Statistics Midterm

Using SPSS

SPSS is a powerful statistical package which is widely used. The basics are easy to learn, and it is capable of almost any type of statistical analysis. I have provided a set of data to practice with. You will be asked to use SPSS to analyze this data to answer some of the quiz questions and to complete your final paper.

The data were collected by me on the Internet using a form at http://www.moneyworkbook.com/finpers.htm

I “cleaned up” the file to make it easier to analyze, eliminating responses with missing or otherwise invalid data. The file contains the responses of 200 men and 200 women in the United States, age 21 or older. All of these participants reported that they were in a romantic relationship with someone of the opposite sex.

The participants reported occupation, income, age, and educational level for themselves and their partners. They also reported their zip code, whether they were financially comfortable or struggling, how happy they were with their current relationship, and the nature of their current relationship: dating several people, dating one person, long-term committed, or married.

Then they took a test of financial personality with three scales: Lifestyle, Dependency, and Risk-Taking. High scores on the Lifestyle scale indicate a desire for luxury. High scores on the Dependency scale indicate a tendency to expect financial support from one’s partner and others. High Risk-Taking scores indicate a tendency to take financial risks, both in careers and in investments.

1) Install SPSS to your computer.

2) Download the SPSS file, DATA540, from Resources and save it on your hard disk.

3) Start SPSS. Load in the data file by clicking FILE >OPEN>DATA and finding the data file.

4) You will see a spreadsheet, similar to Excel. Each row is a “case” and represents one participant. Each column is a variable, such as gender, age, or other pieces of information.

5) At the bottom of the page, you will see tabs that say “Data View” and “Variable View.” In Data View, each row shows data for an individual participant. In Variable View, each row shows information about a variable.

6) In “Data View” click VIEW>VALUE LABELS. This will toggle the way the data is seen. With Value Labels turned on you will see the label for different values of a variable. With Value Labels turned off you will see the numbers used to code that variable. For instance, FCOMFORT (Financial Comfort) appears as 1’s and 2’s with Value Labels off, it appears as “Comfortable” and “Struggling” with Value Labels on. I prefer to keep Value Labels on.

7) Click the “Variable View” tab at the bottom of the screen. You will see information about each variable. Here is the most important information:

NAME: This is the brief name for the variable.

LABEL: This is the longer, descriptive name.

VALUES: These are value labels. Notice that some variables, such as age, don’t have value labels. Other variables, such as EDUC1, do have value labels. Clicking on this field will show you what each value of the variable represents.

Analyses

Here are the steps for doing statistical analyses:

In general, to do an analysis you will click on ANALYZE, then click on a subcategory, and then click on the specific routine you are interested in. A dialogue box will open. You will click on the variables of interest, move then into the appropriate places in the dialogue box, choose appropriate options and statistics, and then press “OK.” Your output will open as a new window. Note that you can cut and paste this output into Excel and Word.

1) FREQUENCY DISTRIBUTIONS:

ANALYZE>DESCRIPTIVE STATISTICS>FREQUENCIES.

Calculate the Mean, Median, and Mode with the STATISTICS button.

Produce bar charts and histograms with the CHARTS button.

2) DESCRIPTIVE STATISTICS:

ANALYZE>DESCRIPTIVE STATISTICS>DESCRIPTIVES.

Default statistics are minimum, maximum, mean, and standard deviation. A few additional statistics are available with the OPTIONS button.

3) CORRELATION:

ANALYZE>CORRELATE>BIVARIATE

Move all of the variables of interest into the dialog box, and SPSS will calculate Pearson r correlations for every possible pair of variables. For instance, putting participant’s education (EDUC1) and partner’s education (EDUC2) in the box gives this output:

Correlations
		Education	Partner's educ.
Education	Pearson Correlation	1.000	.381**
	Sig. (2-tailed)		.000
	N	400.000	400
Partner's educ.	Pearson Correlation	.381**	1.000
	Sig. (2-tailed)	.000
	N	400	400.000
**. Correlation is significant at the 0.01 level (2-tailed).

Pearson r is .381, a moderate relationship. The two-tailed significance is .000. This actually means that the actual p value is less than .0005, but SPSS rounds off to three decimal places. One of the great advantages of computerized statistical calculations over hand calculations is that you don’t need to look up values in a table to see if they’re significant. If the p value is less than your alpha level, then you have significance. This p value indicates that p < .01. Of course, you could also say that p < .001.

4) ONE SAMPLE T-TEST:

ANALYZE>COMPARE MEANS>ONE SAMPLE T-TEST

Bring the variable you’re testing into the “Test Variable(s)” box. Type the population mean that you’re testing it against into the “Test Value” box.

5) INDEPENDENT-SAMPLES T-TEST

ANALYZE>COMPARE MEANS>ONE SAMPLE T-TEST

Bring your dependent variable into the “Test Variable(s)” box and your independent variable into the “Grouping Variable” box. Then click “DEFINE GROUPS” and put in the values that represent your two groups. Note that if this is a numeric variable, you will need to put in the values, not the labels. For instance, if you’re using GENDER1 as your grouping variable you will put in “1” and “2” not “Male” and “Female.”

Take your results from the first line under “T-test for equality of means.” For instance, with “Lifestyle” as my dependent variable and gender as my groups I get:

t = 1.215, df = 398. It shows the two-tailed significance as .225. Since this is greater than the conventional .05 alpha level, we would say that p > .05, the results are non-significant.

6) PAIRED-SAMPLES T-TEST

ANALYZE>COMPARE MEANS>PAIRED-SAMPLES T-TEST

Bring the two variables that you are comparing into the “Paired variables” box. The output shows the correlation between the two variables and tests in for significant. Below this, the box labeled “Paired Samples Test” shows the t, df, and significance level to the right.

7) CHI-SQUARE TEST

ANALYZE>DESCRIPTIVE STATISTICS>CROSSTABS

Bring your row variable and your column variable into the boxes. Click the STATISTICS button and check “Chi-square.”

8) ONE-WAY ANOVA

ANALYZE>COMPARE MEANS>ONE-WAY ANOVA

Bring your dependent variable into the “Dependent List” box and your factor or independent variable into the “Factor” box. Click the OPTIONS button and check “Descriptives” to see group means.

9) TWO-WAY ANOVA

GENERAL LINEAR MODEL>UNIVARIATE

Bring your dependent variable into the “Dependent Variable” box and your factors or independent variables into the “Fixed Factor(s)” box. Click the OPTIONS button and check “Descriptives” to see group means.

10) REGRESSION

ANALYZE>REGRESSION>LINEAR

Put your dependent variable, the one you are predicting, into the “Dependent Variable” box, and your predictor or independent variable in the “Independent Variable” box. With income as the dependent variable and age as the predictor variable you get this output:

Coefficientsa
Model	Unstandardized Coefficients	Standardized Coefficients	t	Sig.
	B	Std. Error	Beta
1	(Constant)	144.257	6974.488		.021	.984
	Age	1418.136	202.515	.331	7.003	.000
a. Dependent Variable: Income

Your regression formula can be generated from the numbers in the “B” column under “Unstandardized Coefficients.” The constant is your Y-intercept and the other number is your slope. So predicting income from age would give this formula:

INCOME = (1418.1*AGE) + 144.3

This formula would indicate that the average 40-year-old is earning:

(1418.1*40)+144.3 = $56,868.

Additional techniques

SPSS provides ways to look at subsets of data and to combine or transform variables. You don’t need to know these techniques for any of the quizzes, but you may wish to use them on your final paper.

Data subsets

Suppose you would like to look only at men who are college graduates. On GENDER1, men are coded as “1” and on the EDUC1 variable college graduates are coded as 4 through 7 (4 = BA/BS, 5 = some grad school, 6 = MA/MS, 7 = doctorate). To select this group I would select:

DATA>SELECT CASES

select “If condition is satisfied” and click “IF.” In the box at the top I’d type

GENDER1 = 1 AND EDUC1>=4

and click CONTINUE. Make sure “Filter out unselected cases” is selected and then click “OK.” You will see a diagonal line through the case number of every case that has been deselected. To return to the full file select:

DATA>SELECT CASES

and click “All cases.”

Data transformations

Suppose you’d like to create a new variable, TOTINC, showing the total income for both members of the couple.

TRANSFORM>COMPUTE VARIABLE

In the “Target Variable” box type TOTINC, the name of your new variable. In the “Numeric Expression” box type INC1+INC2, and then hit OK. The new variable will be added to your file, at the far right.