'

Chapter 8 Comparing Two Groups

&
1

'

Tests comparing two groups Two independent samples Two-sample t-test(normal populations) Wilcoxon rank-sum test (non-parametric) Two related samples Paired t-test (normal population) Sign Test or Wilcoxon signed rank test (non-parametric)

&
2

'

Two independent samples Subjects are randomly assigned to a control or treatment group (where a drug is administered). There response times to a stimulus are then measured.

&
3

'

Data Response time in millisecond Control 80 93 83 89 98 Treatment 100 103 104 99 102

&
4

' Two-sample t-test: SAS Test H0 : t c = 0 against H0 : t c = 0 /Example 8.1 Two sample test / data ex8 1; inle F:\ST2137\lecdata\ex8 1.txt rstobs=2; input group $ time; proc ttest data=ex8 1; title t-test Example; class group; var time; run; /proc ttest uses a class statement to identify the two groups of subjects./ &
5

'

Two-sample t-test: SAS output

&
6

'

Two-sample t-test: SAS output

&
7

' Two-sample t-test: R >ex8.1=read.table(F:/ST2137/lecdata/ex8 1.txt,header=T) >attach(ex8.1) >cont=time[group==c] >treat=time[group==t] >var.test(cont,treat)#Test if the variables are equal F test to compare two variances data: cont and treat F=12.3953, num df=4, denom df=4, p-value=0.03177 alternative hypothesis: true ratio of variances is not equal to 1 95 percent condence interval: 1.290573 119.051498 sample estimates: ratio of variances 12.39535 &
8

'

Two-sample t-test: R >t.test(cont,treat,mu=0,var.equal=FALSE) Welch Two Sample T-Test data: cont and treat t=-3.8302 df=4.641, p-value=0.01410 alternative hypothesis: true dierence in means is not equal to 0 95 percent condence interval: -21.931655 -4.068345 sample estimates: mean of x mean of y 88.6 101.6

&
9

'

Two-sample t-test: SPSS Analyze Compare Means Independent Sample T tests... Move time to the Test Variables panel and group to the Grouping Variable panel. Click Dene Groups... Input the values for the two groups ContinueOK

&
10

' Two-sample t-test: SPSS

&
11

' Two-sample t-test: SPSS

&
12

' Two independent samples: Nonparametric tests Assumptions of two-sample t-test not met The data are not normally distributed and the sample size is small

&
13

'

Two independent samples: Nonparametric tests E.g. The following numbers in a psychology experiment that measured the response to a stimulus 0605769480 7056600 It may be due to a threshold eect. The response is either 0 (the stimulus is not detected), or, once the stimulus is detected, the average response is about 6.

&
14

'

Two independent samples: Nonparametric tests The data values may only represent ordered categories. E.g. Scales such as 1=very mild, 2=mild, 3=moderate, 4=strong, 5=severe reect the strength of a response We cannot say that a score of 4(strong) is worth twice the score of 2 (mild). We need a nonparametric test to analyze dierences in central tendencies for ordinal data. For very small samples, nonparametric tests are often more appropriate since assumptions concerning distributions are dicult to determine.

&
15

'

Wilcoxon rank sum test (Mann-Whitney U-test) Consider the following experiment Group A: No treatment. Group B: Treated with a drug to prevent tumor formation Both groups are exposed to a chemical that encourages tumor growth The masses (in grams) of tumors in Groups A and B are A: 3.1 2.2 1.7 2.7 2.5 B: 0.0 0.0 1.0 2.3 Mass: 0.0 0.0 1.0 1.7 2.2 2.3 2.5 2.7 3.1 Group:B B B A A B A A A Rank: 1.5 1.5 3 4 5 6 7 8 9

&
16

'

Wilcoxon rank sum test (Mann-Whitney U-test) Sum ranks of Group A=4+5+7+8+9=33 Sum ranks of Group B=1.5+1.5+3+6=12 If there were smaller tumors in Group B, we would expect the Bs to be at the lower rank ordering and therefore have a smaller sum of ranks then the As.

&
17

'

Wilcoxon rank sum test: SAS data ex8 2; inle F:\ST2137\lecdata\ex8 2.txt rstobs=2; input group $ mass; proc npar1way data=ex8 2 wilcoxon; title Nonparametric Test to Compare Tumor Masses; class group; var mass; exact wilcoxon; run;

&
18

'

Wilcoxon rank sum test: SAS

&
19

' Wilcoxon rank sum test: SAS

&
20

' Wilcoxon rank sum test: R >ex8.2=read.table(F:/ST2137/lecdata/ex8 2.txt,header=T) >attach(ex8.2) >gp.a=mass[group==A] >gp.b=mass[group==B] > >wilcox.test(gp.a,gp.b) Wilcoxon rank sum test with continuity correction data: gp.a,gp.b W=18, p=value=0.06506 alternative hypothesis: true mu is not equal to 0 Warning message: cannot compute exact p-value with ties in:wilcox.test.default(gp.a,gp.b) &
21

'

Wilcoxon rank sum test: SPSS We create a numeric variableGroupnoto represent the groups Transform Recode into dierent variable... Move group to String variable panel Complete Output Variable window Change Click Old and New values and input the values Continue OK

&
22

' Wilcoxon rank sum test: SPSS

&
23

' Wilcoxon rank sum test: SPSS

&
24

'

Wilcoxon rank sum test: SPSS Analyze Nonparametric Tests Legacy Dialogs 2 Independent samples... Move mass to the Test Variable List and Groupno to the Grouping Variable Click Dene Groups... Input the values for the two groups Choose Mann-Whitney U in the Test-Type

&
25

' Wilcoxon rank sum test: SPSS

&
26

' Wilcoxon rank sum test: SPSS

&
27

'

Paired t-test (Related Samples) There are many situations where each subject receives both treatments Each subject could have been measured in the absence of drug and after receiving the drug The response time for the control and treatment groups would no longer be independent. 2-sample t-test cannot be used since the groups are no longer independent. A paired t-test can be used if the dierences between before and after treatments follow a normal distribution.

&
28

'

Paired t-test: SAS data ex8 3; inle F:\ST2137\lecdata\ex8 3.txt rstobs=2; input subject ctime ttime; proc ttest data=ex8 3; title A Paired t-test; paired ctime*ttime; run;

&
29

'

Paired t-test: SAS output

&
30

'

Paired t-test: R >ex8.3=read.table(F:/ST2137/lecdata/ex8 3.txt,header=T) >attach(ex8.3) >t.test(control,treatment,mu=0,paired=TRUE) Paired t-test data:control and treatment t=-4.3481,df=5,p-value=0.007372 alternative hypothesis: true dierence in means is not equal to 0 95 percent condence interval: -11.66874 -2.997923 sample estimates: mean of the dierences -7.333333 &
31

' Paired t-test: SPSS Analyze Compare Means Paired Sample T test... Highlight the two variables control and treatment simultaneously Move these 2 variables to Paired Variables panel OK

&
32

' Paired t-test: SPSS output

&
33

' 2 Related Samples: Nonparametric tests We apply the one sample nonparametric tests to the dierence of the paired observations. One sample nonparametric tests: Sign test Wilcoxon Signed rank test Example 8.4 Consider an experiment that each subject tries each of the two drugs The time span to pain relief is measured Subject 1 2 3 4 5 6 7 8 drug A 20 40 30 45 19 27 32 26 drug B 18 36 32 46 15 22 29 25 &
34

'

Sign test Consider example 8.4 The dierences in time span between Drug A and Drug B are Subject 1 2 3 4 5 6 7 8 Dierence 2 4 -2 -1 4 5 3 1 Sign + + - - + + + + Number of positive signs=6 and number of negative signs=2 If there was no dierence in the two drugs, we expect the number of positive signs(i.e. A<B) should be more or less the same as the number of negative signs( i.e. B<A)

&
35

'

Wilcoxon signed rank test The dierences in time span between Drug A and Drug B are Subject 1 2 3 4 5 6 7 8 Dierence 2 4 -2 -1 4 5 3 1 Rank of absolute 3.5 6.5 3.5 1.5 6.5 8 5 1.5 Sign + + - - + + + + Sum of positive ranks: 3.5+6.5+3.5+8+5+1.5=31 Sum of negative ranks: 3.5+1.5=5 If there was no dierence in the two drugs, we would expect the sum of the ranks of positive signs should be more or less the same as the ranks of negative signs.

&
36

'

2 Related Samples Nonparametric tests: SAS data ex8 4; inle F:\ST2137\lecdata\ex8 4.txt rstobs=2; input subject drug A drug B; di=drug A-drug B; proc univariate data=ex8 4; title Nonparametric Test for 2 related samples; var di; run;

&
37

'

2 Related Samples Nonparametric tests: SAS output

&
38

' 2 Related Samples Nonparametric tests: SAS output

&
39

' 2 Related Samples Nonparametric tests: R >ex8.4=read.table(F:/ST2137/lecdata/ex8 4.txt,header=T) >attach(ex8.4) >di=drug A-drug B >ncount=sum(sign(di[di>0]))#Get the number of positive signs >binom.test(ncount,length(di),0.5)#binom.test(obs x,n,H 0:p=0.5) Exact binomial test data: ncount and length(di) number of successes = 6, number of trials = 8, p-value = 0.2891 alternative hypothesis: true probability of success is not equal to 0.5 95 percent condence interval: 0.3491442 0.9681460 sample estimates: probability of success 0.75 &
40

'

2 Related Samples Nonparametric tests: R >wilcox.test(di) Wilcoxon signed rank test with continuity correction data: di V = 31, p-value = 0.07895 alternative hypothesis: true location is not equal to 0 Warning message: In wilcox.test.default(di) : cannot compute exact p-value with ties

&
41

'

2 Related Samples Nonparametric tests: SPSS Analyze Nonparametric Tests Legacy Dialogs 2 Related Samples... Highlight the two variables drug A and drug B simultaneously Move these 2 variables to Test Pair(s) List panel OK Check Wilcoxon and Sign OK

&
42

'

2 Related Samples Nonparametric tests: SAS output

&
43

' 2 Related Samples Nonparametric tests: SPSS output

&
44

' 2 Related Samples Nonparametric tests: SPSS output

&
45