UCLA Academic Technology Services HomeServicesClassesContactJobs
Stat Computing > SPSS > Textbook Examples > Experimental Design, 3rd, by Kirk
 Search

SPSS Textbook Examples
Experimental Design by Roger Kirk
Chapter 5: Completely Randomized Design

This shows how to get the results of Chapter 5 using SPSS. Below is how to read the first data file used in this chapter. 
data list free / y a order.
begin data.
4 1 1
6 1 2
3 1 3
3 1 4
1 1 5
3 1 6
2 1 7
2 1 8
4 2 1
5 2 2
4 2 3
3 2 4
2 2 5
3 2 6
4 2 7
3 2 8
5 3 1
6 3 2
5 3 3
4 3 4
3 3 5
4 3 6
3 3 7
4 3 8
3 4 1
5 4 2
6 4 3
5 4 4
6 4 5
7 4 6
8 4 7
10 4 8
end data.
Table 5-2.1, part iii.
means tables=y by a.
Case Processing Summary

Cases
Included Excluded Total
N Percent N Percent N Percent
Y A 32 100.0% 0 .0% 32 100.0%

Report
Y
A Mean N Std. Deviation
1.00 3.0000 8 1.5119
2.00 3.5000 8 .9258
3.00 4.2500 8 1.0351
4.00 6.2500 8 2.1213
Total 4.2500 32 1.8837

Table 5.3-2, ANOVA table. 

ONEWAY y BY a.
ANOVA
Y

Sum of Squares df Mean Square F Sig.
Between Groups 49.000 3 16.333 7.497 .001
Within Groups 61.000 28 2.179

Total 110.000 31


Top of page 173, t-tests for three contrasts. 

ONEWAY y BY a 
 /CONTRAST 1 -1 0 0 
 /CONTRAST 0 0 1 -1 
 /CONTRAST 1 1 -1 -1.
ANOVA
Y

Sum of Squares df Mean Square F Sig.
Between Groups 49.000 3 16.333 7.497 .001
Within Groups 61.000 28 2.179

Total 110.000 31



Contrast Coefficients

A
Contrast 1.00 2.00 3.00 4.00
1 1 -1 0 0
2 0 0 1 -1
3 1 1 -1 -1

Contrast Tests

Contrast Value of Contrast Std. Error t df Sig. (2-tailed)
Y Assume equal variances 1 -.5000 .7380 -.678 28 .504
2 -2.0000 .7380 -2.710 28 .011
3 -4.0000 1.0437 -3.833 28 .001
Does not assume equal variances 1 -.5000 .6268 -.798 11.603 .441
2 -2.0000 .8345 -2.397 10.155 .037
3 -4.0000 1.0437 -3.833 19.431 .001
Table 5.4-1, Kirk illustrates all pairwise comparisons using fisher hayter test. SPSS does not have this test, but the most similar test is a "Tukey" test, which you can get with the commands below. This does produce a table like 5-4.1, but you would need to compute the "critical difference" by hand using the formula on page 174.
ONEWAY y BY a 
 /POSTHOC = TUKEY ALPHA(.05).
ANOVA
Y

Sum of Squares df Mean Square F Sig.
Between Groups 49.000 3 16.333 7.497 .001
Within Groups 61.000 28 2.179

Total 110.000 31






Multiple Comparisons
Dependent Variable: Y
Tukey HSD

Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval
(I) A (J) A Lower Bound Upper Bound
1.00 2.00 -.5000 .7380 .905 -2.5150 1.5150
3.00 -1.2500 .7380 .346 -3.2650 .7650
4.00 -3.2500(*) .7380 .001 -5.2650 -1.2350
2.00 1.00 .5000 .7380 .905 -1.5150 2.5150
3.00 -.7500 .7380 .741 -2.7650 1.2650
4.00 -2.7500(*) .7380 .005 -4.7650 -.7350
3.00 1.00 1.2500 .7380 .346 -.7650 3.2650
2.00 .7500 .7380 .741 -1.2650 2.7650
4.00 -2.0000 .7380 .052 -4.0150 1.499E-02
4.00 1.00 3.2500(*) .7380 .001 1.2350 5.2650
2.00 2.7500(*) .7380 .005 .7350 4.7650
3.00 2.0000 .7380 .052 -1.4986E-02 4.0150
The mean difference is significant at the .05 level.


Y
Tukey HSD

N Subset for alpha = .05
A 1 2
1.00 8 3.0000
2.00 8 3.5000
3.00 8 4.2500 4.2500
4.00 8
6.2500
Sig.
.346 .052
Means for groups in homogeneous subsets are displayed.
a Uses Harmonic Mean Sample Size = 8.000.


On the middle of page 175, Kirk illustrates how to do a Scheffe test. We can compute FS as illustrated by Kirk as shown below. This gives a "t" value of -2.766, and we can square that value to get the "FS" value of 7.651 You would need to compute the critical value of "FS" as shown on page 175 of Kirk by hand. 

ONEWAY y BY a 
 /CONTRAST= 3 -1 -1 -1.
ANOVA
Y

Sum of Squares df Mean Square F Sig.
Between Groups 49.000 3 16.333 7.497 .001
Within Groups 61.000 28 2.179

Total 110.000 31



Contrast Coefficients

A
Contrast 1.00 2.00 3.00 4.00
1 3 -1 -1 -1

Contrast Tests

Contrast Value of Contrast Std. Error t df Sig. (2-tailed)
Y Assume equal variances 1 -5.0000 1.8077 -2.766 28 .010
Does not assume equal variances 1 -5.0000 1.8371 -2.722 11.459 .019


Section 5.5, pages 177-182. The only measure of strength of effect that SPSS automatically computes is eta squared (page 180) as illustrated below. For other measures of strength of effect or effect size, you will need to compute these as described by Kirk. 

UNIANOVA y BY a 
 /PRINT = ETASQ.
Between-Subjects Factors

N
A 1.00 8
2.00 8
3.00 8
4.00 8

Tests of Between-Subjects Effects
Dependent Variable: Y
Source Type III Sum of Squares df Mean Square F Sig. Eta Squared
Corrected Model 49.000(a) 3 16.333 7.497 .001 .445
Intercept 578.000 1 578.000 265.311 .000 .905
A 49.000 3 16.333 7.497 .001 .445
Error 61.000 28 2.179


Total 688.000 32



Corrected Total 110.000 31



a R Squared = .445 (Adjusted R Squared = .386)


In Table 5.7-2, page 194 Kirk shows how to test for linear trend, and departure for linear trend. This can be done in SPSS as shown below. The test of linear trend is shown by "Linear Term, Contrast" and the departure from linearity is shown right below labeled "Deviation". This also shows the tests shown in table 5.7-3 for the test of "quadratic" and "cubic" trend, and is summarized in Table 5.7-4, page 196. 

ONEWAY y BY a 
 /POLYNOMIAL= 3.
ANOVA
Y

Sum of Squares df Mean Square F Sig.
Between Groups (Combined) 49.000 3 16.333 7.497 .001
Linear Term Contrast 44.100 1 44.100 20.243 .000
Deviation 4.900 2 2.450 1.125 .339
Quadratic Term Contrast 4.500 1 4.500 2.066 .162
Deviation .400 1 .400 .184 .672
Cubic Term Contrast .400 1 .400 .184 .672
Within Groups 61.000 28 2.179

Total 110.000 31


Figure 5.7-3, page 199. skipped for now.

Table 5.7-6 shows how to test for departure from linearity. This can be done using the previous table, inspecting the "Linear Term, Deviation" to get F=1.125. 








	

		
		
How to cite this page

		
		
Report an error on this page

	





UCLA Researchers are invited to our Statistical Consulting Services

We recommend others to our list of Other Resources for Statistical Computing Help

These pages are Copyrighted (c) by UCLA Academic Technology Services


The content of this web site should not be 
construed as an endorsement of any particular web site, book, or software 
product by the University of California