Multiple-Choice Exam Questions

The designation preceding each question (e.g., Ch 1.1) indicates the chapter and section of the text that the question draws upon. Separate test items for the same section of the text are distinguished by letters. For example, Ch 1.5A is the first test item pertaining to Chapter 1, Section 4, and Ch 1.5B is the second item pertaining to that section. You will probably want to delete these designations from your exam. The Þ signs below indicate correct answers.

Chapter 1: Statistics and Variables

Ch 1.1

1. A research calculated the percentages of students who are female at a sample of colleges and universities. The unit of analysis in this study was:

a) females.

b) gender.

Þ c) the college or university.

d) percentages.

e) a sample.

Ch 1.2

2. Generalizing from a survey of American colleges and universities to all American colleges and universities is an example of:

a) measurement.

b) a continuous variable.

c) a census.

Þ d) inferential statistics.

e) descriptive statistics.

Ch 1.3

3. A parameter is to a statistic as:

a) collectively exhaustive is to mutually exclusive.

b) discrete is to continuous.

c) a nominal variable is to an ordinal variable.

Þ d) a population is to a sample.

e) descriptive statistics is to inferential statistics.

Ch 1.5A

4. The General Social Survey asked respondents the meanings of each of ten words. The variable #CRCT.WORD reports the number of correct answers. The level of measurement of this variable is:

a) nominal.

b) ordinal.

Þ c) interval/ratio.

d) none of the above.

e) can't tell because there is not enough information.

Ch 1.5B

5. The General Social Survey asked respondents to assess their own health as excellent, good, fair, or poor. The level of measurement of this variable is:

a) nominal.

Þ b) ordinal.

c) interval/ratio.

d) none of the above.

e) can't tell because there is not enough information.

Ch 1.5C

6. The General Social Survey asked respondents what region of the country they lived in at age 16. The level of measurement of this variable is:

Þ a) nominal.

b) ordinal.

c) interval/ratio.

d) none of the above.

e) can't tell because there is not enough information.

Ch 1.6A

7. Suppose that a researcher conducting a survey asks respondents their annual incomes using these values: \$20,000 or less, \$20,000 thru \$60,000; \$60,000 or more. A problem with this set of values is that:

a) they are measured at the nominal level.

b) they are not continuous.

c) they are population data.

d) they are not collectively exhaustive.

Þ e) they are not mutually exclusive.

Ch 1.6B

8. Which of the following sets of categories is not collectively exhaustive?

a) Frequency of newspaper reading: Daily; Few times a week; Once a week; Less often; Never.

Þ b) Favorite TV shows: Drama; Comedy; Sports.

c) Employment status: Employed; Not employed.

d) Self-reported health: Excellent; Good; Fair; Poor.

e) Education: Years of school completed.

Ch 1.9A

9. The CIA’s World Factbook summarizes information about each country. Information for each country includes such variables as population size, gross national product, and mortality rates. Variables described in the World Factbook are:

a) nominal variables.

b) discrete variables.

c) dichotomous variables.

Þ d) ecological variables.

e) bivariate.

Ch 1.9B

10. Consider the argument: "States with higher percentages of drivers under 20 years old have more pickup trucks per 1,000 population.” Therefore, pickup trucks are mostly driven by teenagers. This argument:

a) uses discrete data.

b) uses inferential statistics.

c) is not collectively exhaustive.

Þ d) involves an ecological fallacy.

e) uses multivariate analysis.

Chapter 2: Frequency and Percentage Distributions

Ch 2.2A

11. We often use percentages because:

a) percentages are more scientific.

b) percentages are more precise.

c) percentages are presentation quality.

Þ d) percentages make comparisons easier.

e) percentages are collectively exhaustive.

Ch 2.2B

12. Compared with a small sample, a large sample has percentages that are:

a) more discrete.

b) more ecological.

c) more cumulative.

Þ d) more reliable.

e) more interval/ratio.

Ch 2.2C, 12.4

13. We can assess the stability or reliability of percentages by considering:

Þ a) the total number of cases on which percentages are based.

b) the size of the percentage.

c) inferential statistics.

d) all of the above.

e) none of the above.

Ch 2.3

14. Consider this table:

Table 2.4. Self-Reported Health

(in percentages)

Health Percent

Excellent 30.2

Good 47.4

Fair 17.0

Poor 5.4

Total 100.0

(N) (493)

Which of the following is true?

Þ a) The cumulative percentage who report at least good health is 77.6.

b) A spot map would be a good way to display these percentages.

c) The frequency is too small for stable percentages.

d) All of the above.

e) None of the above.

Ch 2.5

15. If values are not mutually exclusive and collectively exhaustive:

a) the variable is a subset.

b) the researcher commits an ecological fallacy.

Þ c) percentages will not sum to 100.

d) the variable is continuous.

e) all of the above.

Ch 2.7

16. A researcher interested in variations in Asian Americans’ political attitudes restricted her analysis to only Asian Americans in the General Social Survey. The Asian Americans that she analyzed are:

Þ a) a subset.

b) outliers.

c) discrete.

d) missing data.

e) ecological.

Ch 2.8A

17. Generally, ordinal variables are best displayed visually with:

a) cumulative percentages.

b) collapsed values.

c) pie charts.

Þ d) bar graphs.

e) subsets.

Ch 2.8B

18. In a bar graph:

a) the width of bars should reflect the number of cases.

b) values of ordinal variables should usually be listed from tallest to shortest.

Þ c) the vertical axis should usually begin at zero if possible.

d) none of the above.

e) all of the above.

Ch 2.9

19. A bar graph is a useful way to spot:

a) significant digits.

b) aggregate data.

Þ c) outliers.

d) none of the above.

e) any of the above.

Ch 2.10

20. In maps showing the distribution of a variable across the 50 American states, a

researcher can reduce the visual effect of differences in geographic size by using:

a) an area map.

Þ b) a spot map.

c) a continuous variable.

d) a discrete variable.

e) raw data.

Chapter 3: Averages

Ch 3.1A

21. A mode:

b) can be found only for discrete variables.

b) is usually less than the mean.

Þ c) is the value of the most frequently occurring score.

d) is usually less for ordinal than for interval/ratio variables.

e) is useful only for symmetrical variables.

Ch 3.1B

22. A distribution that has one "hump" is:

a) symmetrical.

b) dichotomous.

Þ c) unimodal.

d) skewed.

e) ecological.

Ch 3.3

23. Consider these scores: 0, 3, 1, 5, 1. The mean is:

a) 1.

Þ b) 2.

c) 3.

d) 5.

e) none of the above.

Ch 3.4A

24. The median is less sensitive than the mean to:

Þ c) the skewness of a variable.

a) the sum of squares.

b) dichotomous variables.

d) all of the above.

e) none of the above.

Ch 3.4B

25. Means can be greatly influenced by:

Þ a) outliers.

b) the sum of squares.

c) interval/ratio variables.

d) all of the above.

e) none of the above.

Ch 3.4C

26. Which of the following always produces the smallest sum of squares?

a) the mode.

b) the median.

Þ c) the mean.

d) a symmetrical distribution.

e) a bimodal distribution.

Ch 3.5

27. The variable LIKESCI reports whether students like or do not like science. “Like science” is coded 1 and “does not like science” is coded 0. If 75 percent of the students at a school like science, the mean is:

a) 0.

b) .25.

c) .50.

Þ d) .75

e) 1.00.

Ch 3.6A

28. In a symmetric, unimodal distribution:

a) the median equals the mean.

b) the mode equals the median.

c) the mean equals the mode.

Þ d) all of the above.

e) none of the above

Ch 3.6B

29. The mean years of the GSS variable AGE KD BRN is higher than the median, so we know that the distribution of respondent’s age when first child was born is:

a) collapsed.

b) normal.

c) leptokurtic.

d) bimodal.

Þ e) positively skewed.

Ch 3 (throughout)

30. Missing data should be included when finding:

a) the mode.

b) the median.

c) the mean.

d) all of the above.

Þ e) none of the above.

Chapter 4: Measures of Variation

Ch 4.1A

31. The variance measures deviation around the:

a) mode.

b) median.

Þ c) mean.

d) sum of squares.

e) standard deviation.

Ch 4.1B

32. The variance:

a) is the square of the standard deviation.

b) can be 0.

c) is never negative.

Þ d) all of the above.

e) none of the above.

Ch 4.3A

33. The number of standard deviations a score lies from the mean is:

Þ a) the case’s Z-score.

b) the case’s kurtosis.

c) the standard error.

d) the confidence interval.

e) the sampling distribution.

Ch 4.3B

34. Z-scores for a variable:

a) are normally distributed.

b) have a standard deviation of 0.

Þ c) sum to zero.

d) all of the above.

e) none of the above.

Ch 4.4

35. All normal distributions:

a) have a mean of 1.

b) have sampling distributions.

Þ c) are symmetrical.

d) all of the above

e) none of the above.

Ch 4.5A

36. The standard deviation of a sampling distribution:

a) applies only to population data.

b) can be larger than the standard deviation for the population.

c) is the sum of squares.

Þ d) is the standard error.

e) is always greater than 1.

Ch 4.5B

37. The central limit theorem tells us that the larger the size of a sample, then:

a) the smaller the variance.

Þ b) the smaller the standard error.

c) the less the mean.

d) the greater the standard score.

e) the more skewed the variable.

Ch 4.5C

38. The central limit theorem tells us that the larger the size of a sample, the closer the standard error is to:

a) the population’s standard error.

b) the mean.

Þ c) .

d) 1.00.

e) the standard score.

Ch 4.6A

39. If the 95 percent confidence interval is between 2.5 and 2.7, we know that:

a) the distribution is normal.

b) the standard error is .2

c) the distribution is symmetric.

d) 95 percent of the sample’s scores are between 2.5 and 2.7.

Þ e) the mean is 2.6.

Ch 4.6B

40. A 99 percent confidence interval includes ______number of scores than/as a 95 percent confidence interval.

a) a smaller

b) the same

Þ c) a larger

d) a more normal

e) a more standardized

Chapter 5: Cross-Tabulation

Ch 5.1

41. The row and column totals in a bivariate frequency table are:

b) outliers.

c) the differences between highest and lowest frequencies.

Þ a) marginals.

d) cell frequencies.

e) the dependent variable.

Ch 5.2A, 5.3A

42. If columns of a percentage table are very different from one another, then:

Þ a) the variables are related.

b) the relationship between the variables is causal.

c) the relationship is asymmetric.

d) the relationship is symmetric.

e) any of the above.

Ch 5.2B

43. In a bivariate percentage table:

a) the independent variable must be the column variable.

Þ b) percentages usually are calculated within categories of the independent

variable.

c) data must be sample data rather than population data.

d) variables are usually dichotomized.

e) the independent variable usually has more values than the dependent variable.

Ch 5.2B

44. Consider this table:

Table 2. Legalization of Marijuana

by Gender (in percentages)

Gender

Marijuana Opinion Male Female

Should Be Legal 38 30

Should Not Be Legal 62 70

Total 100 100

(N) (792) (988)

The independent variable in this table is:

a) opinion about legalization of marijuana.

b) male.

Þ c) gender.

d) should be legal.

e) none of the above.

Ch 5.3B

45. The table in the previous question shows that the relationship between marijuana opinion and gender is:

a) nonexistent

Þ b) weak.

c) strong.

d) aggregate.

e) curvilinear.

Ch 5.3C

46. Consider this table based on General Social Survey data:

Table 45. Change in Financial Situation

by Family Income (in percentages)

Financial Family Income

Situation Low Medium High

Better 29 49 66

The Same 43 37 27

Worse 28 14 7

Total 100 100 100

(N) (888) (938) (619)

The relationship between change in financial situation and family income is:

a) strong.

b) positive.

c) monotonic.

Þ d) All of the above.

e) None of the above.

Ch 5.3D

47. The larger the percentage difference(s) across categories of the independent variable:

a) the weaker the relationship between two variables.

Þ b) the stronger the relationship between two variables.

c) the less likely the researcher is to commit an ecological fallacy.

d) the more likely the relationship is positive.

e) the more likely that the independent and dependent variables' marginals are different.

Ch 5.3E, 5.4A

48. The relationship between two ordinal variables can be:

a) positive.

b) strong.

c) curvilinear.

Þ d) all of the above.

e) none of the above.

Ch 5.4B

49. If a relationship is strongly positive, we know that:

a) there is a causal relationship between the variables.

b) the column marginals are skewed.

c) the N is large.

d) there are few cases in the diagonal.

Þ e) high dependent variable scores are associated with high independent variable scores.

Ch 5.8 (throughout)

50. If there is no association between two variables in a bivariate percentage table, then:

a) the row and column marginals are the same.

b) most cases lie on the diagonal

c) the N is small.

d) all of the above.

Þ e) none of the above.

Chapter 6: The Chi-Square Test of Statistical Significance

Ch 6.1A

51. Which of the following risks a Type II error?

a) Rejecting a null hypothesis that is true.