, that could affect the DV if not controlled.
• Confounding Variables: An uncontrolled extraneous variable that has systematically changed with the IV, meaning we cannot know whether the IV or the confound caused the change in the DV.
• Demand Characteristics: Cues in an experiment that lead participants to guess the research aim, causing them to alter their natural behaviour.
• Investigator Effects: Any conscious or unconscious influence of the researcher's behaviour, appearance, or expectations on the outcome.
• Standardisation & Randomisation: Using identical, formalised procedures/instructions for all participants, and using chance (randomisation) to determine trial order or material presentation to minimise bias.
Ethical Issues and BPS Guidelines
The British Psychological Society (BPS) Code of Ethics outlines key ethical standards:
• Informed Consent: Making participants aware of the aims, procedures, and right to withdraw so they can make an informed decision. Dealt with via consent forms, presumptive consent (asking a similar group), prior general consent (agreeing to be deceived beforehand), or retrospective consent (asking for consent after the study during debriefing).
• Deception: Deliberately misleading or withholding information from participants. Dealt with through a comprehensive debriefing at the end, explaining the true aims and offering the right to withdraw data.
• Protection from Harm: Ensuring participants are not exposed to greater physical or psychological risk (stress, embarrassment) than in daily life.
• Privacy and Confidentiality: Protecting personal data. Dealt with by keeping participant identities anonymous (using numbers or false names).
6. Data Handling and Analysis
Types of Data
• Qualitative Data: Descriptive, non-numerical data expressing thoughts, feelings, and experiences (rich in detail, difficult to analyse statistically).
• Quantitative Data: Numerical data that can be counted and statistically analysed (objective and easy to compare, but lacks contextual meaning).
• Primary Data: First-hand data collected directly by the researcher specifically for the current investigation.
• Secondary Data: Pre-existing data collected by someone else (e.g., government statistics, journal articles).
• Meta-analysis: A research method that pools and statistically analyses secondary data from multiple published studies on the same topic to identify overall trends.
Descriptive Statistics: Measures of Central Tendency
• Mean: The arithmetic average (sum of all scores divided by the total number of scores).
- Strength: Most sensitive measure as it includes every score.
- Limitation: Easily distorted by extreme outlier scores.
• Median: The middle score when data is ordered from lowest to highest.
- Strength: Not affected by extreme outliers.
- Limitation: Less sensitive as it does not consider all actual values.
• Mode: The most frequently occurring score in a data set.
- Strength: Easy to calculate and the only measure suitable for categorical (nominal) data.
- Limitation: Uninformative if there are multiple modes or no repeating scores.
Descriptive Statistics: Measures of Dispersion
• Range: The difference between the highest and lowest score (often calculated as \((\text{Highest} - \text{Lowest}) + 1\)). Easy to calculate, but distorted by extreme values.
• Standard Deviation (SD): A single value representing the average distance/spread of scores around the mean. A higher SD means the data is widely spread; a low SD indicates the scores are clustered closely around the mean.
Mathematical Requirements
You must be confident with basic mathematical conversions:
• Percentages: \(\frac{\text{Part}}{\text{Whole}} \times 100\)
• Percentage Change: \(\frac{\text{New Value} - \text{Original Value}}{\text{Original Value}} \times 100\)
• Ratios: Expressing relative parts in simplest form (e.g., \(12 : 4 \implies 3 : 1\)).
• Fractions & Decimals: Converting between fractions and decimals (e.g., \(\frac{3}{4} = 0.75\)).
Data Presentation & Graphs
• Bar Charts: Used for discrete, categorical data (bars do not touch).
• Histograms: Used for continuous data (bars touch to represent intervals).
• Scattergrams: Used to display the relationship between two co-variables in correlational research.
• Line Graphs / Frequency Polygons: Used to show continuous data and trends over time.
Distributions
• Normal Distribution: Symmetrical, bell-shaped curve. The mean, median, and mode all fall at the exact same central midpoint.
• Positive Skew: Data distribution where the tail extends to the right (higher scores). Occurs when a test is very difficult and most people score low.
- Order of averages: Mode < Median < Mean
• Negative Skew: Data distribution where the tail extends to the left (lower scores). Occurs when a test is very easy and most people score high.
- Order of averages: Mean < Median < Mode
Memory Trick for Skewed Distributions: Look at where the long "tail" is pointing! If the tail points to the positive/right side, it is Positive Skew. If the tail points to the negative/left side, it is Negative Skew.
7. Inferential Testing: The Sign Test
The Sign Test is the only statistical test requiring direct calculation at AS Level. It tells psychologists whether an observed difference between two conditions is statistically significant (unlikely to have occurred by chance).
When to Use the Sign Test (The 3 Conditions)
You must use the Sign Test when all three criteria are met:
1. The study is looking for a difference (not an association/correlation).
2. The study uses a repeated measures or matched pairs design (related data).
3. The data is nominal (categorical data, or converted into signs of difference: \(+\) and \(-\)).
Step-by-Step Calculation of the Sign Test
Step 1: For each participant, subtract Condition B from Condition A. Record whether the change is positive (\(+\)), negative (\(-\)), or no change (\(0\)).
Step 2: Count the total number of \(+\) signs and the total number of \(-\) signs.
Step 3: Discard any participants who showed no difference (\(0\)). The remaining number of participants is your adjusted \(N\) value.
Step 4: Find the calculated value of \(S\). \(S\) is simply the less frequent sign (whichever count is smaller between \(+\) and \(-\)).
Step 5: Locate the critical value from the critical values table provided in the exam, using your calculated \(N\), the significance level (usually \(p \le 0.05\)), and whether the hypothesis is directional (1-tailed) or non-directional (2-tailed).
Step 6: Apply the Decision Rule:
For the Sign Test, the result is statistically significant if:
\(\text{Calculated } S \le \text{Critical Value}\)
If \(S\) is less than or equal to the critical value, we reject the null hypothesis and accept our research hypothesis!
Sign Test Example Walkthrough:
Suppose in a memory experiment (\(N = 10\)), 8 participants recalled more words with imagery (\(+\)), 1 recalled fewer (\(-\)), and 1 showed no change (\(0\)).
• Discard the zero: Adjusted \(N = 9\).
• Count of \(+\) signs = 8; Count of \(-\) signs = 1.
• Calculated \(S = 1\) (the less frequent sign).
• If the critical value table for \(N = 9\) at \(p \le 0.05\) (one-tailed) gives a critical value of \(1\):
Since calculated \(S = 1\) is \(\le 1\), the result is statistically significant!
8. Top Pitfalls & Examiner Tips
• Avoid Vague Operationalisation: Never just write "we will measure aggression" or "memory will be recorded." Always specify units: "The number of aggressive acts recorded in 15 minutes" or "The number of words recalled out of 20."
• Pilot Study Purpose: Never write that a pilot study is to "check if the hypothesis is true." It is purely a practical trial run to check instructions, timings, and materials.
• Sign Test Decision Rule: Do not mix up greater than and less than! For the Sign Test: Calculated \(S \le \text{Critical Value}\).
• Correlation Causality: Never state that a correlation "proves" or "causes" something. Always use terms like "positive association" or "negative relationship".
Key Takeaway: Master your experimental designs, sampling techniques, and the Sign Test decision rule, and you will secure top marks across Section C of Paper 2!