Standardisation Flashcards

1
Q

Define psychometrics and psychological testing

A

It’s the science of psychological measurement; psychological tests are a measuring device or procedure designed to measure psychology-related variables

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

List 5 different types of psychological tests and provide examples

A
  1. Mental ability (IQ tests, memory, vocabulary, spatial ability);
  2. Achievement (educational, course assessments, competence, experimental performance measures);
  3. Personality-type (no correct answer; assessing traits);
  4. Interests and attitudes (vocational, social psych questionnaires);
  5. Neuropsychological tests (memory, psychomotor coordination, abstract thinking, IQ, personality, etc)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What are the four key assumptions behind psychological tests?

A
  1. People differ on traits
  2. These traits are measurable
  3. These traits are relatively stable over time
  4. These traits relate to actual behaviour
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What are the advantages of standardisation?

A

It helps us interpret what the scores mean relative to the appropriate sample of people; raw scores alone are uninformative.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Explain what norm and normative sample are

A

Normative sample is a standardised score relative to a sample of people (reference sample/standardised sample); the norm is the data they yield

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What issues might we need to consider when recruiting a standardised sample?

A

A bigger sample is more likely to lead to an accurate and stable representation of the population (but
stability doesn’t guarantee representativeness); may need equal ratio of males to females, same age, socio-economic and educational distribution; same geographic origins of the population

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Define stratified cluster sampling

A

Deliberately recruiting representatives to get particular ratios of subgroups (as opposed to random
sampling)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

If stratified cluster sampling fails or is not possible, what’s another option?

A

Weighting (e.g. If male to female ratio is uneven, count each female as 1.5 people, and each male as .67 of a person)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Why is the normal curve (aka the Laplace-gauss curve) important in psychology?

A

If we can assume something has a normal curve, then knowing only the mean and standard deviation can tell us how someone’s score compares with everyone else

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

When does a distribution come closer to a normal curve?

A

When a sample is larger, and with a wider range of things measured

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

When a sample is normally distributed, what is the mean equal to?

A

Both median and mode (so 50% of people are above/below the mean)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

In a normal distribution, what percentage of scores are +/-1 SD around the mean?;
What about +/-2 SDs?

A

68%;

95%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What’s the general definition of cognitive impairment and what’s notable about the properties of this
definition?

A

An IQ of below 2 SDs below the mean (less than 70); this means “mentally retarded” is always in comparison with the rest of the population (it’s not absolute)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Why is it convenient for our distributions to be normal?

A

We can do more powerful (parametric) statistical tests, and makes scales more comparable

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What are the different strategies we can use to make a skewed distribution normal?

A

Do a non-linear transformation to make it more normal (e.g. take the square root or logarithm); or redesign the measure if possible (e.g. change wording of items)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Does performing a non-linear transformation change the rank order of scores?

A

No, it just stretches some bits of the scale more than others

17
Q

Why do we use standard scores?

A

Makes the interpretation simpler; anchors the mean and SD of the scale; can potentially compare performance across different scales

18
Q

What are the means and SD of:
z score?;
T score?;
IQ score?

A

Z score: mean = 0, SD = 1;
T score: mean = 50, SD = 10;
IQ score: mean = 100, SD = 15

19
Q

How do we covert a raw score into a z score?

A

z = X - Xbar / SD

20
Q

How do we covert a z score into a T score?

A

T = z(10) + 50

21
Q

How do we covert a z score into an IQ score?

A

IQ = z(15) + 100

22
Q

Does calculating a T score change the shape of the distribution?

A

No, it is a linear transformation (also avoids negative numbers)

23
Q

A person scores 70 on a test, where the mean is 85, and SD is 20. You have created a standardised
scale where the mean is 100 and the SD is 50. How would you find their score according to your
scale?

A

z = (70 - 85)/20 = -0.75;

-0.75(50) + 100 = 62.5

24
Q

Give 3 examples of linear transforms, and 3 of non-linear transforms?

A

Linear: z, T and IQ tests;

Non-linear: logarithms, square roots and percentile ranks

25
Q

How do we work out a percentile rank?

A

Calculate the z score, then look up a table of the standard normal distribution (larger portion for positive and smaller portion for negative)

26
Q

What are the advantages of using percentile ranks?;

Disadvantages?

A

Easy to understand for laypeople and easier to calculate; Confusion between percentile rank and percentage correct; location on scale means different things - ranks close to 50 include a smaller range of scores and are bunched up (those near 0 or 100 are more spread out)

27
Q

What are the properties of the Stanine scale?

A

Often used in school tests; has 9 divisions - each .5 SDs wide; the middle band (5) is from -.25 - +.25 SDs; 20% of people are in this middle band; Neale analysis of reading uses Stanines

28
Q

How can we create a narrative report from raw test scores or vice versa?

A

By coding certain scores into qualitative labels (e.g. 7 = outstanding) or coding labels into quantitative data (e.g. Outstanding = 7)

29
Q

Describe how norm-referenced scoring works

A

Score is relative to a norm; protected from the possibility that the test may be easier or harder; will yield a good distribution (can discriminate between good and bad test-takers); score is affected if the norm changes (disadvantaged if in a smart cohort)

30
Q

Describe how criterion-referenced scoring works

A

Score is relative to an absolute score (set standard); unaffected if the norm changes; can set absolute standards based on what people can actually do; score is affected by test difficulty; more likely a skewed distribution (fails to discriminate between people); possible for everyone to pass or to fail

31
Q

Why is converting your percentage score to a grade (1-7) a non-linear transformation?

A

It changes the shape of the distribution because different grades cover different ranges