* ************************************************
* SPSS RUN 4—ONEWAY AND TWOWAY
* FACTORIAL ANOVA
* LOGISTIC REGRESSION
* *************************************************.
GET FILE='c:\temp\werner.sav'.
FREQUENCIES
VARIABLES=agegroup .
*------Compare Means--One-Way Anova.
ONEWAY
chol BY agegroup
/POLYNOMIAL= 1
/STATISTICS DESCRIPTIVES HOMOGENEITY
/PLOT MEANS
/MISSING ANALYSIS
/POSTHOC = TUKEY BONFERRONI ALPHA(.05).
*------General Linear Model--Univariate for twoway anova--GLM can also be used for oneway.
UNIANOVA
chol BY agegroup pill
/METHOD = SSTYPE(3)
/INTERCEPT = INCLUDE
/POSTHOC = agegroup ( TUKEY )
/PLOT = PROFILE( agegroup*pill )
/PRINT = DESCRIPTIVE HOMOGENEITY
/CRITERIA = ALPHA(.05)
/DESIGN = agegroup pill agegroup*pill .
*------General Linear Model--Include Covariates in Model.
UNIANOVA
chol BY agegroup pill WITH height weight
/METHOD = SSTYPE(3)
/INTERCEPT = INCLUDE
/SAVE = PRED SEPRED RESID ZRESID
/PRINT = DESCRIPTIVE HOMOGENEITY
/CRITERIA = ALPHA(.05)
/DESIGN = height weight agegroup pill agegroup*pill .
*------Check Residuals for Normality -- Descriptive Statistics---Explore.
EXAMINE
VARIABLES=zre_1
/PLOT BOXPLOT STEMLEAF HISTOGRAM NPPLOT
/COMPARE GROUP
/STATISTICS DESCRIPTIVES
/CINTERVAL 95
/MISSING LISTWISE
/NOTOTAL.
*------Scatterplot of Residuals vs. Predicted Value.
GRAPH
/SCATTERPLOT(BIVAR)=pre_1 WITH zre_1
/MISSING=LISTWISE .
*------Setup for LOGISTIC REGRESSION.
RECODE
chol
(Lowest thru 249=0) (250 thru Highest=1) INTO hichol .
EXECUTE.
FREQUENCIES
VARIABLES=hichol .
CROSSTABS
/TABLES=agegroup BY hichol
/FORMAT= AVALUE TABLES
/STATISTIC=CHISQ
/CELLS= COUNT ROW.
*------Regression--Binary Logistic.
LOGISTIC REGRESSION VAR=hichol
/METHOD=ENTER acid age alb calcium height weight
/SAVE PRED
/CLASSPLOT
/CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .
*------AGE is Categorical.
LOGISTIC REGRESSION VAR=hichol
/METHOD=ENTER acid alb calcium height weight agegroup
/CONTRAST (agegroup)=Indicator(1)
/SAVE PRED
/CLASSPLOT
/CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .
LOGISTIC REGRESSION VAR=hichol
/METHOD=FSTEP(LR) acid alb calcium height weight agegroup
/CONTRAST (agegroup)=Indicator(1)
/CLASSPLOT
/PRINT=CI(95)
/CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .
LOGISTIC REGRESSION VAR=hichol
/METHOD=ENTER age calcium acid
/SAVE PRED
/CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .
*------Graphs.
COMPUTE xbeta1 = -15.5828+.0687*age+1.1046*calcium+.3493*acid .
EXECUTE .
GRAPH
/SCATTERPLOT(BIVAR)=xbeta1 WITH pre_4
/MISSING=LISTWISE .
COMPUTE logit = ln(pre_4/(1-pre_4)) .
EXECUTE .
GRAPH
/SCATTERPLOT(BIVAR)=xbeta1 WITH logit
/MISSING=LISTWISE .
Frequencies
Oneway—Oneway Anova
Post Hoc Tests
Homogeneous Subsets
Means Plots
Univariate Analysis of Variance
Post Hoc Tests
AGEGROUP
Homogeneous Subsets
Profile Plots
Univariate Analysis of Variance
Explore
Standardized Residual for CHOL
Standardized Residual for CHOL Stem-and-Leaf Plot
Frequency Stem & Leaf
1.00 -2 . 7
1.00 -2 . 3
7.00 -1 . 5555777
22.00 -1 . 0000111122222222233344
26.00 -0 . 55555566666777777778888889
37.00 -0 . 0000111111111111222222233333333444444
30.00 0 . 000000000111111222233333333444
29.00 0 . 55555566666666777777888899999
18.00 1 . 000000011122233444
8.00 1 . 55668899
1.00 2 . 0
2.00 Extremes (>=2.9)
Stem width: 1.00
Each leaf: 1 case(s)
Graph
Frequencies
Crosstabs
Logistic Regression
Block 0: Beginning Block
Block 1: Method = Enter
Step number: 1
Observed Groups and Predicted Probabilities
16
F
R 12
E H
Q L
U L H H
E 8 L L L H H
N L L L HH H
C L L L HH H HL H
Y LLL LHLL H LL H H HH
4 LLLLLLLL HLHLL LH H H H HH H
LLLLLLLLLHLLLLLH LH L HHHHH L H L HH H
LLLLLLLLLLHLLLLLLHLLHLHLHHHHLHLHH L LH HL H H
HLLLLLLLLLLLLLLLLLLLLLLLLLHLLLLLHLHLLHLHLHHHL LHHH H
Predicted
Prob: 0 .25 .5 .75 1
Group: LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH
Predicted Probability is of Membership for HIGH
The Cut Value is .50
Symbols: L - LOW
H - HIGH
Each Symbol Represents 1 Case.
Logistic Regression
Block 0: Beginning Block
Block 1: Method = Enter
Step number: 1
Observed Groups and Predicted Probabilities
16
F
R 12
E
Q H
U L HH
E 8 L LL
N LL LL HHH H
C LL LLLL HHHH L
Y LLLLLLL LLHH HL H H
4 LLLLLLL HLLHL HL HHHH H H H H
LLLLLLLHHLLLL HLL HH LHHH H L H H HH H
LLLLLLLLHLLLLL LLL LLH LHLH HHHLLL HHHH HLHH H
HLLLLLLLLLLLLLLLLLLHLLLHHLLLLHLLLLLLLHLHLLHLLH H H H
Predicted
Prob: 0 .25 .5 .75 1
Group: LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH
Predicted Probability is of Membership for HIGH
The Cut Value is .50
Symbols: L - LOW
H - HIGH
Each Symbol Represents 1 Case.
Logistic Regression
Block 0: Beginning Block
Block 1: Method = Forward Stepwise (Likelihood Ratio)
Step number: 1
Observed Groups and Predicted Probabilities
80
F
R 60 H H
E H H H
Q L H H
U L L H
E 40 L L H
N L L H
C L L H
Y L L H
20 L L L
L L L
L L L
L L L
Predicted
Prob: 0 .25 .5 .75 1
Group: LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH
Predicted Probability is of Membership for HIGH
The Cut Value is .50
Symbols: L - LOW
H - HIGH
Each Symbol Represents 5 Cases.
Step number: 2
Observed Groups and Predicted Probabilities
16
H
H
F H
R 12 H H
E HL H
Q HL HH H
U LL HL HH H H
E 8 LL HL HHL H H
N LL LL LHL L H H
C LL LL LHL LH H H
Y LL LL LHL LL H L H H H
4 LLHLLL LLL LL H L H H L H H
LLLLLLL LLL LL L L HH H HL H HL H HH H
LLLLLLL LLL LL LL L LLHHLHHL H HL HL HLH HH
LLLLLLLL LLL LL LLHL LLHLLLLL LHHL LL HLH LLH H
Predicted
Prob: 0 .25 .5 .75 1
Group: LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH
Predicted Probability is of Membership for HIGH
The Cut Value is .50
Symbols: L - LOW
H - HIGH
Each Symbol Represents 1 Case.
Step number: 3
Observed Groups and Predicted Probabilities
16
F
R 12
E
Q H
U L HH
E 8 L L LH H
N L LHLL H H H
C LLLLLL HHH HH H
Y LLLLLL HHHLHL H H H
4 HLLLLLLLHHHHLHL L L H H H H
LLLLLLLLLLLLLLLH L L H H HH L HH HHH HH
LLLLLLLLLLLLLLLLHL LLH LHLLHLHHL LHH LH HHHH
LLLLLLLLLLLLLLLLLLL LLL LLLLHLHLL LLHHLHLLHHHLH H H
Predicted
Prob: 0 .25 .5 .75 1
Group: LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH
Predicted Probability is of Membership for HIGH
The Cut Value is .50
Symbols: L - LOW
H - HIGH
Each Symbol Represents 1 Case.
Logistic Regression
Block 0: Beginning Block
Block 1: Method = Enter
Graph
Graph
1