What do I need to know about validity and how can I understand it?

Validity is a complex issue to understand as it is closely linked with reliability.

Ultimately, it can be defined as the extent to which something measures what it is supposed to. In simple terms, this means you want to ask the question how accurate is a piece of research in terms of its sample, measures used and the design?

There are a number of types of validity but the main ones to think about when evaluating are as follows:

• External validity refers to how far you can apply the results of an experiment to the outside world, beyond the original test environment.

Subtypes of external validity include population and ecological validity.

POPULATION validity is whether or not the sample of people that participated in the study are representative of the general population. Can the findings be generalized to other groups of people or was the original sample too limited?

For example, say that 10 university students aged 18-20 took part in research about memory. First of all, this is a very small sample size and cannot account for much human variation. It also does not represent what you might find if your selected sample included people of an older age. Students are more likely to have a sharper memory retrieval and a more extensive memory capacity than those who are older and perhaps no longer working. This means your sample is not valid or generalisable to the general population.

ECOLOGICAL validity is the extent to which research can be applied to real life. This is apparent in terms of the environment the research was carried out in and the task used to measure variables.

e.g. A laboratory environment is likely to be an unusual and unfamiliar setting for most people and so their behaviour may not be the same as it is normally.

When talking about the ecological validity of test materials be careful not to confuse this with internal construct validity (as explained later). An example of ecological validity with the task could be shown in Milgram's famous experiment. Participants believed they were giving electric shocks to another person if they got a question wrong. This is a highly distressing task and one that is uncommon and unfamiliar to everyday life.

• Internal validity is the extent to which the study was conducted without error and without confounding variables. This is not the same as reliability! (remember reliability is how replicable an experiment is).

It is good to remember the difference between internal and external like this: internal = the accuracy INSIDE an experiment and external = the accuracy generalizing to OUTSIDE the experiment.

A subtype of internal validity that you might want to know about is construct validity.

CONSTRUCT validity is essentially the same as the general definition of validity. It is the extent to which the dependent variables measure what they are supposed to.

For example, if you were to use a questionnaire with only closed questions to measure happiness it would be limited. It is restricted by social desirability, the questions included, the day/time of day and collecting only quantitative data may not reflect the complexity of an emotion like happiness.

Other keywords defined •GENERALISABLE = the extent to which research can be applied to other settings •DEPENDENT VARIABLE = the variable that is measured in an experiment to see the effect of the independent (manipulated) variable •SOCIAL DESIRABILITY = behaving in a way that you think is favourable and expected of you