SPSS Exercise 9

One-Way ANOVA

Analysis of Variance is used when you have more than two means that you want to compare such that the t -test is no longer applicable. The null hypothesis is the same as it was with two means only expanded to more than two means. It could be stated that "There are no differences between the means" or "Any differences between the means are due to chance". This would apply for, however, many means that you are considering so long as they are from different levels of one variable. Consider the example an analysis of the attitudes of students from three colleges in the eastern United States to student participation in determining college curricula. The following data are given:

In order to do the analysis, you must first recognize that it is a between groups one-way analysis of variance. This is similar to the independent t -test where the people in the groups are independent of each other. The data will be entered like the independent t -test only you have more than two groups. You will have one column for the independent variable (or factor) and you will have one column for the dependent variable.

Open SPSS, name and enter the data according to previous exercises. In the variable view, label row 1 ‘ college ’ (independent variable) and label row 2 ‘ score ’ (dependent variable). In the label column enter ‘C ollege ’ . Tab over to the ‘ Label ’ column and click on the gray box on the right portion of the cell. The ‘Value label’ window should appear. In the ‘ Value ’ window enter 1. Tab down to "Value label" and enter "College 1" (no quotes). Do the same procedure for the remaining variables (College 2 and College 3) and then click OK. Enter your data as you would for the independent t -test only when you finish with the second set of data continue with the third set.

Your " college " column will begin with 1's, then 2's, and then 3's and the score column will contain the corresponding scores.

Alternate method: It is also possible to do the analysis using the general linear model. Go back to you data spreadsheet and click on Analyze , General Linear Model and then univariate . In the dialog box, select score as the dependent variable. Then click on college and move to the fixed factor box. Click on Posth oc and in the Post h oc box, click on college and move to the " Post hoc T ests for " box. Then select Tukey , continue, OK. Your results should be the same as the first analysis.

1. Conduct a one-way ANOVA to investigate the differences between college attended and student attitudes on curriculum design
2. Create a boxplot to show the differences among the distributions for the 3 colleges
3. Print out your results and identify the following:
   1. F ratio for the college effect (between groups) _______
   2. Sums of squares for the error term (within groups) ______
   3. Mean for college 3 students _____
   4. p value for the college effect (between groups) __________
4. Write out your conclusions regarding the students’ attitudes in curriculum design
5. Refer to the Tukey posthoc test to see if there are differences in attitude between the colleges? If so which ones

SPSS Exercise 10

Repeated Measures ANOVA

A researcher is interested in how many times students laugh in a lecture class. She is interested in knowing if there is a difference in different class subjects. She randomly selects 10 students who have 4 classes on a MWF schedule. Observers record the number of times a student laughs during the class. In order to be considered a laugh a traditional laugh sound must be made accompanied by mouth and head movement. The data are as follows:

Create the file and enter the data. Click on Analyze and then click on " General Linear Model ". Next select Repeated Measures since the same subjects are used in each class.

The Repeated Measures Define Factor(s) window will open. It must now be made clear to SPSS that the subjects are not 4 different variables but 4 levels _, of the same variable. In the within-subject factor name dialog box (factor1) enter the name that you have chosen for your variable; in this case call it "content".

Next enter the number of levels of the within subject variable; in this case 4. Click on Add and then Define . In the new dialog box, you need to name each level of the variable. Click on Statistics in the left box and use the arrow to move it to the right. Notice that it is defined as level 1.

Repeat the process for each of the levels that are left. Now click on Options and then click on the box next to Descriptive statistics in order to get the descriptive statistics. Then click on the box next to Estimates of effect size to get the strength of association between the variables. Note that you are not able to run post hoc tests for within-subjects comparisons. If you wish to review the main effects click on Content in the Factor(s) and Factor Interactions window and move the variable to the Display means for window. Then click on Compare main effects while in the "options" tab. This output table will give you a variety of post hoc tests related to the problem. Click on Continue . Finally click on OK to run the analysis.

From the output identify the following information:

1. multivariate F value based on the Wilk’s lambda ________
2. p value for Wilk’s lambda _________
3. mean of the chemistry scores _______

Analyze the results of the multiple comparisons

Write a results section (what happened).

Create an error bar chart to show the distribution of the content scores (graphs, error bar, summaries for separate variables, define). Move all variables to the error bars window (except name).

SPSS Exercise 1 1 a

2 x 2 Factorial ANOVA

You know that women live longer than men do. What about right-handed and left-handed people? Coren and Halpern’s (1991) ¹ review included data on gender and handedness, as well as death. An analysis of the age-at-death numbers in the problem will produce conclusions like those reached by Coren and Halpern.

Using SPSS conduct a 2 x 2 factorial ANOVA to analyze the effect of gender and handedness on the age of men and women at death.

Open SPSS, name and enter the data. Click on the "Variable" tab and in the first cell of the first row name your first independent variable (Factor A ) "gender". Tab over to "decimals" and change the default setting to "zero". Tab over to "label" and enter "Gender". Finally, tab over to "values" and type 1 = women and 2 = men.

In the first cell of the second row, name your second independent variable (Factor B ) "handedness". Tab over to "decimals" and change the default setting to "zero". Tab over to "label" and enter "Handedness". Finally, tab over to "values" and type 1 = left-handed and 2 = right-handed. In the first cell of the third row name your dependent variable "age". Tab over to "decimals" and change the default setting to "zero". Tab over to "label" and enter "Age at Death" (Figure 1).

After entering the data, click on "Analyze", "General Linear Model" and "Univariate". When the dialog box appears, click on "age at death" and move it to the dependent rectangle. Click on "gender" and "handedness" and move them to the fixed factors rectangle. Next click on the "options" button and select "gender", "handedness", and "gender * handedness" and move them to the "display means for" box. Check the box by the words "descriptives", "estimates of effect size" and "homogeneity tests" then click continue. Click on "continue" and then OK.

Create a boxplot to visually represent the interaction between gender and handedness on age at death by clicking on graphs, boxplot, summaries for groups of cases, and define. When the window appears on your screen move the "handedness" variable to the "category axis" and the "age at death" variable to the variable rectangle. Click OK.

Identify the independent variables (factors) and dependent variable.

Factor A _________________________________________

Factor B ________ _________________________________

Dependent Variable ________________________________

Identify the design using R x C notation_____________

Identify the following information from the output:

1. F value for the gender main effect ________
2. Mean age at death for left-handed men ________
3. Mean age for right-handed women _______
4. Partial Eta Squared for the interaction effect ________
5. p value for the handedness main effect ________

Briefly discuss your findings. Is there a significant difference between gender and handedness on age at death? (yes/no is not a complete answer)

SPSS Exercise 1 1 b

2 X 3 Factorial ANOVA ²

Clinical depression occurs in approximately 15 percent of the population. Many treatments are available. The data that follow are based on Hollon, Thrase, and Markowitz’s (2002) comparison of: (1) psychodynamic therapy , which uses free association and dream analysis to explore unconscious conflicts from childhood, (2) interpersonal therapy , which progresses through a three-stage treatment that alters the patient’s response to recent life events, and (3) cognitive-behavioral therapy , which focuses on changing the client’s thought and behavior patterns. The second factor in this data set is gender. The numbers are the improvement scores for the 36 individuals receiving therapy.

Using SPSS conduct a 2 x 3 factorial ANOVA to analyze the effect of gender on the type of behavioral therapy provided to an individual.

Click on the "data" tab and begin entering your data (Figure 2). For SPSS to be able to compute a factorial ANOVA, the data need to be entered in a specific way. Make sure that you enter the data in the proper sequence. If you look at the data in the table above you will notice that each cell represents the intersection of Factor A and Factor B for each level of the factor. For example cell A ₁ B ₁ contains the scores 22, 42, 30, 49, 15, & 34. You need to make sure that the data are entered into SPSS so that the data are matched up with the two Factors (see the figure below for more info). After the data are entered, click on "Analyze", "General Linear Model" and "Univariate". When the dialog box appears, click on "improve_score" and move it to the dependent rectangle. Click on "type of therapy" and "gender" and move them to the fixed factors rectangle. Next click on the "options" button and select "therapy", "gender", and "therapy * gender" and move them to the "display means for" box. Check the box by the words "descriptives", "estimates of effect size" and "homogeneity tests" then click continue. Now Click on the "Posthoc" box, select the "therapy" variable and move it to the "posthoc tests for" box. Place a check in the box next to the "Tukey" test. Click on "continue" and then OK.

Create a boxplot to visually represent the interaction between gender and type of therapy provided to an individual by clicking on graphs, boxplot, summaries for groups of cases, and define. When the window appears on your screen move the "therapy" variable to the "category axis" and the "improve_score" variable to the "variable" rectangle. Click OK.

Identify the independent variables (factors) and dependent variable.

Factor A _________________________

Factor B ____________ _____________

Dependent Variable _________________

Identify the design using R x C notation __________________

1. SS _tot ___________ SS _cells _______ SS _therapy ________
   SS _gender ____ SS _AB ________ SS _error _________

What follow-up tests should be conducted? _____ __________________

Why?

Briefly discuss your findings. Is there a significant difference between gender and the type of therapy a person receives? (yes/no is not a complete answer)

Price: $45.6

Solution: The downloadable solution consists of 20 pages, 2560 words and 2 charts.
Deliverable: Word Document