Week 4 - Psychometrics II

Question 1

checking whether the internal structure of your test is ‘as expected’ is testing what? And how would you check that?

Answer 1

Construct Validity: running a ‘sweeeet’ factor analysis baby! (Said Jo).

Question 2

construct validity is an umbrella term for all kinds of validity except for _______?

Answer 2

face validity

Question 3

What two numbers do you need to calculate a person’s SEM (standard error of measurement)?

Answer 3

SD of test-takers’ scores
reliability of test

SEM = SD x square root of (1-reliability)

https://www.youtube.com/watch?v=PZDDWd-jUzM

Question 4

What is the 95% CI of a person who scores 123 on an IQ test (given that the SD is 15 and the reliability is .95)?

Answer 4

SEM = 15 x square root of (1-.95) = 3.35

3.35 x 2 = 6.7 (remember Jo said it wouldn’t matter if we used 2 or 1.96).
123 - 6.7 = 116.3
123 + 123 = 129.7

95% CI = 116 - 130

Question 5

To find out if the following scores are statistically different, work out the standard error of the difference for a man who first scored 133 on a test of happiness (yay for positivity!) and 143 on the same test after the intervention. The reliability of the test is .96 and the SD is 16.

Answer 5

SEdiff = 16*sqrt (2- .96 - .96) = 4.5

The man’s scores differ by 10 points, which is 10/4.5 = 2.2

Therefore we are 95% confident that the scores are statistically different. The man is significantly more satisfied after the intervention.

Question 6

Reliable change index

Answer 6

is the difference between two scores (143-133 = 10) divide by the SEdiff = 4.5
= 2.2

Question 7

what are the two ways to adjust for practice effects?

Answer 7

1. Re-standardise people’s scores for their 2nd attempt against a sample of 2nd attempt scores
2. Add or subtract a constant to an individual’s second score that reflects the average change in a standardisation sample

Question 8

A ROC curve plots the ____ on the Y axis and the ____ on the X axis.

Answer 8

A ROC curve plots the correct hit rate (sensitivity) on the Y axis, and the false hit rate (1-specificity) on the X axis

Question 9

true or false, a steeper ROC curve indicates that the test has better diagnostic power than a ‘straighter’ ROC curve?

Answer 9

True, this indicates that the test is good at distinguishing between those who have the disorder/disease and those who do not

Question 10

what is the best way to choose the best pass mark for a test when using a ROC curve?

Answer 10

The point on the curve where the sum of sensitivity and specificity is highest.

Question 11

Which area of psychology was signal detection theory first applied to?

Answer 11

Perception studies, but now it is also applied whenever a person is discriminating between two stimuli

Question 12

How do you disentangle sensitivity and response bias?

Answer 12

Hit rate - false positives

Question 13

List the five factors that the Stanford-Binet Intelligence test tests for

Answer 13

1. fluid reasoning (fluid intelligence)
2. knowledge (crystallised intelligence)
3. Quantitative reasoning
4. Visual-spatial reasoning
5. Working memory

Question 14

List the four factors that the Wechsler Intelligence scale for children (WISC-IV) tests for

Answer 14

1. verbal comprehension
2. perceptual reasoning
3. working memory
4. processing speed

Question 15

If a clinician who is giving a client an intelligence test notices that the client does exceptionally well in 2 of the 4 categories, but really poorly in the other 2, what should the clinician NOT do?

Answer 15

The clinician should avoid adding the scores together, as this averages out/ ignores the client’s deficits.

Question 16

Name the Big Five Factors of personality

Answer 16

OCEAN
openness 
conscientiousness
extroversion
agreeableness
neuroticism

Question 17

The MMPI has how many items and what sort of scale are they on?

Answer 17

567 true-false items