Deck 15: Data Analysis II-- Descriptive and Inferential Statistics

Question 1

Which of the following is (are)among the special properties of the data that influence the choosing of appropriate statistics

A)level of measurement
B)the shape of the data distribution
C)whether observations are made independently of one another
D)all of the answers represent such special properties

Answer 1

D

Question 2

A more common term for measures of central tendency is:

A)correlators.
B)averages.
C)coefficients.
D)variances.

Answer 2

B

Question 3

The function of descriptive statistics is to:

A)allow generalizations from sample data to populations.
B)aid in organizing, summarizing, and interpreting sample data.
C)infer sample data from entire populations.
D)avoid the expense of coding the data.

Answer 3

B

Question 4

A researcher has nominal data in a 2 x 2 table and wishes to apply a measure of association. According to the text, which of the following would be the best choice

A)Goodman and Kruskal's tau
B)lambda
C)phi
D)Somer's D

Question 5

Of the following measures of association for ordinal data, a positive value for this statistic allows the researcher to conclude only that "as X increases, Y does not decrease" rather than that "as X increases, Y also increases."

A)Somer's D
B)Spearman's rho
C)Kendall's tau
D)gamma

Question 6

One of the weaknesses of the mean as a measure of central tendency is that:

A)there can be more than one mean in a distribution.
B)it produces a misleading result when a distribution is symmetrical.
C)it is less stable than the median.
D)it does not take into account the actual value of all scores in a distribution.

Question 7

What is the difference between fully ordered and partially ordered data

A)Fully ordered data has numerous ties and partially ordered data does not.
B)The categories of fully ordered data are naturally occurring whereas the categories of partially ordered data are created by the researcher.
C)Partially ordered data has numerous ties and fully ordered data does not.
D)Only fully ordered data is considered ordinal level.

Question 8

Measures of dispersion indicate:

A)the strength of relationships.
B)the average value in a distribution.
C)the amount of spread of a distribution.
D)the amount of skewness of a distribution.

Question 9

Which of the following is NOT one of the considerations described in the text that are important in choosing appropriate statistics

A)goals of the data analysis
B)the size of the population
C)number of variables
D)the audience

Question 10

According to the text, the primary factors involved in selecting a measure of central tendency are:

A)level of measurement and skewness.
B)level of measurement and audience.
C)skewness and audience.
D)level of measurement and alpha level.

Question 11

A researcher applies measures of association to five sets of variables. Which of the following indicates the strongest relationship

A) .36
B).05
C) .76
D).63

Question 12

Of the following kinds of statistics, which one does the text suggest is most likely to be beyond the comprehension of audiences without some prior knowledge of statistics

A)graphical forms of presenting data
B)univariate statistics
C)percentages
D)multivariate statistics

Question 13

The function of inferential statistics is to:

A)do exploratory analysis.
B)make findings look impressive.
C)aid in organizing, summarizing, and interpreting sample data.
D)allow generalizations from sample data to populations.

Question 14

Measures of association are used to indicate:

A)the average score of a set of data.
B)the strength of a relationship between two variables.
C)the amount of dispersion in a set of scores.
D)whether findings from a sample can be generalized to a population.

Question 15

The category in a distribution that contains a larger number of cases than any other category is called the:

A)mean.
B)median.
C)mode.
D)standard deviation.

Question 16

If a statistic is designed for nominal level of measurement, then it would be appropriate to use that statistic on data at all these levels of measurement EXCEPT:

A)ordinal.
B)interval.
C)ratio.
D)all of the answers are levels of measurement upon which it would be appropriate to use a statistic designed for nominal level.

Question 17

Which of the following is NOT a measure of dispersion

A)range
B)semi-interquartile range
C)gamma
D)standard deviation

Question 18

When choosing statistics and considering the level of measurement, one may:

A)apply any statistic to any level of measurement.
B)apply a statistic designed for a higher level of measurement to lower level data.
C)apply a statistic designed for a lower level of measurement to higher level data.
D)none of the answers is correct.

Question 19

The semi-interquartile range indicates the deviation of a set of scores from the:

A)mode.
B)median.
C)mean.
D)none of the answers is correct.

Question 20

Several measures of association can be interpreted as indicating proportional reduction in error. What does this mean

A)the amount error is reduced by eliminating missing cases
B)how much knowing the value of the independent variable helps to reduce error in predicting values of the dependent variable
C)the amount measurement error is reduced by applying statistical controls
D)how much knowing the value of the dependent variable helps to reduce error in predicting values of the independent variable

Question 21

In inferential statistics, type II or beta error is:

A)a common misinterpretation of the statistic beta.
B)rejecting null hypotheses that are actually true.
C)failing to reject null hypotheses that are actually false.
D)increased by using alpha levels of .20 or greater.

Question 22

A sampling distribution is:

A)the distribution of scores from a single sample.
B)used to test the null hypothesis.
C)a theoretical distribution of statistics from an infinite number of samples.
D)used to test the null hypothesis, and a theoretical distribution of statistics from an infinite number of samples.

Question 23

Which of the following is true of a normal distribution

A)It is a symmetrical distribution.
B)It is a unimodal distribution.
C)The mean and median of the distribution are identical.
D)All of the answers are true of normal distributions.

Question 24

The statistical procedure called multiple regression:

A)is another name for analysis of variance.
B)uses one independent variable to predict values of more than one dependent variable.
C)uses more than one independent variable to predict values of one dependent variable.
D)generates a value called lambda.

Question 25

Sampling distributions are important for:

A)descriptive statistics.
B)inferential statistics.
C)measures of association.
D)all of the answers are correct.

Question 26

In computing chi square, expected frequencies are:

A)the distribution of cases if the two variables were not related to one another.
B)frequencies predicted by our research hypothesis.
C)the actual distribution of cases from the sample.
D)the distribution of cases if the two variables were perfectly associated.

Question 27

Inferential statistics are based on:

A)intuition.
B)the Pythagorean theorem.
C)the theory of differential association.
D)probability theory.

Question 28

According to Research in Practice 15.1, statistical power in data analysis can be improved by:

A)using a larger sample.
B)changing the alpha level, say from .05 to .001.
C)increasing the size of the effect of the independent variable.
D)using a larger sample, and increasing the size of the effect of the independent variable would improve statistical power.

Question 29

The central limit theorem tells us about the properties of:

A)sampling distributions.
B)population distributions.
C)population parameters.
D)contingency tables.

Question 30

In statistical hypothesis testing, a Type II error is the same as:

A)beta error.
B)alpha error.
C)systematic error.
D)random error.

Question 31

The central limit theorem tells us that:

A)sampling distributions based on small samples approximate a normal distribution.
B)the chi-square statistic can be used only with ratio-level variables.
C)sampling distributions based on large samples approximate a normal distribution.
D)inferential statistics are almost always preferred to descriptive statistics.

Question 32

A chi-square statistic indicates:

A)whether a relationship exists between two variables in a population.
B)the strength of the relationship between two variables in a population.
C)the direction of the relationship between two variables in a population.
D)both the strength and direction of the relationship between two variables in a population..

Question 33

The statistical procedure called chi square:

A)is a measure of association for nominal data.
B)is a descriptive statistic for ordinal data.
C)is a descriptive statistic for interval data.
D)is an inferential statistic for nominal data.

Question 34

In inferential statistics, type I or alpha error is:

A)failing to reject a null hypothesis that is actually false.
B)a common error in calculating the correlation coefficient.
C)increased by using a very stringent alpha level such as .001.
D)none of the answers are correct.

Question 35

A scattergram is:

A)an ungrouped frequency distribution of one variable.
B)a preliminary look at the data before it is organized.
C)another name for a histogram.
D)a table with the scores of two variables plotted on it.

Question 36

If the null hypothesis is true and we set the alpha level at .05, then we will make type I errors:

A)5 times out of 100.
B)1 time out of 100.
C)5 times out of 10.
D)it can't be determined from this information.

Question 37

Indicating that a given result is "statistically significant" means:

A)we have obtained an important finding.
B)the null hypothesis was rejected.
C)the findings support our theory.
D)both type I and type II errors were avoided.

Question 38

The mean of a standard normal distribution will always be equal to:

A)1.
B)0.
C)10.
D)This cannot be determined without seeing the actual data in the distribution.

Question 39

Probability theory tells us:

A)the likelihood with which certain events or outcomes will occur.
B)which statistics are appropriate to use with which levels of measurement.
C)the strength and direction of the relationship between two variables.
D)whether a raw data distribution is normal or not.

Question 40

According to the conventions discussed in the text, social researchers would be least inclined to reject a null hypothesis if a test showed which of the following alpha levels

A).10
B).05
C).001
D)the alpha levels specified in all of the answers would lead to rejection of the null hypothesis.

Question 41

Define and clearly distinguish between inferential and descriptive statistics.Give an example of when each might be appropriately used.

Question 42

Describe what probability theory is and what sampling distributions are. In what ways are they important in statistics

Question 43

Identify the level of measurement and the type of research question for which these inferential statistical procedures might be used: chi square, ANOVA, multiple regression.

Question 44

Identify and describe three measures of central tendency and three measures of dispersion.Indicate what considerations are made regarding their appropriate use.

Question 45

Define the normal distribution; list its properties, and explain its significance in statistics.

Question 46

For nominal, ordinal, and interval levels of data, identify and briefly describe an appropriate measure of association.

Question 47

Explain the steps in statistical hypothesis testing. Include the concepts of null hypothesis, research hypothesis, alpha level, and beta level in your answer.

Question 48

What is meant by the concept of statistical power What can be done to improve the statistical power of a research project

Question 49

Why review the scattergram when using the correlation coefficient (Pearson's r)

Question 50

Describe considerations for choosing statistics appropriate to a particular set of data.