Deck 15: Non-Parametric Tests

Question 1

non-parametric and parametric tests make different assumptions about data

Answer 1

True

Question 2

Non-parametric tests assume that the data are on a nominal scale

Answer 2

False

Question 3

Nonparametric tests are still inferential tests.

Answer 3

True

Question 4

All measurement scales can be manipulated with mathematical operations like addition and division

Question 5

Statistics can be manipulated for the author's benefit, so sometimes we must be careful how must we trust them.

Question 6

Nonparametric tests do not necessarily test differences between means .

Question 7

skewed data is measured on a scale where 0 indicates very little to skew.

Question 8

The Mann-Whitney U-test is the nonparametric equivalent of a dependent samples t-test.

Question 9

The null hypothesis for a Mann-Whitney U-test states that the distributions of ranked data for both groups are different.

Question 10

The Mann-Whitney U test is a nonparametric method that compares means between two groups

Question 11

When conducting a Mann-Whitney U-test, if your calculated U-value is smaller than that of the critical value, than the test is considered statistically significant.

Question 12

The degrees of freedom in a Wilcoxon Signed-Rank is based on the number of matched pairs as opposed to the number of individual data points.

Question 13

The Wilcoxon Signed-Rank test produces a t value. If the observed t -value is greater than the critical value, then the results are deemed to be statistically significant.

Question 14

The Wilcoxon signed-rank test should be used in place of its parametric equivalent when the data is non-normal or on an ordinal scale

Question 15

One method of dealing with skewed data is to use its rank order instead of actual values.

Question 16

There are three types of skewed data: positive, negative, and one-dimensional

Question 17

If the data are skewed, one should try use a parametric test instead of resorting to nonparametric tests.

Question 18

Nonparametric tests do not produce a significance level. Instead, they generate effect sizes.

Question 19

Ranked data is best described to be on an ordinal scale.

Question 20

Generally speaking, if a nonparametric test is used on rank data, the dataset is on a ___________ scale.

A) ratio.
B) nominal.
C) interval.
D) ordinal.

Question 21

Why is it the case that skewed data poses a problem for parametric tests?

A) There mean is larger than the median.
B) The assumption of normality is violated.
C) The data does not represent the population.
D) The assumption of independence is violated.

Question 22

Which of the following is not an assumption of parametric tests?

A) Observations are independent
B) The data are either on an interval or ratio scale
C) Central limit theorem holds
D) The data are skewed positive.

Question 23

Which of the following characteristics would lead you to choose a non-parametric test over a parametric test?

A) Skewed data
B) Having data that is not normally distributed
C) Data that are ranked
D) All of the above

Question 24

Why might a spearman correlation be appropriate for a given dataset?

A) The data are on a ratio scale
B) The sample size is too large
C) The sample is skewed with a small sample size
D) The data are on a nominal scale

Question 25

The Perception of Competitiveness Scale (PCS) developed by Carter and Weissbrod (2011) is a likert scale. If you assume that the scale is ordinal and are examining its relation to age or some other ranked data, you should…

A) use a Spearman correlation
B) use a Mann Whitney U test
C) use a Wilcoxon Signed-Rank test
D) use a Pearson correlation

Question 26

In a spearman correlation, how do you determine the strength of the correlation coefficient?

A) Negative correlations are weaker than positive correlations
B) Correlations are weaker as they approach 0.
C) The strength of the correlation depends on the sample size. Smaller samples tend to be stronger
D) Correlations become stronger as they approach 0.

Question 27

The Mann-Whitney U test makes assumptions about which of the following?

A) The data are quantitative
B) The data are normally distributed and generalize to the population
C) Observations are independent
D) Equal variances in both groups

Question 28

In which of the following situations should a Wilcoxon signed-rank test be used?

A) When matched data is normally distributed and ranked
B) When observations are non-independent and the data are skewed
C) When you wanted to use an independent samples t-test but the data ends up being heavily skewed.
D) When the sample size is very large and on a ratio scale.

Question 29

Why is it a problem if data are skewed when considering parametric tests?

A) The assumption of independence is violated
B) The data is qualitative
C) The assumption of normality is violated.
D) There is no method to calculate estimates of population parameters

Question 30

After conducting a Mann-Whitney U test, how do you determine statistical significance?

A) If the U-statistic is smaller than the critical value, the result is non-significant.
B) If your calculated U-value is smaller the critical value, than the result is statistically significant.
C) If the U-value is smaller than median rank, the result is statistically significant
D) If the U-value is greater than the median rank, the result is statistically significant.

Question 31

How is a Mann-Whitney U test and a Wilcoxon Signed-Rank test different?

A) the Mann Whitney U test is used on matched data and the Wilcoxon Signed-Rank test is not.
B) The Mann-Whitney U test is a parametric procedure, and the Wilcoxon Signed-Rank test is a nonparametric procedure.
C) The Mann-Whitney U test is used on 3 or more groups, and the Wilcoxon Signed-Rank test is used to test for relationships
D) The Mann-Whitney U test is used for two independent groups and the Wilcoxon Signed-Rank test is used for paired or matched data.

Question 32

If you have 12 pairs of participants what is your critical value for a two-tailed Wilcoxon Signed-rank test at $\alpha$ = .01?

A) 7
B) 3
C) 5
D) 9

Question 33

A researcher has 16 matched pairs for a Wilcoxon Signed-Rank test. For this two-tailed Wilcoxon Signed-Rank test to be significant at $\alpha$ = .05, the test statistic must be…

A) greater than 30
B) less than 30
C) greater than 25
D) less than 25

Question 34

A researcher has two independent groups with ranked data with 20 participants in each group. When using a Mann-Whitney U test (assume assuming α = 0.05), the researcher has data that is statistically significant if…

A) the U statistic is less than 127
B) the U statistic is greater than 127
C) the U statistic is less than 113
D) The U statistic is greater than 113

Question 35

Consider the following dataset on standardized test scores: 134, 145, 142, 145, 146, 152, 155, 155, 159. There are _______ ties and the score 152 is in the _____ rank.

A) 2; 6
B) 3; 6
C) 2; 5.5
D) 3; 5.5

Question 36

A researcher conducted a correlational study. He studied the relationship between individuals' perceptions of their own bodies on a rating scale (higher scores meant most positive self-perception) and their perception of the college (higher scores meant more positive perception of the college). He found a spearman correlation of -.68 and the result is statistically significant. What does this mean?

A) Body image and college ratings are positively correlated: those who have better self-perception of their body had worse perceptions of the college.
B) Body image and college ratings are positively correlated: those who have better self-perception of their body had better perceptions of the college
C) Body image and college ratings are negatively correlated: as self-perception of one's body got higher, perception of the college got lower.
D) Body image and college ratings are negatively correlated: as self-perception of one's body got higher, perception of the college got higher.

Question 37

Consider the following data: 6, 8, 10, 10, 10, 11, 12, 15, 18, 18, 19, 20, 24, 25. The data in rank 4 is ___ and rank 11 is ____.

A) 10; 19
B) 12; 19
C) 10; 20
D) 12; 20

Question 38

If you have quantitative data that is skewed, which of the following can be used to "unskew" it?

A) rank the data
B) use the mode
C) use the median
D) convert it to a ratio or interval scale.

Question 39

If using a two-tailed spearman correlation with 10 pairs and an alpha level of .05, what is the critical value?

A) .745
B) .700
C) .564
D) .648

Question 40

A statistical method to measure hidden or latent variables from observed data is…

A) spearman correlation
B) factor analysis
C) eugenics
D) Wilcoxon sign-ranked test

Question 41

Supporters of eugenics would likely agree that…

A) intelligence is developed via environmental influences
B) intelligence is inherited
C) intelligence cannot be measured objectively
D) None of the above.

Question 42

With regards to skewed distributions, which best describes them?

A) There are two types of skewed distribution, and if a dataset is heavily skewed enough, parametric tests are not suitable for it.
B) There are two types of skewed distributions, and skewed data can be on interval, ratio, or ordinal scales
C) Skewed data are ranked or qualitative, making it unsuitable for parametric analyses
D) Skewed data do not pose issues for parametric tests, but are not suitable for nonparametric tests.

Question 43

When using nonparametric analyses, you are generally dealing with __________ instead of ___________ and __________.

A) rank; means; standard deviations
B) rank; medians; interquartile range
C) means; rank; range
D) rank; means; sample size

Question 44

Data on an ordinal scale are generally analyzed by…

A) nonparametric analyses
B) parametric analyses
C) either parametric or nonparametric analyses
D) there is insufficient information about the data.

Question 45

The data below that records the number of inches of snowfall in different parts of the country
Rank the data.

Question 46

The data below that records the number of inches of snowfall in different parts of the country
What is the 7^th rank?

Question 47

The data below that records the number of inches of snowfall in different parts of the country
Are there any ties? If so, at which rank?

Question 48

A researcher hypothesizes that individuals' interest in politics is related to the amount of television they watch. He asks participants to rate on a scale from 1 (not interested at all) to 10 (extremely interested) the extent to which they are interested in politics, and also asks for the amount of television they watch. Below are the data
If calculating a spearman correlation, what is the null hypothesis?

Question 49

A researcher hypothesizes that individuals' interest in politics is related to the amount of television they watch. He asks participants to rate on a scale from 1 (not interested at all) to 10 (extremely interested) the extent to which they are interested in politics, and also asks for the amount of television they watch. Below are the data
What is the critical value, assuming a two-tailed test and alpha of .05?

Question 50

A researcher hypothesizes that individuals' interest in politics is related to the amount of television they watch. He asks participants to rate on a scale from 1 (not interested at all) to 10 (extremely interested) the extent to which they are interested in politics, and also asks for the amount of television they watch. Below are the data
What is the observed correlation?

Question 51

A researcher hypothesizes that individuals' interest in politics is related to the amount of television they watch. He asks participants to rate on a scale from 1 (not interested at all) to 10 (extremely interested) the extent to which they are interested in politics, and also asks for the amount of television they watch. Below are the data
What is the conclusion?

Question 52

Do men and women differ in the way they rate the cuteness of babies? In a study, a team of researchers examined whether gender differences existed in ratings for infants. The data below show average ratings by male and female participants for infant pictures. The team used a Mann-Whitney U test to compare the two groups. Higher scores indicate that the participant thought the image represented a cuter baby.
What is the critical value, if you assume alpha = .05 and a two-tailed test?

Question 53

Do men and women differ in the way they rate the cuteness of babies? In a study, a team of researchers examined whether gender differences existed in ratings for infants. The data below show average ratings by male and female participants for infant pictures. The team used a Mann-Whitney U test to compare the two groups. Higher scores indicate that the participant thought the image represented a cuter baby.
Calculate the Mann-Whitney U-test U-value and determine the smaller of the two values.

Question 54

Do men and women differ in the way they rate the cuteness of babies? In a study, a team of researchers examined whether gender differences existed in ratings for infants. The data below show average ratings by male and female participants for infant pictures. The team used a Mann-Whitney U test to compare the two groups. Higher scores indicate that the participant thought the image represented a cuter baby.
Is the difference significant if the test were two-tailed?

Question 55

In 2017, Nintendo (a Japanese video game company) released a new console called the Nintendo Switch. Nintendo was interested to see how existing customers with their previous generation console (i.e., the Wii U) rate the new product. Below are data of individuals who rated the first console and now rated the Nintendo Switch. Nintendo decided that given the small sample and ranked data, they would use a Wilcoxon Ranked-Sign test.
$\begin{array}{cc}\hline\text { Wii U } & \text { Switch } \\\hline 7 & 8 \\5 & 9 \\6 & 7 \\4 & 2 \\8 & 6 \\9 & 10 \\2 & 9 \\4 & 8 \\7 & 9 \\7 & 8 \\\hline\end{array}$ What is the critical value for this test, assuming two tails and $\alpha$ = .05?

Question 56

In 2017, Nintendo (a Japanese video game company) released a new console called the Nintendo Switch. Nintendo was interested to see how existing customers with their previous generation console (i.e., the Wii U) rate the new product. Below are data of individuals who rated the first console and now rated the Nintendo Switch. Nintendo decided that given the small sample and ranked data, they would use a Wilcoxon Ranked-Sign test.
What is the test statistic after calculating the Wilcoxon Signed Rank test?

Question 57

In 2017, Nintendo (a Japanese video game company) released a new console called the Nintendo Switch. Nintendo was interested to see how existing customers with their previous generation console (i.e., the Wii U) rate the new product. Below are data of individuals who rated the first console and now rated the Nintendo Switch. Nintendo decided that given the small sample and ranked data, they would use a Wilcoxon Ranked-Sign test.
What is the conclusion?

Question 58

Explain why non-parametric tests are important. Be sure to discuss when they are used, and any important controversies that might influence their use.