Statistical Analysis of Thesis

Running Descriptives on SPSS

The Descriptives procedure allows you to get descriptive data about any of your scale level

variables. The most common use of the procedure is to find the mean and standard deviation

for a variable.

The procedure is used with scale level variables, most likely scores on some measure. If you

run Descriptives on a nominal level variable, you get information that is meaningless. For

example, if your variable is gender with values coded as 0 = male and 1 = female and you run a

Descriptives, you might get a mean of 0.7. What does that mean? It doesn’t tell you anything

about the participants in your study. For nominal or ordinal variables, you must run Frequencies.

Running Descriptives

Follow these commands:

 Analyze  Descriptive Statistics  Descriptives

 Highlight each variable of interest from the box on the left side of the screen, then click

on the arrow button.

 Click on Options.

 Mean, standard deviation, minimum and maximum should be checked as the defaults.

Click on Skewness and Kurtosis.

 Click on Continue.

 Finally, click OK.

Practice Descriptives

Use the High School and Beyond data set (hsbdata.sav) to practice running Descriptives.

 Analyze  Descriptive Statistics  Descriptives

 Highlight math achievement, then click on the arrow button.

 Highlight scholastic aptitude test - math, then click on the arrow button.

 Click on Options.

 Mean, standard deviation, minimum and maximum should be checked as the defaults.

Click on Skewness and Kurtosis.

 Click on Continue.

 Click OK.

Reading a Descriptives Output

The following is the Descriptives Output for the practice session and the presentation.

Descriptives

This is all there is to the output, but there is a lot of information. Look at the first five columns in

blue. The first column has the number of cases with valid data for each test. The next two

columns tell you the minimum and maximum score that were earned on each test. Then the

mean and standard deviation for each test is listed.

Descriptive Statistics

N Minimum Maximum Mean Std.

Deviation

Skewness Kurtosis

Statistic Statistic Statistic Statistic Statistic Statistic Std.

Error

Statistic Std.

Error

math achievement test 75 -1.67 23.67 12.5645 6.67031 .044 .277 -.940 .548

scholastic aptitude test - math 75 250 730 490.53 94.553 .128 .277 .943 .548

Valid N (listwise) 75

Now look at the last two columns in purple. These are the skewness and kurtosis statistics.

These statistics are more precise than looking at a histogram of the distribution. The rule to

remember is that if either of these values for skewness or kurtosis are less than ± 1.0, then the

skewness or kurtosis for the distribution is not outside the range of normality, so the distribution

can be considered normal. If the values are greater than ± 1.0, then the skewness or kurtosis for

the distribution is outside the range of normality, so the distribution cannot be considered

normal.

This column tells you the

number of cases with

scores.

These two columns tell

you the minimum and

maximum score that

were earned on each

test.

These two columns tell

you the mean score and

the standard deviation

for each test.

These two columns give

the skewness and

kurtosis statistics. Values

should be less than ± 1.0

to be considered normal.

For both of these variables the skewness is very close to 0, indicating that the distribution of

scores in not skewed. But look at the kurtosis. The math achievement test has a negative

kurtosis, meaning that the distribution is slightly flatter than normal or platykurtik. Just the

opposite is true for the SAT math test. While it is not outside the normal range, the distribution is

tall, it is leptokurtik, hence the positive kurtosis value.

For skewness, if the value is greater than + 1.0, the distribution is right skewed. If the value is

less than -1.0, the distribution is left skewed.

For kurtosis, if the value is greater than + 1.0, the distribution is leptokurtik. If the value is less

than -1.0, the distribution is platykurtik.