9.9. Summary

Key experimental principles in psychology include manipulating independent variables, controlling extraneous variables, and using random assignment to ensure reliability and rule out alternative explanations. Experimental designs, such as single-factor two-level and multi-level setups, often include control conditions like placebos or wait-list groups to isolate treatment effects. The placebo effect demonstrates how expectations alone can influence outcomes, as shown in studies like placebo knee surgeries for osteoarthritis. Between-subjects experiments expose participants to only one condition, requiring random assignment to balance groups, while within-subjects designs test all conditions on the same participants, controlling variability but introducing order effects that can be mitigated through counterbalancing. Researchers must weigh these designs’ benefits and limitations, selecting the most suitable approach based on the research question and practical considerations.

Validity is a cornerstone of psychological research, encompassing internal validity (establishing cause-and-effect relationships), external validity (generalisability), construct validity (accurately measuring intended concepts), and statistical validity (appropriate use of data analysis). While experiments excel in internal validity due to controlled conditions, external validity may be limited unless field studies are conducted. Construct validity relies on the alignment between experimental manipulations and research questions, while statistical validity requires proper sample sizes and analysis methods. Correlational research complements experiments by exploring statistical relationships between variables without manipulation, enabling description and prediction when experimental designs are impractical or unethical. However, its inability to establish causation due to directionality and third-variable problems underscores the importance of integrating experimental methods to reinforce theoretical insights.

Qualitative research offers a contrasting approach, focusing on understanding human experiences through rich, non-numerical data collection methods like interviews and thematic analysis. This approach generates deep insights, often identifying patterns and hypotheses for future quantitative exploration. For instance, Lindqvist et al. used unstructured interviews with families of teenage suicide victims to uncover the enduring emotional impact of unanswered questions. Qualitative research captures lived experiences, while mixed-methods designs bridge the depth of qualitative insights with the statistical rigour of quantitative analysis, providing a comprehensive understanding of human behaviour.

Survey research exemplifies flexibility in psychological studies, relying on self-reports and often large random samples to describe variables or assess relationships. With roots in early 20th-century social surveys, innovations like the Likert scale have expanded its applications to topics like mental health prevalence and public attitudes. Surveys can be experimental or non-experimental, requiring careful design to minimise biases and errors. Features like randomisation and clear question ordering enhance reliability, while qualitative and quantitative analyses enable a balance between depth and breadth. Mixed-methods approaches often combine these strengths to capture complex phenomena effectively.

Quasi-experimental research, which lacks random assignment or counterbalancing, addresses directionality problems by manipulating independent variables but remains vulnerable to confounding variables. Common designs like one-group posttest-only and pretest-posttest setups offer preliminary insights but face threats to internal validity, such as history and maturation effects. Interrupted time series designs strengthen causal interpretations by using repeated measurements before and after interventions, distinguishing treatment effects from random fluctuations. These designs are valuable in real-world settings where true experiments are impractical but require cautious interpretation to avoid overstating causal claims.

Factorial designs allow researchers to explore complex interactions between multiple independent variables, combining their levels into unique conditions. For example, Schnall et al. demonstrated how disgust (manipulated by room cleanliness) and private body consciousness interacted to influence moral judgements. These designs, such as 2 × 2 or more complex setups, investigate main effects and interactions, with simple effects analyses offering nuanced insights into how variables interact. While factorial designs enhance understanding of multifactorial influences, they require careful planning to manage complexity and control confounding variables.