Question 1

There are four possible outcomes of a researcher's decision to accept or reject the null hypothesis. If a null hypothesis is accepted and it turns out to be the wrong decision, what kind of error has occurred?
A)Type I error
B)Type II error
C)Type III error
D)Power error

Accepted Answer

A type II error occurs when null hypothesis is accepted and that turns out to be the wrong decision. In this error, the specific experiment found no statistically significant results, but future experiments do obtain significant results.

Question 2

There are four possible outcomes of a researcher's decision to accept or reject the null hypothesis. A correct decision wherein the null hypothesis is rejected is called power. Which of the following values does power have an inverse relationship with?
A)Alpha
B)Beta
C)Standard error of the mean
D)z-score

Accepted Answer

Power has an inverse relationship with the beta value (specifically, power=1-beta). Beta corresponds to the probability of making a type II error and is different for every experiment.

Question 3

Z-scores are a type of standard score (scores based on the standard deviation of the sample). Why is it useful to transform raw scores into z-scores?
A)To normalize distribution of a sample
B)To determine population parameters
C)Because they correspond to percentile ranks
D)Because they can be used to calculate standard error of the mean

Accepted Answer

Z-scores are useful because they correspond to percentile ranks and because they are relatively easy to calculate from raw scores. However, they only correspond to percentile ranks in normal distributions and can only be defined when population parameters are known.

Question 4

Choosing the most appropriate statistical test depends on what type of question the research is addressing. In the study with post-cardiac patients, where a number of measure of cardiac and physical functioning were related to the participation in one of many psychological treatment models, what is the best test choice?
A)ANOVA
B)Chi-square
C)MANOVA
D)t-test

Accepted Answer

MANOVA is the best test choice in this study. Similar to ANOVA, which is best used in research where data are collected for more than two groups (in this case, groups for each different psychological treatments), MANOVAs are useful when there is more than one dependent variable.

Question 5

Choosing the most appropriate statistical test depends on what type of question the research is addressing and what type of data is collected. ANOVAs and MANOVAs are both useful in analyzing interval and ratio data. Using multiple ANOVAs instead of one MANOVA where there is more than one dependent variable has what effect on results?
A)Increases the likelihood of a type I error
B)Increases the likelihood of a type II error
C)Decreases the likelihood of a type I error
D)Decreases the likelihood of a type II error

Accepted Answer

Running an ANOVA on each dependent variable, instead of one MANOVA on all data, increases the probability of a type I error (that is, of mistakenly rejecting the null hypothesis).

Question 6

Choosing the most appropriate statistical test depends on what type of question the research is addressing and what type of data is collected. Which of the following studies would you least likely use a chi-square test to analyze?
A)Percentage of subjects who have continued or discontinued substance use after a substance abuse treatment program as compared to pre-treatment baseline behavior.
B)Proportion of male and female voters who describe themselves as liberal, moderate, or conservative
C)Number of high school students who admitted to anorexic eating behavior as compared to the number of college freshman who admitted to anorexic eating behavior.
D)Frequency of passing the EPPP in a group of males and females by ethnicity

Accepted Answer

A chi-square test is not a good choice to analyze the percentage of subjects who have continued or discontinued substance use after a substance abuse treatment program as compared to pre-treatment baseline behavior. Chi-square tests are used to analyze nominal data for one or two independent variables.

Question 7

Two students take a test. Student F scores in the 92^nd percentile and student K scores in the 47^th percentile. If three points are added to each student's raw score, whose percentile will be more impacted? A)Student K B)Student F C)Both students K and F D)Neither student

Accepted Answer

Student K's percentile rank will be more greatly impacted by the addition of 3 points to the raw score. Percentile rank varies much more in the middle of a distribution than at either end (smaller fluctuations in raw score result in bigger fluctuations in percentile).

Question 8

Choosing the most appropriate statistical test depends on what type of question the research is addressing and what type of data is collected. Which of the following factors does NOT play a significant role in determining choice of statistical test?
A)Number of dependent variables
B)Number of independent variables
C)Group correlation
D)Sample size

Accepted Answer

Sample size is not one of the main factors considered when choosing statistical tests. Sample size can impact results by causing skewed distributions, but that is addressed by including a sufficient number of subjects in the group. Factors that are considered: number of dependent and independent variables, levels of independent variables, type of data, and group correlation or independence.

Question 9

A study collects information from 50 people about their ethnicity, where options are Caucasian, Hispanic, Native American, Asian, and African; and their religious views, where options are agnostic, Christian, atheist, and other. How many degrees of freedom are there?
A)9
B)12
C)20
D)3

Accepted Answer

There are 12 degrees of freedom. A multiple sample chi-square is used to analyze this data, since there are two independent variables. The formula for degrees of freedom in a multiple sample chi-square is df=(#rows-1)x(#columns-1). There are 5 rows for ethnicity and 4 rows for religious views. So, degrees of freedom is 5 minus 1 (4) times 4 minus 1 (3).

Question 10

When analyzing a one-way ANOVA, what value is used to determine significance?
A)Standard error of the mean
B)F ratio
C)t-scores
D)Percentiles

Accepted Answer

F ratio is the number used to determine significance in a one-way ANOVA. F ratio is calculated by dividing the mean square between groups by the mean square within groups.

Question 11

In statistical analysis, what is the Kuder-Richardson 20 used to analyze?
A)Potential discrimination in test construction
B)Internal reliability of test
C)Standard error of measurement
D)Factors that might affect validity

Accepted Answer

The Kuder-Richardson 20 is a measure of internal reliability, specifically of internal consistency, for one test. It is used in tests containing measures with dichotomous choices.

Question 12

In statistical analysis, a z-score is a standard score used to evaluate the standard deviation of factors. What information would you use to calculate a z-score?
A)Raw score and mean of the population
B)Raw score and standard deviation of the mean of the population
C)Raw score, standard deviation of the mean of the population, and mean of the population

Accepted Answer

A z-score is calculated using raw score, standard deviation of the mean of the population, and mean of the population. For instance, if the raw score is 130, mean of the population is 100, and standard deviation is 15, z-score is calculated by subtracting the population mean from the raw score (30) and dividing the result by the standard deviation for a z-score of 2. That score indicates 2 standard deviations about the mean for the sample.

Question 13

In statistical analysis, when would trend analysis likely be used?
A)When there are two qualitative independent variables
B)When there is one dependent variable and one independent variable
C)When there is a large number of samples to compare
D)When there is one quantitative independent variable

Accepted Answer

Trend analysis is useful when there is a quantitative independent variable that yields a non-linear outcome. Differences between groups are less important than the trend of the data for this type of analysis.

Question 14

Tests of relationship and predication are divided into bivariate and multivariate tests. Bivariate correlations look at the relationship of two variables-one is a predictor and the other is the criterion. What does the correlation coefficient describe?
A)The effect of the predictor variable on the criterion variable
B)The effect of the predictor variable on the criterion variable in terms of strength and direction
C)The relationship between the predictor variable and the criterion variable in terms of strength and direction
D)The relationship between the predictor variable and the criterion variable in terms of absolute value

Accepted Answer

The correlation coefficient describes the relationship between the predictor variable and the criterion variable in terms of both strength and direction. Correlation cannot be used to determine causal relationships; it is used to determine relationships between variables when there is no other obvious explanation.

Question 15

One of the values used in bivariate correlation tests is the coefficient of determination. If there is a correlation coefficient of .50 between education and income, what is the coefficient of determination?
A).50
B).25
C).07
D)75%

Accepted Answer

The coefficient of determination is the square of the correlation coefficient. Squares of decimals are always smaller than the original value; 5 squared is 25, and the decimal point remains in the same place. This number suggests that 25% (the equivalent percentage) of variability in income is explained by education.

Question 16

On a correlational graph of a bivariate test, a lot of scatter around the regression line implies that there is:
A)A low correlation coefficient
B)A negative correlation coefficient
C)A curvilinear relationship
D)A positive correlation coefficient

Accepted Answer

Scatter on a correlational graph, rather than cluster around a regression line, suggests that there is a low correlation between two variables and therefore, a low correlational coefficient.

Question 17

You have conducted a study to test the relationship between personality and sex appeal in which there are two indicators for each latent variable (personality and sex appeal). When analyzing the results of this study, you might want to use LISREL in order to:
A)Assess a causal model involving observed variables that are believed to operate unidirectionally
B)Assess a causal model involving observed and latent variables that are believed to operate unidirectionally
C)Assess a causal model involving observed variables that are believed to operate bidirectionally
D)Assess a causal model involving observed and latent variables that are believed to operate bidirectionally

Accepted Answer

LISREL (linear structural relationships) is a computer program designed to analyze relationships in each direction (bidirectional rather than unidirectional) between latent variables and between latent variables and their indicators.

Question 18

Choosing what type of bivariate correlation coefficient to use in statistical analysis depends on what type of data variable X and variable Y are. You are analyzing the results of a survey about satisfaction with current political happenings in which subjects answered questions on a scale of 0-5, 0 being not at all satisfied and 5 being completely satisfied. They also answered questions about amount of time spent reading or watching news on a similar scale. What bivariate correlation coefficient might you use?
A)Pearson r
B)Spearman's Rho
C)Phi
D)Tetrachoric

Accepted Answer

Spearman's Rho is the best choice for this study because both variables are ordinal-they are discrete values that can be ranked. Kendall's Tau is another coefficient used with this type of data.

Question 19

Choosing what type of bivariate correlation coefficient to use in statistical analysis depends on what type of data variable X and variable Y are. In a study where X is gender and Y is weight, what correlation coefficient might you use?
A)Pearson r
B)Kendall's Tau
C)Point-Biserial
D)Phi

Accepted Answer

Point-Biserial correlation coefficient is the best choice for analysis of this data. X is a true dichotomy variable (there are two possible answers), and Y yields ratio data (numerical values with an absolute zero); this correlation coefficient is indicated when one variable is a true dichotomy and one is either interval or ratio data.

Question 20

Multiple regression tests are used when there are several predictors (or independent variables) and one or more criterion (dependent variables). What does multicollinearity refer to?
A)Predictors are highly correlated, and therefore redundant
B)Predictors are too lowly correlated to be useful
C)Criterion are highly correlated, and therefore redundant
D)There is a high correlation between each predictor and criterion

Accepted Answer

Multicollinearity is a problem that can occur in multiple regression tests when predictors are highly correlated and therefore redundant.

Quiz 6: Research Methods and Statistics

There are four possible outcomes of a researcher's decision to accept or reject the null hypothesis. If a null hypothesis is accepted and it turns out to be the wrong decision, what kind of error has occurred?

There are four possible outcomes of a researcher's decision to accept or reject the null hypothesis. A correct decision wherein the null hypothesis is rejected is called power. Which of the following values does power have an inverse relationship with?

Z-scores are a type of standard score (scores based on the standard deviation of the sample) . Why is it useful to transform raw scores into z-scores?

Choosing the most appropriate statistical test depends on what type of question the research is addressing and what type of data is collected. Which of the following studies would you least likely use a chi-square test to analyze?

Two students take a test. Student F scores in the 92^nd percentile and student K scores in the 47^th percentile. If three points are added to each student's raw score, whose percentile will be more impacted?

Choosing the most appropriate statistical test depends on what type of question the research is addressing and what type of data is collected. Which of the following factors does NOT play a significant role in determining choice of statistical test?

A study collects information from 50 people about their ethnicity, where options are Caucasian, Hispanic, Native American, Asian, and African; and their religious views, where options are agnostic, Christian, atheist, and other. How many degrees of freedom are there?

When analyzing a one-way ANOVA, what value is used to determine significance?

In statistical analysis, what is the Kuder-Richardson 20 used to analyze?

In statistical analysis, a z-score is a standard score used to evaluate the standard deviation of factors. What information would you use to calculate a z-score?

In statistical analysis, when would trend analysis likely be used?

Tests of relationship and predication are divided into bivariate and multivariate tests. Bivariate correlations look at the relationship of two variables-one is a predictor and the other is the criterion. What does the correlation coefficient describe?

One of the values used in bivariate correlation tests is the coefficient of determination. If there is a correlation coefficient of .50 between education and income, what is the coefficient of determination?

On a correlational graph of a bivariate test, a lot of scatter around the regression line implies that there is:

You have conducted a study to test the relationship between personality and sex appeal in which there are two indicators for each latent variable (personality and sex appeal) . When analyzing the results of this study, you might want to use LISREL in order to:

Choosing what type of bivariate correlation coefficient to use in statistical analysis depends on what type of data variable X and variable Y are. In a study where X is gender and Y is weight, what correlation coefficient might you use?

Multiple regression tests are used when there are several predictors (or independent variables) and one or more criterion (dependent variables) . What does multicollinearity refer to?