Statistical Methods - and Interpretation, 7e Craig Mertler, Rachel Va...

Advanced and Multivariate Statistical Methods Practical Application and Interpretation, 7e Craig Mertler, Rachel Vannatta, Kristina LaVenia (Answer Key) 1 / 3

Answer Key Chapter 2

• Bivariate regression
• One-way MANOVA
• T-test of independent samples
• Bivariate correlation
• Discriminant Analysis
• One-way ANCOVA
• One-way MANCOVA
• Multiple regression
• One-way ANOVA
• Two-way MANOVA
• Logistic regression
• Two-way ANOVA
• Factor Analysis

Chapter 3 Output for Exercise 3.1.

1a. There are no missing values for either variable.1b. Peduc2 has a large split in that n=121 (77%) with a master’s degree and n=36 (23%) with a post- master’s degree. However, it is not extreme (90/10).1c. Boxplots identify two outliers (cases 38 and 40) for the post-master’s group in Humility.1d. Normality may be assumed as tests of normality are not significant. No transformation is necessary.1e. Homogeneity of variances is not assumed since Levene’s is significant (p = .021) 1f. Remove the outliers for the post-master’s group, where HumilityMean < 3.0. This may improve homoscedasticity. The authors did NOT conduct this transformation.

• / 3

Output for Exercise 3.2.

2a. There are no missing data. The box plots do not identify an outliers. However, since normality is violated for HumilityMean and MLQMean, we will examine the histograms to identify extreme values. Histograms do not reveal extreme outliers. No transformations will be made at this time.2b. Chi-square criteria for three quantitative variables (df = 3) at p = .001 is  2 = 16.266. The box plot of multivariate outliers for Mahalanobis distance identifies two values that exceed 16.266. Use select cases.2c. Scatter plot matrix indicates no curvilineary. Linearity will be assumed. Plots are fairly elliptical.Normality will be assumed.2d. Residuals plot shows not extreme clustering. Homoscedasticity will be assumed.

• / 3