Analysis of Variance ANOVA COMM 420 8 Fall

  • Slides: 35
Download presentation
Analysis of Variance ANOVA COMM 420. 8 Fall, 2008 Nan Yu 1

Analysis of Variance ANOVA COMM 420. 8 Fall, 2008 Nan Yu 1

Data Collect for your group project On-line questionnaires? Paper questionnaires? What you need to

Data Collect for your group project On-line questionnaires? Paper questionnaires? What you need to do… 2

Warming Up… l Download the following files from ANGEL (folder: week 11). ¡ review

Warming Up… l Download the following files from ANGEL (folder: week 11). ¡ review practice answers ¡ ANOVAdata. sav l Please complete the questions in “review practice. ” l Compare your answers to “review practice answers. ” 3

T-test Review 4

T-test Review 4

One-Way ANOVA 5

One-Way ANOVA 5

Analysis of Variance (ANOVA) One-Way ANOVA - ANOVA is used to test differences among

Analysis of Variance (ANOVA) One-Way ANOVA - ANOVA is used to test differences among more than two groups. - Independent variables must be nominal, more than 2 levels. - Dependent variables must be interval or ratio-level and measured on the same metric. 6

One-Way ANOVA 7

One-Way ANOVA 7

8

8

Assumptions of ANOVA 9

Assumptions of ANOVA 9

10

10

11

11

Hypothesis (ANOVAdata. sav) Attitude toward the ad will vary as a function of the

Hypothesis (ANOVAdata. sav) Attitude toward the ad will vary as a function of the type of ads that are featured on the web site. IV? DV? How many levels (groups) are in the IV? 12

SPSS and ANOVA 13 GSS 93 Subset. sav

SPSS and ANOVA 13 GSS 93 Subset. sav

Put the interval or ratio-level variable here. (DV) Put the variable representing the groups

Put the interval or ratio-level variable here. (DV) Put the variable representing the groups here. (IV) Click "Options" 14

Put the group variable in the right window. Then click "continue" and "Ok. "

Put the group variable in the right window. Then click "continue" and "Ok. " Select Compare main effects, Descriptive Statistics, Estimates of effect size, Homogeneity tests, Bonferroni 15

SPSS Output Each group has 20 people 16

SPSS Output Each group has 20 people 16

Means and standard deviations for each group 17

Means and standard deviations for each group 17

Homogeneity of Variance is good! Meaning that variances between groups are not significantly different.

Homogeneity of Variance is good! Meaning that variances between groups are not significantly different. 18

Effect size These are degrees of freedom. This is the obtained p-value. This is

Effect size These are degrees of freedom. This is the obtained p-value. This is the test statistic. F(3, 76) = 21. 74, p <. 001. Reject the null 19

This tells you the group of Animated Ad is different than the other three

This tells you the group of Animated Ad is different than the other three groups. Bonferroni post hoc comparisons. 20

Interpret One-Way ANOVA with Graphs F(3, 76) = 21. 74, p <. 001, partial

Interpret One-Way ANOVA with Graphs F(3, 76) = 21. 74, p <. 001, partial η 2 =. 46 21

Interpret One-Way ANOVA with tables The results shows that participants reported significantly more favorable

Interpret One-Way ANOVA with tables The results shows that participants reported significantly more favorable attitudes toward the ad in the animated condition (M = 5. 33, SD = 0. 92) compared to the text condition (M = 3. 59, SD = 0. 92), static condition (M = 3. 53, SD = 0. 68), or the pop-up condition (M = 3. 43, SD = 0. 93), F (3, 76) = 21. 74, p <. 001, partial η 2 =. 46. 22

Factorial Designs More than one IV, both are nominal, DV is interval or ratio-level

Factorial Designs More than one IV, both are nominal, DV is interval or ratio-level Behavioral intention will vary as a function of the type of ads that are featured on the web site and the gender of the participants. IVs? DV? 23

Put the interval or ratio-level variable here. (DV) Put the variables representing the groups

Put the interval or ratio-level variable here. (DV) Put the variables representing the groups here. (IVs) Then Click "Options" 24

Put the group variable in the right window. Then click "continue" and "Ok. "

Put the group variable in the right window. Then click "continue" and "Ok. " Select Compare main effects, Descriptive Statistics, Estimates of effect size, Homogeneity tests. 25

Main Effects Interaction Degrees of freedom P-values Test statistics Effect size Significant main effect

Main Effects Interaction Degrees of freedom P-values Test statistics Effect size Significant main effect for Gender, F (1, 72) = 31. 55, <. 001, partial η 2 =. 31 No significant Main Effect for Condition, F (3, 72) =. 74, p =. 53, partial η 2 =. 03 26 No significant Gender X Condition interaction, F (3, 72) = 1. 19, p =. 32, partial η 2 =. 05

Creating Graphs Highlight these two numbers with mouse, right click Create Graph Bar 27

Creating Graphs Highlight these two numbers with mouse, right click Create Graph Bar 27

Significant main effect for Gender, F (1, 72) = 31. 55, <. 001, partial

Significant main effect for Gender, F (1, 72) = 31. 55, <. 001, partial η 2 =. 31 28

No significant main effect for Condition, F (3, 72) =. 74, p =. 53,

No significant main effect for Condition, F (3, 72) =. 74, p =. 53, partial η 2 =. 03 29

How to Create Line Graphs for Interaction Effects? Double click this table, then right

How to Create Line Graphs for Interaction Effects? Double click this table, then right click Select Pivoting Trays 30

Move the small square from bottom to the left Then, the means table will

Move the small square from bottom to the left Then, the means table will look at this. 31

Highlight these numbers with mouse, then right click Create Graph Line 32

Highlight these numbers with mouse, then right click Create Graph Line 32

Graphs for the Interaction Effects No significant Gender X Condition interaction, F (3, 72)

Graphs for the Interaction Effects No significant Gender X Condition interaction, F (3, 72) = 1. 19, p =. 32, partial η 2 =. 05 These lines are not parallel, we can suspect that there might be interaction effects. But they are not statistically significant. 33

In-class practice (ANOVAdata. sav) Attitude toward the brand will vary as a function of

In-class practice (ANOVAdata. sav) Attitude toward the brand will vary as a function of the type of ads that are featured on the web site. -Can we reject the null? -Please report the test statistics and create a bar chart for this result. -Which condition seems to yield least favorable attitude toward the brand? 34

In-class practice l Can we reject the null? Yes l Please report the test

In-class practice l Can we reject the null? Yes l Please report the test statistics and create a bar chart for this result. F (3, 72) = 31. 45, p <. 001, partial η 2 =. 56 l Which condition seems to yield least favorable attitude toward the brand? Text-only 35