Computer Exercise: ANOVA
Question: Do person’s marital status differ significantly in their positive attitudes toward health outcomes?
1. Each student is assigned to a different SPSS dataset from Book 5 for the assignments (MINE IS 380-680). For information about the Book 5 dataset, please go to Learning SPSS–Inventory of Personal Attitude (IPA) and read the description. Download the document, “Individual Dataset Assignment” attached on this assignment.
Download Book 5 (attached on this assignment) and read the document, How to Create Individual Dataset, for setting up the dataset for your specific subjects. This document is also posted in the folder, “Learning SPSS” in the left side menu column. Once you create your dataset, save it in your convenient place. You will use this dataset for all the computer assignments (except correlation) throughout the semester.
Click for more options
2. Open your dataset (380-680).
First, create a copy of the dictionary so you will have a description of the variables and their value labels. Do this by clicking on File—-Display Data File Information——Working File. Print out (or save) the results. Review all the variables in the dataset and get familiar with them.
3. Recode marital (marital status) variable.
The variable, “marital”, has 6 groups—“Never Married”, “Married”, “Living with Significant Other”, “Separated”, “Widowed”, and “Divorced”. To make your analysis simple, recode into three groups: married, never married, and living alone. The groups of “Married” and “Living with Significant Other” will be recoded as “Married”, and the groups of “Separated”, “Widowed”, and “Divorced” will be recoded as ”Living alone”. Read the attached document, How to Recode into Different Variables.
•“Never Married”—>“Never Married”
•“Married”, “Living with Significant Other”—-> “Married”
•“Separated”, “Widowed”, and “Divorced” —-> “Living alone”.
4. Run frequencies on the two variables, recoded marital and total (=total IPA score). Read the attached document, How to Run Frequencies.
Inspect your SPSS output and review the variables for invalid values. Invalid (or incorrect) values are values outside of your expectations. For example, if you have an age value of 4, this is an “invalid value” because 4 year old cannot participate in the survey. If you see a value of 2 in a gender variable, this is also invalid value as the gender variable has only 2 values-0 for male and 1 for female. Invalid values should be removed (deleted) from the dataset. Review all the variables for invalid values thoroughly. We call this procedure, “Data Cleaning”. This is an important step to do before you begin your data analysis. Describe your experience with data cleaning briefly—did you see any invalid values?
5. Check the assumptions of ANOVA.
•Assumption of independence
•Assumption of homogeneity of variance
•Assumption of normality–examine the skewness of the dependent variable (total IPA score). For this assignment, use Pearson’s measure (see p. 9 in Descriptive statistics lecture notes). To calculate a Pearson’s skewness coefficient, do a simple hand calculation using the formula;
Skewness = (mean-median)/SD
According to the Hildebrand guideline:
1.Pearson’s skewness coefficient ≤ .1 —minor skewness
2.Pearson’s skewness coefficient ≤ .2 —moderate skewness
3.Pearson’s skewness coefficient > .2 —severe skewness
If total IPA score is considered skewed (moderate or sever skewness), try data transformation of the variable. read the document, How to Recode Variables for Data Transformation: Square Root and Log Base 10.
6. Run one-way ANOVA. Read the document, How to Run ANOVA.
7. Interpret SPSS output and write up your findings. Your report includes:
•Results of assumption check
•Your answer to the question. In your paper, you should cite the value of test statistics, probability levels, and any other values necessary to determine what the results mean. A sample answer to the question might be: The results of the ANOVA are significant (F (between group degree of freedom, within group degree of freedom) = XX, p < .05). Therefore, we reject the null hypothesis. The means on personal positive attitude scores are different among three groups. The posthoc results show that there is a mean difference between the group A and B (p =YY). Group A (mean total IPA score =ZZ) has more positive attitudes than group B (mean total IPA score = AA). There are no significant mean differences between the group B and C, and A and C. Please note. If your ANOVA results indicate that three groups are not significantly different, DO NOT report posthoc results. Posthoc results are only reported when ANOVA results indicate three groups are different.
•ANOVA summary table—use the following template.
Source of variance ss df Ms f p
Between groups
Within groups
Total
_____________________________________________________________________________________SS= Sum of square ; Ms= mean square
Click for more options
6. Submit the followings:
•A text file with one table
•A SPSS output file (.spv)
