Correlation, survey, tests and case study

Correlational research asks how two variables go together, without changing either. The correlation coefficient runs from +1.0 through 0 to −1.0:

• Positive: both rise and fall together, such as hours of sleep and alertness next day.
• Negative: one rises as the other falls, such as late-night screen time and hours of sleep.
• Zero: no relation, such as shoe size and honesty.
• The closer the value is to +1 or −1, the stronger the relation; a value near 0 is weak. Correlation helps prediction, but it does not prove cause.

Survey research collects people's opinions, attitudes or facts about them, as a company does when it asks which brands of snack you have heard of.

• Interviews: face to face. Structured interviews have set questions in a set order (opening, main and closing questions); unstructured ones are flexible and the questions arise in conversation. They can be one-to-one, one-to-group, group-to-one or group-to-group.
• Questionnaires: printed or online forms. Open-ended questions let people answer freely ('Which city would you like to live in, and why?'); closed-ended questions have fixed answers (yes/no, choose one).
• Telephone surveys are quick and cheap, but people may not take them seriously.
• Surveys collect a lot of information fast, but people may give careless or untrue answers.

Psychological tests are not quick internet quizzes ('the first thing you see in this picture reveals your personality'). Real tests are:

• Objective: given and scored the same way for everyone, friend or stranger.
• Standardised, like a big chain restaurant whose food tastes the same in every city. This needs reliability (consistent scores, checked by test-retest, giving the test again later, or split-half, comparing the two halves), validity (it measures what it claims to; a psychology test full of maths questions is not valid) and norms (the average performance of a comparable group, against which a score is judged).

Kinds of tests:

• By language: verbal (needs reading and writing), non-verbal (symbols and pictures, usable by people who cannot read) and performance (doing a task, such as fitting blocks together).
• By administration: individual or group.
• By time: speed tests (strict time limit, easy items), power tests (no time pressure, items get harder). Most tests combine the two.

A case study is an in-depth study of one person, group or event, often by clinical psychologists: present problems, life history, family relationships, friendships, health. It uses many methods at once (interviews with the person and family, observation, tests), which is why it is discussed last.