Psychology Research Methods Mind Map

The researcher manipulates an independent variable and measures its effect. Laboratory, field, natural and quasi-experiments differ in how much control the researcher has.

Measure the relationship between two variables without manipulating either. A correlation does not show cause and effect.

Naturalistic or controlled, covert or overt, participant or non-participant. Behaviour categories and time or event sampling make them more systematic.

Questionnaires and interviews, from structured to unstructured. They reach thoughts and feelings, but answers can be affected by social desirability bias.

In-depth studies of one person or a small group. Rich detail, but hard to generalise.

The independent variable is changed by the researcher; the dependent variable is measured.

Defining variables so they can be measured, for example memory as the number of words recalled out of twenty.

Extraneous variables could affect the dependent variable. If one varies systematically with the independent variable, it becomes confounding and threatens the conclusions.

The null hypothesis predicts no difference or relationship. The alternative hypothesis predicts one: directional (one-tailed) or non-directional (two-tailed).

Different participants in each condition. No order effects, but participant variables can differ between groups. Random allocation helps.

The same participants take part in every condition. Controls participant variables, but risks order effects such as practice and fatigue. Counterbalancing reduces these.

Participants are paired on key characteristics, then one of each pair goes into each condition. Fewer participant differences, but matching takes time.

Every member of the population has an equal chance of selection. Unbiased in principle, but hard to achieve.

Systematic: every nth person from a list. Stratified: subgroups sampled in proportion to their size in the population, so the sample is more representative.

Whoever is available, often students. Quick and easy, but likely to be unrepresentative.

People who choose to take part, for example by answering an advert. Easy, but volunteers may differ from everyone else.

Consistency. Test-retest reliability checks results over time; inter-rater reliability checks agreement between observers.

Whether the independent variable really caused the change. Threatened by demand characteristics and investigator effects; improved by standardised procedures, randomisation and single or double-blind designs.

Whether results generalise to other settings (ecological validity), people (population validity) and times.

The BPS Code of Human Research Ethics: informed consent, avoiding unnecessary deception, protection from harm, confidentiality, the right to withdraw and debriefing.

Nominal (categories), ordinal (ranked data) and interval or ratio (equal units on a scale).

Measures of central tendency (mean, median, mode) and of spread (range, standard deviation), plus graphs.

Psychology usually uses p < .05: a less than 5% probability of the result if the null hypothesis were true. A Type I error is a false positive; a Type II error is a false negative.

Ask: difference or correlation? Independent or related design? What level of data?

• Nominal: chi-square (independent), sign test (related)
• Ordinal: Mann-Whitney U (independent), Wilcoxon (related), Spearman's rho (correlation)
• Interval, normally distributed: independent t-test, related t-test, Pearson's r (correlation)

University courses add ANOVA for more than two conditions, regression to predict one variable from others, and effect sizes and confidence intervals alongside p values.