4.1 inferential statistics Flashcards
(11 cards)
1
Q
what is a nominal level of measurement? with an example (3)
A
- shows categories of data represented by frequencies
- no relative numerical value
- e,g hair colour, sex, age
2
Q
what is the ordinal level of measurement? (3)
A
- data can be ranked in either ascending or descending order
- intervals are not equal
- 1st, 2nd, 3rd in a race
3
Q
what is the ratio level of measurement? (2)
A
- equal numerical intervals between scores AND has an absolute 0 point
- e.g temperature, speed
4
Q
what is the interval level of measurement? (2)
A
-equal numerical intervals between scores BUT LACKS an absolute 0 point
- scales, rating
5
Q
use sign test when…
A
- data is related (repeated measures + matched pair design)
- nominal level of measurement
- test of difference (directional hypothesis)
6
Q
use chi square when…
A
- independent group design
- AT LEAST ordinal (interval or ratio incl.)
- hypothesis predicts difference
7
Q
use willcoxon when…
A
- independent group design, matched pairs
- AT LEAST ordinal (interval or ratio incl.)
- hypothesis predicts difference
8
Q
use mann whitney u when…
A
- data is independent
(independent group design) - ordinal or above (interval/ratio) level of measurement
- test of difference (directional hypothesis)
9
Q
explain what is meant by the significance level (2).
A
- numerical value that is usually expressed in value including two decimal
places. - level that tells you the margin of error that could occur in your results and shows the probability of rejecting the null hypothesis.
10
Q
explain what is meant by ‘p≤/<0.01’.
A
the probability that the results will be down to chance is 1% (or less).
11
Q
explain why a psychological researcher might use a p value of p≤0.01 rather
than p≤0.05.
A
- stricter level of measurement such as p<0.01 reduces the margin of error
- 0.05 has more probability that the results are due to chance
- not acceptable in clinical trials
