Stat Computing > SPSS > Textbook Examples > Applied Regression Analysis, by Fox

SPSS Textbook Examples
Applied Regression Analysis by John Fox
Chapter 8: Analysis of variance

page 160 The table in the middle of the page.

GET FILE='D:\duncan.sav'.

EXAMINE
  VARIABLES=prestige BY occ_type
  /PLOT NONE
  /STATISTICS DESCRIPTIVES.

page 161 Figure 8.1 Parallel boxplots for occupational prestige by type of occupation.

EXAMINE
  VARIABLES=prestige BY occ_type
  /PLOT=BOXPLOT
  /STATISTICS=NONE
  /NOTOTAL.
 

The table in the middle of page 161.

ONEWAY
  prestige BY occ_type.

page 167 Table 8.2 Conformity by authoritarianism and partner's status, for Moore and Krupat's (1971) experiment. Each cell shows (from top to bottom) the conformity mean, standard deviation, and cell frequency.

GET FILE='D:\moore.sav'.

MEANS
  TABLES=conform  BY status by fcat
  /CELLS MEAN COUNT STDDEV. 

page 169 Figure 8.5 Cell means for the Moore and Krupat conformity experiment.

compute fcat1 = fcat.
recode fcat1 (1 = 3) (2 = 1) (3 = 2).
execute.

UNIANOVA
  conform  BY status fcat1
  /PLOT = PROFILE( fcat1*status )
  /DESIGN = status fcat1 status*fcat1. 

page 170 Figure 8.6 Data from Moore and Krupat's experiment on conformity and authoritarianism. The horizontal coordinates of the points have been jittered to separate overlapping points. The means are plotted as X's, with the profiles connected by broken lines. Note the two unusual points, for observations 16 and 19, in the high-authoritarian, low-status-partner group.

NOTE: SPSS does jittering only if the igraph command is used, but it still can be difficult to see how many points are in one place. Therefore, we would use "sunflowers" for this graph.  Also, we know how to put the predicted values on the graph with the data points. 

NOTE: SPSS will allow you to see how many observations are represented by a single point on the scatterplot with "sunflowers."  To add sunflowers to a scatterplot, create the scatterplot, and then use SPSS's chart editor.  To use the chart editor, double click on the graph. This will open the chart editor. Next, select "chart" from the list across the top. Next, select "options". In the bottom left, there is a check box labeled "show sunflowers". Single click in the box to active the use of sunflowers. By clicking on the "sunflower options" button, you can control how many observations each pedal represents, the resolution and whether the point is at the center of the pedals or at the mean. When you are finished, click on "OK" and then close the chart editor. The changes that you made in the chart editor will than take effect on you graph in the output window.

page 177 The sums of squares listed in the middle of the page.

compute c1 = fcat.
recode c1 (1 = 1) (2 = 0) (3 = -1).
compute c2 = fcat.
recode c2 (1 = 0) (2 = 1) (3 = -1).
compute r = status.
recode r (2 = -1).
compute rc1 = r*c1.
compute rc2 = r*c2.
execute.

Sums of squares (alpha, beta, gamma) and TTS.

regression 
 /dep=conform 
 /method=enter r c1 c2 rc1 rc2. 

Sums of squares (alpha, beta)

regression 
 /dep=conform 
 /method=enter r c1 c2. 

Sums of squares (alpha, gamma)

regression 
 /dep=conform 
 /method=enter r c1 c2. 

Sums of squares (beta, gamma)

regression 
 /dep=conform 
 /method=enter c1 c2 rc1 rc2. 

Sums of squares (alpha)

regression 
 /dep=conform 
 /method=enter r. 

Sums of squares (beta)

regression 
 /dep=conform 
 /method=enter c1 c2. 

page 178 Table 8.6 Analysis of variance table for Moore and Krupat's conformity experiment. Alternative tests are shown for the Partner's status and authoritarianism main effects.

NOTE: This yields the values for residuals (called "error" in the SPSS output) and the total (called "corrected total" in the SPSS output).

Partner's status:

regression 
 /dep conform 
 /method=enter c1 c2 rc1 rc2 
 /method=test(r). 

regression 
 /dep conform 
 /method=enter c1 c2 
 /method=test(r).

Authoritarianism:

regression 
 /dep conform 
 /method=enter r 
 /method=test(c1 c2). 

regression 
 /dep conform 
 /method=enter r rc1 rc2 
 /method=test(c1 c2). 

Partner's status and authoritarianism:

regression 
 /dep conform 
 /method=enter r c1 c2 
 /method=test(rc1 rc2). 

page 192 The R-squared in the middle of the page.

compute d = 0.
if status = 2 d = 1.
compute intfd = fscore*d.
execute.

REGRESSION
  /DEPENDENT conform
  /METHOD=ENTER fscore d intfd. 

page 194 The R-squared in the middle of the page.

compute s = status.
if status = 1 s = -1.
compute intfs = fscore*s.
execute.

REGRESSION
  /DEPENDENT conform
  /METHOD=ENTER fscore s intfs. 

page 197 Means and standard deviations in the middle of the page.

GET FILE='D:\friendly.sav'.

MEANS
  TABLES=correct  BY cond
  /CELLS MEAN COUNT STDDEV. 

page 198 Figure 8.8 Number of words recalled (out of 40) by experimental condition, from Friendly and Franklin's (1980) memory experiment. The horizontal coordinates of the points are jittered to separate overlapping points. The mean of each group is plotted as an "X" and the means are connected by a broken line.

NOTE: SPSS has option for jittering if you use the igraph command and the editor in the output window.  You cannot get jittering using syntax.  Another way to see how many points are plotted in one place is to use "sunflowers".  To add sunflowers to a scatterplot, create the scatterplot, and then use SPSS's chart editor.  To use the chart editor, double click on the graph. This will open the chart editor. Next, select "chart" from the list across the top. Next, select "options". In the bottom left, there is a check box labeled "show sunflowers". Single click in the box to active the use of sunflowers. By clicking on the "sunflower options" button, you can control how many observations each pedal represents, the resolution and whether the point is at the center of the pedals or at the mean. When you are finished, click on "OK" and then close the chart editor. The changes that you made in the chart editor will than take effect on you graph in the output window.

if cond = "SFR" g1 = 1.
if cond = "Before"  g1 = 2.
if cond = "Meshed" g1 = 3.
execute.

UNIANOVA
  correct BY g1
  /PLOT = PROFILE(g1)