- Social: Practical issues when researching prejudice, such as social desirability or demand characteristics. Issues with interviews/self-reports when measuring obedience/prejudice. Importance of sampling ppts into conditions. Sherif – naturalistic observation so high in ecological validity. Stratified sampling, variables (ie situational) needed controlling and data was subjective.
- Cognitive: Relies on labs experiments a lot so raises issues around ecological validity, mundane realism in tasks, control and operationalisation of variables. Also, how can we actually measure memory? Baddeley lacked mundane realism and ecological validity but had high reliability and controls.
- Biological: Issues around cost and equipment when scanning and measuring complexity of brain. Cause and effect when trying to infer conclusions from studying a complex organ. Difficulty in finding samples of MZ and DZ twins (issue of generalisability) and whether samples are valid. Brain scans are objective. Raine – PET scans, expensive. Lab, not ecologically valid. Objective.
- Learning: Generalising from animals to humans is questionable, problems with using overt and structured observations. Watson and Rayner – lack of ecological validity, high control so good reliability.
- Clinical: Focuses more on qualitative data from case studies/interviews. Hard to analyse – conclusions may be unreliable and subjective. Rosenhan – qualitative data, subjective. Ecologically valid but not as reliable.
- Child: Observations – risk of observer effects, demand characteristics, observer bias/subjectivity. Meta-analysis – often inconsistent, no primary data. Different methods, hard to compare. Van Ijzendoorn and Kroonenberg – biased sample, good generalisability.
