Experimental Designs and Research Methods

Choosing how to find out

Psychologists do not have one way of investigating things. They have several, and each one trades control against realism. A tightly controlled study tells you clearly what caused what, but the situation may be nothing like ordinary life. A study that watches people in ordinary life is realistic, but you can no longer be sure what caused what. There is no method that wins on both counts. The question you are asking decides which trade is worth making.

The methods and what they do

Seven methods, each suited to a different kind of question.

Method to strength

Three kinds of experiment

All three look for the effect of one variable on another. They differ in where the study happens and in who changes the variable.

Which experiment is this?

A school moves its year 9 lessons from morning to afternoon for timetabling reasons. A researcher, who had nothing to do with the decision, measures how much the pupils remember in each. Which kind of experiment is this?

• A natural experiment
• A field experiment
• A laboratory experiment
• A case study

Watching behaviour properly

An observation study has to be more than just watching. Select the TWO things that make an observation study reliable.

• Deciding in advance which categories of behaviour will be recorded
• Checking that two observers using those categories agree with each other
• Watching for as long as possible so that more behaviour is seen
• Letting each observer decide what counts as interesting behaviour

Asking people, or studying one

Questions asked out loud, where the researcher can follow up an answer, make up an _____. The same questions in writing, sent to many people at once, make up a _____. Both rely on what people are willing and able to _____ about themselves. When a researcher wants depth instead of numbers, studying one person or a small group closely over time is a _____ study, which gives rich detail but is about too few people to _____ to everyone.

Association is not cause

A correlation measures whether two variables are associated, meaning they tend to vary together. Plotting one against the other on a scatter diagram shows the pattern: as screen time rises, hours slept might fall. Nothing has been changed by the researcher, so nothing has been shown to cause anything. Less sleep might follow from more screen time, more screen time might follow from lying awake, or something else entirely might be behind both. You do not need to calculate anything for this specification. You do need to say what an association can and cannot tell you.

Reading the scatter

A researcher records the screen time and the hours slept of 200 students, and finds that students with more screen time tend to sleep fewer hours. What can be concluded?

• The two variables are associated, but the cause has not been established
• Screen time reduces sleep, because the pattern is clear in the data
• Nothing at all, because a correlation is not a real research method
• Sleeping less must be what causes students to spend longer on screens

Design to drawback

Run the study the same way

A researcher tests whether people finish a puzzle faster standing up or sitting down, using the same participants in both conditions. Put the steps of a standardised procedure into order.

• Write the instructions that every participant will be given, word for word
• Decide by chance which half of the participants stands first and which sits first
• Read the same instructions to each participant and give the same puzzle
• Time the first condition, then run the second condition the same way
• Compare the two sets of times to see whether the position made a difference

Keeping it fair

In the standing and sitting study, select the TWO steps that help stop something other than the position affecting the result.

• Having half the participants stand first and half sit first
• Giving every participant exactly the same instructions and puzzle
• Using a harder puzzle for the sitting condition than for the standing one
• Letting participants choose the condition they would rather do first

A method choice, justified

Question: a psychologist wants to know whether children behave more helpfully in a classroom where the teacher praises helpfulness. Suggest a suitable method and justify it. Model answer: a field experiment would suit this question. The variable being changed is whether the teacher praises helpfulness, and the classroom is where the behaviour of interest actually happens, so the finding is more likely to hold in real classrooms than one produced in a room built for the study. The cost is control, because a classroom contains many things the researcher cannot keep equal between the two conditions. An observation study would be more realistic still, but because nothing would be changed it could not show that the praise made the difference. The marks are for weighing what each method gives up, not for naming one.

Where the plan goes wrong

Psychologists check a research plan against the guidelines published by the British Psychological Society, which set out how people taking part must be treated. Select the ONE sentence in this plan that those guidelines exist to prevent.

• Participants will be recruited from two year groups at a local school.
• Each participant will be given the same written instructions to read.
• Anyone who asks to stop once the study has begun will be told to continue.
• The results will be stored without any names attached to the records.

Designing the sleep study

A psychologist wants to know whether using a screen in the hour before bed affects how well people sleep. Work through the design decisions.

• The psychologist first considers simply recording how much screen time and how much sleep 200 people already get. What would that study be able to show?
• To test cause, the psychologist decides to control the screen use instead. Which method now fits?
• Every participant will spend one night with screen use and one night without. Which design is that, and what has to be guarded against?
• Before running it, the psychologist checks the plan against the BPS guidelines. What does that check require?

Justify your method

A psychologist wants to know whether pupils concentrate better in a quiet room or a busy one. Explain how you would investigate this, and justify the choices you make.

• Name the method you would use and say why it suits this particular question
• Explain what your method gives up, using the idea of control against realism
• State which experimental design you would use and give its main drawback
• Describe two things you would do to stop anything other than the noise affecting the result
• Explain why the plan would be checked against the BPS guidelines before it is run