Research Methods

Research Methods

1. The Scientific Method

   • Empirical Methods
     • Information is gained through direct observation or experiment
   • Objectivity
     • Scientists' expectations should not affect what they record
   • Replicability
     • To affirm the validity of an experiment or observation, it is necessary to repeat it, affirming the 'truth' of the original results
   • Theory Construction
     • Theories help us understand and predict natural phenomena. A theory is a collection of general principles that explain observations and facts
   • Control Testing
     • Scientists seek to demonstrate causal relationships, so must control all conditions other than the ones they are testing to make it a fair test

2. The Scientific Process

   • Induction
     • Reasoning from the particular to the general
   • Deduction
     • Reasoning from the general to the particular

3. Psychology: A Science?

   • Lack of objectivity and control
     • Human behaviour cannot be measured objectively. The object of study reacts to the researcher, leading to experimenter bias and demand characteristics
   • Nomothetic vs Idiographic
   • The Scientific Method
     • Psychology uses the scientific method, generating falsifiable models and conducting well-controlled experiments
     • Scientific methods haven't worked. Unlike in other sciences, there is no single paradigm
   • Falsification
   • Reductionism and Determinism
     • The aim of science is to reduce phenomena to their most simple explanation. In Psychology, this isn't possible due to the complexities of human behaviour
   • Qualitative Research
     • Qualitative research can be objective and follow the scientific method
   • Nomothetic vs Idiographic
     • In Psychology, generalisations cannot be made as they can in other sciences. This may mean that the goals of science are inappropriate for Psychology

4. The Psychological Report

   • Abstract
     • A summary of all the other sections of the report, including the aim of the study
   • Introduction
     • Introduction of topic. Review of related theories and studies. Explanation of what the research wants to investigate and why. Statement of hypotheses
   • Method
     • Exact description of method to allow replication, including experimental design, participants, procedures, instructions and materials
   • Results
     • Descriptive and inferential statistics
   • Discussion
     • Application of research to hypothesis. Comparison with background studies. Acknowledgement of limitations. Implications of findings and suggestions for future research. Explanations of results and conclusions
   • References

5. Peer Review

   • Purposes
     • Helping funding bodies decide how to allocate research funding
     • Ensuring high quality of publications of research in scientific journals
     • Assessing the research rating of university departments
   • Criticisms
     • It isn't always possible to find an appropriate expert, causing some research to be badly reviewed
     • The anonymity of the reviewers may be used to bury rival research
     • Peer review favours the publication of positive results because editors want significant results in their journals
     • Peer review results in a preference for research that goes with the existing theory

6. Reliability

   • Definition
     • Consistency
   • Inter-rater Reliability
     • The extent to which observers, experimenters or interviewers agree, or produce the same outcome
   • External Reliability
     • A measure of consistency over several different occasions
   • Internal Reliability
     • A measure of the extent to which something is consistent within itself
   • Split-half Method
     • Compares a person's performance on two halves of a test to measure internal reliability
   • Test-retest Method
     • Compares a person's performance on a test on two different occasions to measure external reliability

7. Sampling

   • Volunteer
     • Uses people who have put themselves voluntarily forward, usually in response to an advertisement
   • Random
     • Everyone in the target population has an equal chance of being selected
   • Stratified
     • Different types of people within a population are identified. A number of individuals are sampled from each category in proportion to their representation in the population
   • Snowball
     • When it is hard to identify suitable partcipants researchers start with a small number of suitable people then ask them where they might find more suitable people
   • Opportunity
     • Uses people from a target population available and willing to take part. It is based on convenience

8. Validity

   • Observer Bias
     • When what someone observes is influenced by their expectations
   • Concurrent validity
     • Comparing performance on a new test with an old test on the same topic
   • Face Validity
     • Whether the test looks like it is measuring what it intends to measure
   • External Validity
     • Made up of population, time and ecological validity
   • Internal validity
     • Whether the test measures what it is supposed to measure without the interference of EVs

9. Descriptive Statistics

   • Measures of Central Tendancy
     • Mean: (+) Incorporates all data (-) Anomalies skew results
     • Median: (+) Not affected by extreme values (-) Does not incorporate all data
     • Mode: (+) Can be used for all types of data (-) Does not incorporate all scores
   • Measures of Dispersion
     • Range (+) Easy to calculate (-) Skewed by anomalous results
     • Standard Deviation: (+) Less affected by extreme scores (-) Data must be interval or ratio

10. Inferential Statistics

    • Type 1 and Type 2 Errors
      • Type 1 Errors: The null hypothesis is rejected when it should have been accepted (a false positive)
      • Type 2 Errors: The null hypothesis is accepted when it should have been rejected (a false negative)
    • Spearman's Rho
      • Used to determine whether a correlation is statistically significant
      • To calculate it, you need to know the number of participants (N) and whether the hypothesis is one-tailed or two-tailed
      • Observed value must be greater than critical value to be significant
    • Chi Squared
      • Used for independent groups with nominal data
      • Count how many occurrences there are in each category and draw up contingency table
      • Degrees of freedom = (no. rows -1)x(no. columns-1)
      • Observed value must be greater than the critical value to be significant
    • Mann Whitney U
      • Used for independent groups with ordinal or interval/ratio data
      • Only works for experiments with 2 conditions
      • U must be less than the critical value to be significant
      • Allocate points by taking a target score from one group, then comparing it with all the scores in the other group, giving 1 point for every score that is higher than the target score and 1/2 if it is equal. Do this for all scores in both groups, one score at a time. Add up points. "U" is the lower total number of points
    • Wilcoxon T
      • Used for repeated measures and matched pair experimental designs
      • Need to know N and whether it is one/two tailed. Can only be done in experiments with 2 conditions
      • Calculate difference between the two scores in each of the two conditions. Rank data from highest to lowest. If difference is 0, omit result and reduce N. "T" = the sum of the LEAST frequent sign
      • T must be less than the critical value to be significant

11. Ethics

    • British Psychological Society Code of Conduct
      • Respect (confidentiality, anonymity, informed consent, deception, right to withdraw)
      • Competence (high quality research, peer review)
      • Responsibility (protection from harm, brief and debrief)
      • Integrity (honesty, avoiding fraudulence)
    • Socially-sensitive Research
      • This is research on controversial or sensitive topics. Psychologists must bear in mind imposed etic in their interpretation of findings
    • Animal Research
      • Arguments for: Animals are interesting to study in their own right. The research may be beneficial to animals. There is increased objectivity and control in research procedures. There are some situations where humans cannot ethically or practically be studied.
      • Arguments against: Animals are sentient beings that can feel emotional and physical pain. Using animals instead of humans is speciesist. Animals have rights - they should have an equal right to respect, yet they have no right to withdraw or informed consent. Generalisation - you cannot always extrapolate from animals to humans.
      • Existing Constraints: Animal research can only take place in licensed labs with licensed researchers on licensed projects (projects each individually undergo a cost-benefit analysis)
      • The 3 Rs: Psychologists should aim for reduction (use fewer animals), replacement (use alternative methods) and refinement (reduce stress and harm) when using animals

12. Qualitative Data

    • Types of qualitative analysis
      • Discourse analysis (studying oral or written discourses)
      • Ethnography (observing people in their natural environments)
      • Interpretative Phenomenological Analysis (understanding how people make sense of their experiences)
    • Coding
      • Thematic analysis (codes are developed from existing theories then looked for in the material)
      • Grounded theory (codes are allowed to emerge from the data, without being influenced by existing theories)
    • Key terms
      • Reflexivity (the recognition that a researcher's attitudes have an unavoidable influence on the research they are conducting)
      • Reality (The qualitative approach denies the existence of a single 'real' world, and sees each individual's subjective perspective as their 'truth')
    • Pros and Cons
      • Quantitative Data Pros: Easier to analyse and produces neat conclusions
      • Quantitative Data Cons: Oversimplifies human experience
      • Qualitative Data Pros: Represents the true complexity of human behaviour, gains access or feelings that may not be assessed with the quantitative method and provides rich detail
      • Qualitative Data Cons: Analysis can be affected by personal expectations. More difficult to detect patterns and draw conclusions