lecture 11 - between subjects t-test

Question 1

when to use a between subjects t-test

Answer 1

Between-subjects tests of difference for two condition experiments with interval data

Question 2

between subjects t-test in the within subjects t-test

Answer 2

the focus was on difference scores - within subjects equation

t = D-bar / SM

sm = S/√N

D = mean difference
SM = Standard error of the mean difference

Question 3

features of between subjects t-test

Answer 3

In the between subjects t test we can’t focus on differences scores because the data aren’t paired….
And because the data aren’t in related pairs, we expected more noisy variability because we aren’t controlling for individual differences….

Question 4

between- subjects t-test

Answer 4

t = Y-bar1 - Y-bar2 / Sm

Y-bar1 - Y-bar2 = difference of means
Sm = standard error of the difference between the means

Sm = √Sp^2/ N1 + Sp^2/N2

Sp^2 = (N1 - 1) S1^2 + (N2 - 1) S2^2/ N1 + N2 - 2

–> how you find the pooled estimate of variance.

variability gets more complicated with more components - more things go into the specific of variability

the equation is conceptually doing same as within subjects t-test but still specifying the variability between the means

Question 5

Comparing the within and between-subjects t-test results - df, t value, value

Answer 5

CONCLUSION: “Sig O happiness was significantly affected by flower type in a within-subjects design, that is, happiness was significantly higher for tulips than roses,
t(5) = 2.825, p < 0.05.”
CONCLUSION: “Sig O happiness was not significantly affected by flower type in a between-subjects design, that is, happiness was not significantly different for tulips and roses,
t(10) = 1.336, p = 0.211.”
Note. In reality you would almost never do a within and between subject t test on exactly the same numbers because you should know(!) whether the data came from a within or between subjects design…..

Question 6

how to see if significant

Answer 6

use a critical value table
df = N1 + N2 - 2
find right df and go to 0.05 significance

t value needs to be bigger than critical value to be significant

Question 7

reporting the t-statistic

Answer 7

t (10) = 1.34, p >0.05, two-tailed

t = type of test
10 = degrees of freedom
1.34 = critical value
p > 0.05 = probability value
two-tailed = type of hypothesis

Question 8

Comparing the wilcoxon rank-sum test and the between-subjects t-test results

Answer 8

CONCLUSION: “Sig O happiness was not significantly affected by flower type in a between subject design, that is, happiness was not significantly different for tulips and roses,
WS(6, 6) = 2.825, p = 0.301, two-tailed.”
CONCLUSION: “Sig O happiness was not significantly affected by flower type in a between-subjects design, that is, happiness was not significantly different for tulips and roses,
t(10) = 1.336, p = 0.211, two-tailed.”
Am I worried about the assumptions of the t-test here?
Does it matter if the data are interval or just ordinal?
Note. In reality you would almost never do a within and between subject t test on exactly the same numbers because you should know(!) whether the data came from a within or between subjects design…..

Question 9

look at SPSS output of t-test