Validity - What you need to know

What is Validity?
Validity refers to the extent to which a measure is measuring what it intends to, or in other words, it refers to extent to which it reflects the reality that it claims to represent. The first definition is more relevant to internal validity, whilst the latter helps to explain external validity.

How many types of validity are there? 
In short, loads. But out of ones that you should know at A-level, there's two categories, and a couple of types under each heading.

Internal validity: the extent to which a measure or test is measuring what it intends to. There are typically thought to be three main types:

1. Face validity - validity at face value, the simplest form. It's basically what gets affected by confounding and extraneous variables and lack of control over these factors. 

2. Construct validity - is the IV or DV appropriate for the concept being investigated? For concepts such as stress, this is a major issue. How do you define stress? And how can it be properly measured if there's no standard definition? One study which is relevant from the clinical psychology bit is Lewinson's study on positive reinforcements and depression - can you really measure depression levels on a depression adjective checklist? 

3. Concurrent (criterion) validity - do other relevant measures get the same result as yours? For example, if you're intending to measure intelligence through a new method, do the IQ scores or school performances of your participants correlate with the results you're getting?

External validity: the extent to which a measure or findings reflect the reality that it claims to represent. Again, there are 3 main types:

1. Ecological validity - refers to how realistic the task and environment is. If your task is similar to the real life behaviour it's attempting to measure, it's going to have high ecological validity.

2. Population validity - refers to how relevant the population is. If you're attempting to generalise your findings to all eyewitnesses, for example, you should ideally have a wide variety of real-life witnesses, who vary in age, gender, occupation, intelligence, ethnicity, etc.

3. Temporal validity - refers to how relevant the time period is in affecting the findings. e.g. A study on attitudes conducted decades ago cannot be expected to have temporal validity due to how quickly attitudes shift in society. 


All of these essentially link back to generalisability, usefulness and applicability to real life situations, so it's a really useful issue to get your head around, and worth mentioning if you can.

The problems with DEFINING DYSFUNCTIONAL BEHAVIOUR

What is dysfunctional behaviour? It's a tricky question that not even the professionals can easily describe. The simplest definition is "behaviour which disrupts normal functioning such as social interaction and sustaining employment", but it's still pretty broad.

Rosenhan and Seligman (1984) attempted to define dysfunctional behaviour. They created four criteria for "abnormality" (this used to be the official term for what is now dysfunctional behaviour):

1. Statistical infrequency (if a behaviour is statistically very rare, it could be dysfunctional)
2. Deviation from social norms (if a behaviour is not "normal", it may be dysfunctional)
3. Failure to function adequately (such as causing observer discomfort and being irrational)
4. Deviation from ideal mental health (Jahoda's 7 criteria including aspects such as being able to act independently and having positive self-esteem)

Some of these criteria may seem fairly logical if you think about the term dysfunctional, however it would label the following behaviours as dysfunctional at one point or another: homosexuality, protesting (e.g. civil rights movement or feminism), inventions, science, having a high IQ, winning the lottery, e.t.c.

So, it's not ideal. It's also pretty ethnocentric. For example, the criterion of "being able to act independently" would actually be seen as dysfunctional in some cultures, as group decisions and democracy are seen as prevalent. Statistical infrequency would render high IQ dysfunctional, but depression functional. And everyone is irrational or has low self-esteem at one point or another; does that mean we necessarily all have a psychological disorder?

Evaluation issues for this section would include subjectivity, cultural bias, ethnocentrism, temporal validity, face validity and construct validity, as well as reliability.