validity

Validity

Validity is concerned with legitimacy, e.g. truth

Concerned with whether the study and findings provide a true representation of life

what are the two types of validity?

Internal validity and external validity

What is internal validity?

• Measures whether the DV is genuinely affected by the IV

• High when the confounding variable are well controlled

• Indicates the study measures what intends to measure

what is external validity?

Extent to which result reflect real life behaviour e.g. Mundane realism.

Assess his general availability across places times and population e.g. population validity

mundane realism

How typical or every day the research task is EG exam equals high MRI scans equals low

representative

representativeness – how similar the sample is to the wider population

• Diverse samples age gender ethnicities = high population validity

• Limited samples = biases such as ethnocentrism culture bias and dualism gender bias

Research location affects validity

• More control – higher reliability and low external validity

• Real life settings – high validity lower reliability

three specific validity issues

• Research or bias

• Demon characteristics

• Social desirability

research bias

• research a bias occurs when the researcher unintentionally influences the studies designed conduct analysis

• solution – use a double blind procedure – neither researcher or participant knows the true aims or group assignment

Benefit

• Prevent research or bias

• Reduce demand characteristics

Commonly used in drug trials with placebo and experimental groups

demon characteristics and social desirability bias

Demand characteristics – what participants guess the studies aims and change behaviour?

Social desirability – participants give answers that make them look good

Screw you affect – participants deliberately act to disrupt or ruin the results

Both reduces the validity of findings as behaviour no longer reflects true thoughts or feelings

demon characteristics and social desirability

Single blind procedures – participants are unaware of the researchers aims until after the study research researchers know the aim

• Reduces the demand characteristics and socio our ability bias

• To further reduce social desirability bias you should

• Emphasise honesty and accuracy in response

• Guarantee anonymity and confidentiality

• You face-to-face questioning to discourage deliberate deception

improving validity

• Changes to location of the research – a nature of the task can improve for ensuring that research is conducted in the field rather than laboratory or realistic task. That person would normally complete in day-to-day life.

• Changes the sampling technique can target population – can also improve external village of the results

five methods to assess validity

• Face validity

• Predictive validity

• Content validity

• Concurrent validity

• Construct validity

Face validity

• faithfully is simply whether the test appears at face value to measure what he claims

• Test where the purpose is clear even to naive you respondent

predictive validity

The degree to which a task accurately forecast of future outcome on the more broadly related to topic

Do the findings apply indifferent and in more varied situations?

For example, do those with high IQ scores gain higher grade than exam?

content validity

The objectively checks whether the method of measuring behaviour is Acura and decides whether it is fair test which achieve the aim of the study

This can be carried out by asking an expert in the specific area of behaviour to check if the test is valid

concurrent validity

Validating a measurement by comparing it with an established one that has non-validity

A similar result occurred in both tests then the new test has concurrent validity

If not, then the new test would have to be redesigned and carried out

construct validity

Construct validity is about how well a test or measurement to can actually be measured the things is supposed to measure

To have good construct village tea you need –

The test should include all the important parts that make up the construct

The test should not include anything that’s unrelated to the construct