SPSS_Examples.doc

SPSS Examples     Page 17

SPSS

Examples

(Version 15)

Topic                            Pages

General Notes              2

Descriptive Statistics              3

One-Sample Tests              4-5

Two-Sample Tests              6-7

One-Way Analysis of Variance              8-9

Factorial Design              10-11

Unbalanced Design              12

Chi-Square Test of Independence              12-14

Correlation              15

Cronbach’s Alpha              16

Simple Linear Regression              17-20

Regression Diagnostics              21-22

Multiple Regression              23-34

Prepared by Bob Schulman, April 2007

General Notes

All of these examples were run on version 15 of SPSS.

SPSS has two main windows, the Data Editor window for input and the Viewer window for output.  Data are entered into the Data Editor window one row per observation and one column per variable.  To assign variable names, use the Variable View tab of the Data Editor window.  Variables created and saved during an analysis are added to the Data Editor window.

To run an analysis, pull down the Analyze menu, select a category, and then a procedure from the submenu.  This action is represented with arrows.  For example, Analyze è Descriptive Statisitcs è Descriptives denotes pulling down the Analyze menu, selecting Descriptive Statistics, and then selecting Descriptives from the submenu.  Variables and options are identified in a dialog box, with additional features selected from subdialog boxes.  Use the arrow buttons to move selected variables into the target lists and click on OK to run the procedure.  The results of the analysis are displayed in the Viewer window.

Descriptive Statistics

Example:  "Sales Contracts"

A large company was interested in the productivity of its sales force.  Six salespersons were randomly selected, and the number of contracts negotiated in the past month was determined for each.  The resulting values were 7, 4, 10, 2, 9, and 4.

Data Editor

One variable, named SALES, is entered in the first column.

Analyze è Descriptive Statistics è Descriptives

Variable = SALES

Options = Mean, Std. deviation, Variance, Range, S.E. mean

Viewer

Descriptive Statistics

One-Sample Tests

Example:  "Sprinkler Activation Time"

A fire prevention sprinkler system is designed to activate within 25 seconds when the temperature reaches a certain threshold.  In 20 trials at the threshold temperature, the following activation times were recorded:

              21              23              33              29              25

              39              34              19              24              30

              24              35              31              41              27

              29              39              18              37              21

Is the system responding within the designed time limit?

One variable, named TIME, is entered in the first column of the Data Editor.  SPSS produces the t test through one procedure, and nonparametric tests (the signed rank test and sign test) through a different set of menu commands.  Accordingly, two separate analyses are shown below.

Data Editor – t Test

The t test allows specification of the hypothesized mean, 25 for our example, but automatically prints a two-tailed p-value.

Analyze è Compare Means è One-Sample T Test

Test Variable(s) = TIME

Test Value = 25

Options = 95% Confidence Interval

Viewer – t Test

T-Test

One-Sample Statistics

One-Sample Test

Data Editor – Nonparametric Tests

Technically, SPSS does not perform nonparametric tests for a single mean.  However, by defining a second variable (CONSTANT) with the value of 25 for each observation, we can use a paired-data approach to produce the one-sample nonparametric tests.  The signed rank test is requested using the keyword Wilcoxon, the name of its developer.

Analyze è Nonparametric Tests è 2 Related Samples

Test Pair(s) List = TIME, CONSTANT

Test Type = Wilcoxon, Sign

Viewer – Nonparametric Tests

NPar Tests

Wilcoxon Signed Ranks Test

Ranks