The two problems every design has to handle

1. An experimental design is your decision about who takes part in which condition of the study.
2. Order effects appear when the same person does both conditions, because practice can lift the second score and boredom or tiredness can lower it, and that belongs to the order rather than to the independent variable.
3. Participant variables appear when different people do each condition, because a gap between the conditions could really be a gap between the two groups of people.

Independent measures

1. Different participants take part in each condition, so nobody is tested twice and you compare one group's scores against the other group's.
2. One strength is that there are no order effects at all, because nobody meets a second condition.
3. A second strength is that each participant sees only one condition, so they are less likely to work out what is being compared.
4. One weakness is that participant variables are wide open, because the two groups may not be evenly matched on ability or motivation.
5. A second weakness is that you need roughly twice as many participants to collect the same amount of data.
6. The usual answer to the participant variable problem is to let chance decide which condition each person goes into, since over a decent sized sample that spreads differences such as ability evenly across the two groups.

Repeated measures

1. The same participants take part in every condition, so each person's score in one condition is compared with their own score in the other.
2. One strength is that participant variables are almost entirely removed, because everyone acts as their own comparison.
3. A second strength is that you need far fewer participants, since every person supplies data in every condition.
4. One weakness is that order effects appear, because completing the first condition leaves people practised, bored or tired for the second.
5. A second weakness is that participants who have seen both conditions are far better placed to guess what the study is testing.
6. The standard answer to order effects is counterbalancing, where half the group does condition A first and the other half does condition B first, so any practice or fatigue lands on both conditions equally.

Matched pairs

1. You pair participants on the characteristics most likely to affect the result, then place one member of each pair in each condition.
2. One strength is that participant variables are cut down without repeating anyone, because the two groups begin level on whatever you matched them on.
3. A second strength is that order effects are still absent, since each person completes only one condition.
4. One weakness is that matching takes time and usually an extra test before the study can begin, which makes it the most expensive design to set up.
5. A second weakness is that you can only match on characteristics you thought of and could measure, so an unmatched difference can still be the real cause.
6. Which member of each pair goes into which condition should still be settled by chance, so that you cannot slide the stronger partners into the condition you are hoping will win.

Choosing between the three

1. If doing the second condition would spoil it, for instance because people would already know the answers, rule out repeated measures.
2. If participants are hard to recruit and the task can safely be done twice, repeated measures gives you the most data from the fewest people.
3. If one characteristic clearly drives the outcome and you can measure it beforehand, matched pairs is worth the extra setup.
4. Otherwise independent measures with people allocated by chance is the simplest workable choice.

Turning this into marks

1. Naming a design earns almost nothing, so say who does which condition in the study you have been given.
2. Name the problem the design invites and the control that handles it, such as counterbalancing for order effects.
3. The mark most often lost comes from giving only a strength when both a strength and a weakness were asked for.