Sunteți pe pagina 1din 10

Using SPSS Windows for Research: An Exploratory Tutorial

Raj Echambadi Friday, the 13th of October, 2000

A. SPSS Windows and Files

You are likely to work with three kinds of windows --

• The Data Editor Window -- to define and enter your data and to perform statistical
procedures. File extension -- *. sav
• The results of the statistical tests appear in the Output Window. File extension -- *.
spo
• The Syntax Window can be used to keep a record of the operations that you perform
on your data. This window will automatically open when you click a Paste function. I
have no clue on this syntax window; to find out how to do this, you can read through
the Help commands. File extension -- *. Sps

Data Issues

• Open new SPSS data file


• Create new data
• Data view
• Variable view

1
Variable View Window

Change Settings Default Value


Type -- allows you to define the type of variable numeric
Width -- total number of characters 8
Decimals -- number of characters beyond the decimal point 2
Label -- allows you to list a more extensive label for your
variable. Eight character variable names are difficult to
none
remember, and we recommend that you always exercise the
option of listing a more descriptive label.
Value -- allows you to provide labels for the various levels of a
none
variable
Missing Values -- enables you to designate certain scores as
none
missing.
Columns --allows you to change the maximum number of
8
characters in a column.
Align -- allows you to determine the alignment of your column right
Measure -- allows you to determine the kind of scale for that
scale
particular variable

Variable description of data used in this tutorial

18 respondents were asked to reveal their attitudes about six types of soft drinks. The
following statements were used to record their attitudes.

v1. Brand X has a refreshing taste.


v2. I prefer Brand X because it has fewer calories than other drinks.
v3. Brand X quenches my thirst immediately.
v4. I like the sweet taste of Brand X.
v5. I prefer drinking Brand X after workouts and sports because it gives me
energy.
v6. I prefer Brand X because it comes in environment friendly packaging.
v7. Brand X has minerals and vitamins that help quench my deep down body thirst.
v8. Brand X has a unique flavor of its own.
v9. Brand X has the right mix of minerals and vitamins that are healthy for my
body.
v10. I prefer to drink Brand X when I am really thirsty.

2
B. Descriptive Statistics for Overall Sample

Descriptive Statistics

N Minimum Maximum Mean Std. Deviation


refreshing taste 95 1.00 7.00 5.0842 1.4341
fewer calories 95 1.00 7.00 4.0737 1.5591
quench thirst 95 1.00 7.00 4.6526 1.3743
Valid N (listwise) 95

3
C. Descriptive Statistics for different brands

Step 1. First click on Data and then Split File. This allows you to split the file according
to a particular variable and conduct separate analyses for each level of the variable.
Step 2. Next you should select Organize output by groups and move Brands from the
variable list to the Groups Based on list. When you have done this, click on OK.
Step 3. Now, click on Analyze, Descriptive Statistics, and Descriptives. Select the
desired variable(s) and Options.

Illustrative output

brand = Allsport
Descriptive Statisticsa

N Minimum Maximum Mean Std. Deviation


refreshing taste 14 1.00 5.00 3.5714 1.1579
fewer calories 14 3.00 6.00 3.9286 .9169
quench thirst 14 3.00 7.00 5.5714 1.1579
Valid N (listwise) 14
a. brand = Allsport

C2. Descriptive Statistics for select cases

D. Frequencies for the Sample

Step 1: From the "Analyze" menu, click "Descriptive Statistics," then move to the sub
menu and click on "Frequencies."

Illustrative output
refreshing taste

Cumulative
Frequency Percent Valid Percent Percent
Valid 1.00 1 1.1 1.1 1.1
2.00 3 3.2 3.2 4.2
3.00 9 9.5 9.5 13.7
4.00 20 21.1 21.1 34.7
5.00 21 22.1 22.1 56.8
6.00 23 24.2 24.2 81.1
7.00 18 18.9 18.9 100.0
Total 95 100.0 100.0

4
E. Computing a New Variable using Transform / Compute

Let's say a reviewer comes back and suggests that the analysis should be repeated by
using a composite variable for taste (v1, v4, and v8) and not for individual variables.

v1. Brand X has a refreshing taste.


v4. I like the sweet taste of Brand X.
v8. Brand X has a unique flavor of its own.

Step 1. Click on Transform and then Compute. This should produce the screen shown
below.

Step 2. Type the name of the variable that you wish to create (in this case, we will name
it "comptast") in the Target Variable field. Now, you need to type in the computation
that you wish to have performed in the Numeric Expression field. You can use all the
operations listed on the bottom of this screen, and it is important to realize that
operations within parentheses are performed first.

Step 3. Now you should click on the Type&Label field. This will enable you to create a
longer label for your variable and modify the type and width of the variable. Once you
have done this, press Continue. Then press OK.

Step 4. Continue with your analysis.

5
F. Computing a New Variable using Transform / Recode
• Reverse coding

G. Replacing missing data with Transform / Missing data

H. Computing a new variable using Transform / Compute / If

I. Visual examination of the distributional characteristics of continuous variables

Step 1. From the "Analyze" menu choose "Descriptive Statistics," drag to the sub menu
and select "Explore."

J. Cross Tabulations

Step 1. From the "Analyze" menu choose "Descriptive Statistics," drag to the sub menu
and select "Cross Tabulations."

K. Computing correlations

Step 1. Click on Analyze, then Correlate, then Bivariate.


Step 2. Move the variables that you wish to have correlated into the Variables field.
Step 3. If you wish to print the means and standard deviations, then click on Options and
select the desired statistics. Then click on Continue and then OK.

Note: SPSS computes all possible correlations among the variables that you specified.
The correlations are listed in the first third of the table, the significance levels of the
correlations are listed in the second third of the table, and the valid number of pairs of
scores are listed in the bottom portion of the table.

6
Illustrative output
Correlations

refreshing
taste fewer calories quench thirst
refreshing taste Pearson Correlation 1.000 .130 -.303**
Sig. (2-tailed) . .208 .003
N 95 95 95
fewer calories Pearson Correlation .130 1.000 .454**
Sig. (2-tailed) .208 . .000
N 95 95 95
quench thirst Pearson Correlation -.303** .454** 1.000
Sig. (2-tailed) .003 .000 .
N 95 95 95
**. Correlation is significant at the 0.01 level (2-tailed).

L. Creating a Scatterplot for Pairs of Variables

Step 1. To create a scatterplot of your data, click on Graphs and then Scatter.
Step 2. In the Scatterplot menu, click on Simple and then Define.
Step 3. Now move the variables into the appropriate axes. When you have done this,
click on OK a scatterplot of the relationship between your variables is produced.

7
Illustrative output
Stupid Chart
8

3
refreshing taste

0
0 1 2 3 4 5 6 7 8

fewer calories

Step 4. As you can see, the SPSS default options for labeling this graph are not ideal.
You can modify your graph by double clicking on it, and this accesses the SPSS Chart
Editor. Once you are in the SPSS Chart Editor, you can basically do whatever you want.
You can make the title fit by clicking on the title and then on Format and then on Text.
With a little work, your graph can look like the one below.

Stupid Chart
8
Refreshing taste

0
0 1 2 3 4 5 6 7 8

Fewer calories

Step 5. You can save the template for your preferred options by selecting File and then
Save Chart Template. This will enable you to apply these options to future graphs.

8
L. Comparing Means between two groups on "Packaging"

Click on "Analyze," then point your mouse at "Compare Means" and then click on
"Independent-Samples T Test."

In d e p e n d e n t S a m p l e s T e s t

L e v e n e 's T e s t fo r
E q u a lit y o f V a r ia n c e s t- te s t fo r E q u a lity o f M e a n s
9 5 % C o n fid e n c e
I n te r v a l o f th e
M e a n S td . E r r o r D iffe r e n c e
F S ig . t df S ig . ( 2 - ta ile dD) iffe r e n c eD iffe r e n c e L o w e r U p p e r
p a c k a g in gE q u a l v a r ia n c e s
3 .0 0 2 .0 8 6 - 1 .1 0 1 93 .2 7 4 - .1 9 2 8 .1 7 5 1 -.5 4 0 6 .1 5 5 0
a ssum e d
E q u a l v a r ia n c e s
- 1 .0 9 9 9 0 .0 5 2 .2 7 5 - .1 9 2 8 .1 7 5 4 -.5 4 1 4 .1 5 5 7
n ot assu m e d

M. Reliability Analysis

Click on "Analyze," then "Scale " and then "Reliability Analysis" and click on "List
Items"

Reliability
****** Method 1 (space saver) will be used for this
analysis ******
Reliability Coefficients

N of Cases = 95.0 N of Items = 3

Alpha = .9460

9
N. Regression

Click on "Analyze," then "Regression " and then "Linear"


Dep. Variable = v1
Ind. Variables = v2, v3, v4

Model Summaryb

Adjusted Std. Error of Durbin-W


Model R R Square R Square the Estimate atson
1 .847 a .717 .708 .7747 2.167
a. Predictors: (Constant), sweet taste, fewer calories, quench thirst
b. Dependent Variable: refreshing taste

ANOVAb

Sum of
Model Squares df Mean Square F Sig.
1 Regression 138.708 3 46.236 77.034 .000 a
Residual 54.618 91 .600
Total 193.326 94
a. Predictors: (Constant), sweet taste, fewer calories, quench thirst
b. Dependent Variable: refreshing taste

Coefficientsa

Standardi
zed
Unstandardized Coefficien
Coefficients ts
Model B Std. Error Beta t Sig.
1 (Constant) 1.886 .435 4.340 .000
fewer calories .169 .059 .184 2.887 .005
quench thirst -.210 .068 -.201 -3.075 .003
sweet taste .705 .054 .767 13.125 .000
a. Dependent Variable: refreshing taste

You can save predicted values to use later.

N. Multivariate analysis can also be done

10

S-ar putea să vă placă și