Designing psychological research — Edexcel GCSE Psychology

In psychological research, variables must be clearly identified. The independent variable (IV) is the factor manipulated by the researcher to observe its effect (e.g., type of music played). The dependent variable (DV) is the factor measured to assess the outcome of the IV manipulation (e.g., number of words recalled). Extraneous variables are other factors that could affect the DV if not controlled. These include situational variables, which are environmental factors like noise or lighting, and participant variables, which are individual differences among subjects, such as age or intelligence. Controlling these extraneous variables ensures that changes in the DV are solely caused by the IV, establishing valid cause and effect.

Extraneous variables are factors other than the independent variable that could affect the dependent variable. Researchers use specific methods to control them. Standardised procedures ensure participants experience identical environments and instructions, preventing situational variables from skewing data. Counterbalancing (half do condition A then B, half do B then A) controls order effects in repeated measures designs. Randomisation reduces bias by using chance to allocate participants. Single-blind techniques, where participants do not know their condition, reduce demand characteristics. Double-blind techniques, where neither participants nor the researcher know the conditions, prevent both demand characteristics and researcher bias.

A null hypothesis predicts there will be no significant difference or relationship between the variables studied, and any observed changes are due to chance. Writing a clear null hypothesis requires explicitly stating both the independent variable (IV) and the dependent variable (DV). For an experiment, it states there will be no difference in the DV between the conditions of the IV. For example, 'There will be no significant difference in the number of words recalled between participants who listen to music and those who study in silence'. For a correlational study, it states there will be no significant relationship between the two co-variables. Variables must be operationalised clearly so the hypothesis is testable and precise.

An alternative hypothesis is a precise, testable statement predicting the outcome of a psychological study. Unlike a null hypothesis, which predicts no effect, the alternative hypothesis states that there will be a significant difference or relationship between variables. To write one effectively, you must clearly operationalise both the independent variable (IV) and the dependent variable (DV). For example, instead of saying caffeine affects memory, a strong alternative hypothesis would state: Participants who consume 50mg of caffeine will recall significantly more words from a list of 20 than participants who consume no caffeine. This clearly identifies the two conditions of the IV and how the DV is measured, making the prediction fully testable.

A target population is the specific group a researcher wants to study. Since testing everyone is rarely possible, researchers select a sample. Random sampling gives every member an equal chance of selection, reducing bias but requiring a full list of the population. Stratified sampling divides the population into subgroups and selects participants proportionally, ensuring high representativeness, though it is time-consuming. Volunteer sampling relies on participants putting themselves forward; it is ethical and easy, but often suffers from volunteer bias. Opportunity sampling selects anyone available at the time, making it quick and convenient, but highly unrepresentative of the wider target population.

Experimental design dictates how participants are allocated to conditions. In an independent measures design, different participants are used in each condition (e.g., Group A drinks caffeine, Group B drinks water). This avoids order effects but introduces participant variables. A repeated measures design uses the same participants in all conditions. This eliminates participant variables and requires fewer people, but order effects like fatigue or practice can occur. A matched pairs design pairs participants based on key variables (e.g., memory ability) and assigns one from each pair to each condition. This reduces participant variables and avoids order effects, but it is time-consuming and exact matching is impossible.

Psychologists evaluate reliability (consistency) and validity (accuracy) when planning research. Sampling methods affect population validity; random sampling is highly valid as it represents the target population, whereas opportunity sampling may lack validity due to bias. Experimental designs impact internal validity; repeated measures can reduce validity through order effects, but increase it by controlling participant variables. Quantitative methods (e.g., closed questionnaires) offer high reliability because they are easily replicated, but may lack ecological validity. Qualitative methods (e.g., unstructured interviews) provide rich data, increasing validity, but their subjective nature and lack of standardisation often reduce reliability.

Ethical guidelines protect participants in psychological research. Informed consent requires researchers to explain the study's aims and procedures so participants can agree to take part. If deception is necessary to prevent demand characteristics, researchers must deal with this by providing a full debrief afterwards. Confidentiality ensures participants' data is kept anonymous and secure. The right to withdraw means participants can leave the study at any time and have their data destroyed. Protection of participants ensures they are not exposed to physical or psychological harm greater than in everyday life. Researchers deal with these issues by using consent forms, debriefing, anonymising data, and conducting risk assessments.

Psychologists use various research methods, each with distinct features, strengths, and weaknesses. Key methods include experiments, self-reports, observations, case studies, and correlational research. Features include the setting, variable control, and data type. Strengths often involve high internal or ecological validity, while weaknesses include demand characteristics, researcher bias, or ethical issues. Understanding these allows researchers to select the most suitable method. For example, a laboratory experiment is suitable for establishing cause and effect due to high control, whereas correlational research is ideal for identifying relationships between two co-variables without manipulating them.

A laboratory experiment is conducted in a highly controlled, artificial environment where the researcher manipulates an independent variable to measure its effect on a dependent variable. This strict control minimises extraneous variables, allowing researchers to establish a clear cause-and-effect relationship. For example, testing memory recall in a quiet room with timed word lists ensures background noise does not affect results. While laboratory experiments offer high internal validity and are easily replicable due to standardised procedures, they often lack ecological validity. Participants may also alter their behaviour due to demand characteristics, knowing they are being observed in an unnatural setting.

A field experiment is conducted in a natural, everyday environment where the researcher deliberately manipulates the independent variable to measure its effect on the dependent variable. Because the setting is natural, such as a school or a high street, participants are often unaware they are in a study. This significantly reduces demand characteristics and increases ecological validity. For example, a researcher might drop a wallet in a shopping centre to test if the dropper's gender affects whether bystanders help. However, field experiments offer less control over extraneous variables, making it harder to establish a definitive cause-and-effect relationship and more difficult to replicate the study exactly.

A natural experiment is a research method where the independent variable (IV) is not manipulated by the researcher but changes naturally in the environment. The researcher simply measures the effect of this naturally occurring IV on a dependent variable (DV). Because the researcher does not control the IV or randomly allocate participants to conditions, natural experiments often have lower internal validity but higher ecological validity than laboratory experiments. For example, a researcher might study the effect of a sudden school closure (the naturally occurring IV) on students' stress levels (the DV). This method is particularly useful for investigating situations where it would be unethical or practically impossible to manipulate the IV directly, such as studying the psychological impact of a natural disaster.

An interview is a self-report method where a researcher asks participants questions directly. A structured interview uses a predetermined set of questions in a fixed order, making it easy to replicate but lacking flexibility. An unstructured interview resembles a guided conversation with no set questions, allowing the researcher to explore responses in depth, though it is harder to analyse. A semi-structured interview combines both approaches, using a list of set questions but allowing the interviewer to ask follow-up questions. For example, a psychologist might use a semi-structured interview to cover specific topics while allowing the participant to elaborate on their experiences, balancing reliability with rich qualitative data.

A questionnaire is a self-report method where participants respond to written questions. It includes closed-ended questions, which restrict the respondent to fixed answers like 'yes/no'. These elicit quantitative data, which is numerical and easy to analyse statistically, for example, calculating the percentage of people who sleep eight hours a night. Questionnaires also include open-ended questions, allowing participants to answer in their own words without restricted options. These elicit qualitative data, providing rich, non-numerical insights, such as asking participants to describe how they feel after a poor night's sleep. Researchers often use a combination of both question types to gather comprehensive data.

A correlation is a research method used to investigate the relationship between two co-variables, rather than looking for a cause-and-effect relationship. Researchers measure both variables to see if they change together. A positive correlation occurs when both variables increase together, such as hours spent revising and test scores. A negative correlation occurs when one variable increases while the other decreases, for example, school absences and exam performance. Zero correlation means there is no relationship between the variables. Correlations are plotted on scatter diagrams to visualise the strength and direction of the relationship, but crucially, they cannot prove that one variable causes the change in the other.

A case study is an in-depth investigation of a single individual, small group, or specific event. It uses a variety of research methods, such as interviews, observations, and psychological tests, to gather rich qualitative and quantitative data. Case studies are useful for investigating rare phenomena that cannot be studied experimentally due to ethical constraints, such as severe brain damage. While they provide deep insights and high ecological validity, their findings are difficult to generalise to the wider population because the sample is unique. Furthermore, researchers may become subjective, leading to researcher bias, and replicating a case study is usually impossible, which significantly reduces the reliability of the findings.

Observation involves watching and recording participants' behaviour in a specific setting. Observations can be naturalistic (in a real-world environment) or controlled (in a structured environment like a lab). They can also be participant (the researcher joins the group) or non-participant (the researcher observes from a distance), and overt (participants know they are being watched) or covert (participants are unaware). Researchers use behavioural categories and tally charts to systematically record data. Naturalistic observations offer high ecological validity but lack control over extraneous variables. Covert observations reduce demand characteristics but raise ethical issues regarding informed consent.