Question 1

You give participants a list of five different sports and ask them to rank order which they would prefer to watch, from most liked through to least. What type of data have you collected?
A) Interval data
B) Ratio data
C) Nominal (categorical) data
D) Ordinal data

Accepted Answer

It is ordinal because, although the numbers exist in a particular order (1-5), the distances between each number can vary.

Question 2

You have collected a set of IQ scores from 500 participants. There are a few participants who got exceptionally high, or outlying, scores. How would you describe the distribution of this dataset?
A) Negatively skewed
B) Positively skewed
C) Normal
D) Bimodal

Accepted Answer

The small number of high scores would create a tail on the right of the graph, meaning the distribution has a positive skew.

Question 3

You have collected reaction time data from 200 participants. If the data are normally distributed, how many would you expect to have scores within ± 2 SD around the mean?
A) 95
B) 5
C) 190
D) 10

Accepted Answer

You would expect 95% of scores to fall within ± 2 SD around the mean. 95% of 200 would be 190 participants.

Question 4

How would you evaluate normality for a repeated t test?
A) Look at the distribution for each condition separately
B) Look at the distribution of the difference scores
C) Look at the distribution of all the data, ignore which condition the data are from
D) By calculating the variance

Accepted Answer

You can revise how to evaluate normality across different designs on p214

Question 5

You have run a study comparing two independent groups. For one of the groups the variance is calculated as 3. The following values represent possible variance scores for the other group. Which would suggest that the assumption of homogeneity of variance has been violated?
A) 3
B) 6
C) 9
D) 12

Accepted Answer

The largest variance should be no more than four times larger than the smallest. 3 * 4 = 12, so values up to 12 would not violate the assumption and values of 12 or more would.

Question 6

You have run a study comparing two independent groups. For one of the groups the variance is calculated as 2. The following values represent possible variance scores for the other group. Which would suggest that the assumption of homogeneity of variance has NOT been violated?
A) 12
B) 8
C) 4
D) 16

Accepted Answer

The largest variance should be no more than four times larger than the smallest. 2 * 4 = 12, so values up to 8 would not violate the assumption and values of 8 or more would.

Question 7

Imagine you are ranking data. You have three values of 12, which would have been ranked 2, 3 and 4. What value would you give as the rank score for these three values of 12?
A) 2
B) 3
C) 4
D) 12

Accepted Answer

Tied ranks are calculated as the mean of the ranks scores: (2 + 3 + 4) / 3 = 3.

Question 8

Imagine you are ranking data. You have four values of 8, which would have been ranked 6, 7, 8 and 9. What value would you give as the rank score for these four values of 8?
A) 7.5
B) 7
C) 8
D) 8.5

Accepted Answer

Tied ranks are calculated as the mean of the ranks scores: (6 + 7 + 8 + 9) / 4 = 7.5.

Question 9

You want to run an independent measure t test, but the data violate the assumption of homogeneity of variance. Which non-parametric equivalent should you use?
A) Spearman's r
B) Wilcoxon signed rank
C) Mann Whitney U
D) Chi square

Accepted Answer

See Figure 13.7 p216 for summary of parametric tests and their non-parametric equivalents

Question 10

You want to run a Pearson's correlation, but the data are not normally distributed. Which non-parametric equivalent should you use?
A) Spearman's r
B) Wilcoxon signed rank
C) Mann Whitney U
D) Chi square

Accepted Answer

See Figure 13.7 p216 for summary of parametric tests and their non-parametric equivalents

Quiz 13: When and Why Might I Need to Use Non-Parametric Statistics

You give participants a list of five different sports and ask them to rank order which they would prefer to watch, from most liked through to least. What type of data have you collected?

You have collected a set of IQ scores from 500 participants. There are a few participants who got exceptionally high, or outlying, scores. How would you describe the distribution of this dataset?

You have collected reaction time data from 200 participants. If the data are normally distributed, how many would you expect to have scores within ± 2 SD around the mean?

How would you evaluate normality for a repeated t test?

You have run a study comparing two independent groups. For one of the groups the variance is calculated as 3. The following values represent possible variance scores for the other group. Which would suggest that the assumption of homogeneity of variance has been violated?

You have run a study comparing two independent groups. For one of the groups the variance is calculated as 2. The following values represent possible variance scores for the other group. Which would suggest that the assumption of homogeneity of variance has NOT been violated?

Imagine you are ranking data. You have three values of 12, which would have been ranked 2, 3 and 4. What value would you give as the rank score for these three values of 12?

Imagine you are ranking data. You have four values of 8, which would have been ranked 6, 7, 8 and 9. What value would you give as the rank score for these four values of 8?

You want to run an independent measure t test, but the data violate the assumption of homogeneity of variance. Which non-parametric equivalent should you use?

You want to run a Pearson's correlation, but the data are not normally distributed. Which non-parametric equivalent should you use?