RS4

Question 1

Variance

Answer 1

How much scores on a variable deviate from the mean

Question 2

Covariance

Answer 2

How much scores on two variables deviate from their respective means

Question 3

Spearman’s rho and Kendall’s Tau-b

Answer 3

Use when we have a smaller data set with a large number of tied ranks, ie: if many scores have the same rank. Kendall’s tau is lower as it takes account of people scoring the same.

Question 4

Point Biserial correlation

Answer 4

When one variable is categorical (dichotomous with just 2 categories eg: male and female) and the other is continuous eg: self-reported IQ.

Question 5

Phi correlation

Answer 5

2 dichotomous categorical variables (each has 2 levels)

Question 6

Face validity

Answer 6

From the appearance of the items, does the scale appear to measure what it claims to measure?

Question 7

Content validity

Answer 7

Do the items of the scale cover the important characteristics of the concept being measured? Good content validity involves researching in depth the literature etc on a topic and needs to reflect the variety of facets of the concept being measured. Consulting expert opinion may also be necessary.

Question 8

Concurrent validity

Answer 8

Does the scale correlate well with other measures of the same concept taken at the same time? eg: a more widely used and accepted test taken at the same time.

Question 9

Predictive valdity

Answer 9

Does the measure accurately predict future behaviour, eg: does IQ predict those that go to university? Usually unimportant in psychology.

Question 10

Construct validity

Answer 10

This involves developing theoretical and conceptual understanding of what is being measured. How well do we understand the construct? It is about theory development and general progress of knowledge and is not a technical type of validity that can be measured easily.

Question 11

Convergent validity

Answer 11

Do different sorts of measures of the same construct tend to intercorrelate? eg: self-reported measures of honesty, behavioural measures of honesty and a polygraph test. However, would not always expect different measures of same construct to correlate eg: behaviour may not always reflect attitudes.

Question 12

Discriminant validity

Answer 12

Do measures of apparently different concepts not correlate with each other? Opposite to convergent validity.

Question 13

Reliability is about:

Answer 13

a) the consistency of the items within the measure and

b) the consistency of a measure over time

Question 14

Cronbach’s alpha, split-half reliability and odd-even reliability are all measures of

Answer 14

Internal reliability

Question 15

Which is preferable to use out of split-half reliability, odd-even reliability and Cronbach’s alpha?

Answer 15

Cronbach’s alpha as this takes an average of every possible half of items correlated with every possible other half of items so therefore takes account of all possible ways of splitting items.

Question 16

Test-retest reliability

Answer 16

Correlating a measure given at different time points.

Question 17

Validity is concerned with

Answer 17

whether a measure actually measures what it is intended to measure.

Question 18

Partial correlation is used when

Answer 18

you want to measure the relationship between 2 variables whilst controlling for the effect that a third variable has on them both.

Question 19

Semi partial correlation or part correlation

Answer 19

Measures the relationship between 2 variables controlling for the effect of a third variable on only 1 of the variables.