Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
29 Cards in this Set
- Front
- Back
Independent Sample T-test
|
We collect data from two independent samples
(control and experimental group) Want to know if there's a difference between the sample means |
|
Assumptions of Test
|
the populations standard deviation and variance are equivalent.
(group 1; SD and V = group 2; SD and V) |
|
Level of Measurement (independent samples t-test)
|
WE GO BY THE GROUP LEVEL
Our sample is not dependent on each other, so we cannot go by the individual level |
|
Sampling distribution of the mean difference
|
The mean is 0
SD = ? |
|
Pooled Variance
|
Done if the sample means are not equivalent
|
|
Degrees of Freedom
|
Pooled as well
sample 1 + sample 2 - 2 = pooled DF this gives us an estimate of their combined degrees of freedom |
|
Null and Alternative Hypothesis
|
Null: Mu1 = Mu2
(we assume the two interdependent samples are equivalent) Alt: Mu1 (doesn't equal) Mu2 (they may also be proven not to be equivalent) |
|
Standard error of the mean difference
|
The pooled variance is divided by sample size 1 and 2, then added together and square rooted.
The variance square rooted gives us the Standard deviation of the two, or the standard error of the mean difference. |
|
T-value for mean difference
|
subtract the means from each other, and divide that by the standard error of the mean difference.
This gives us the value that represents where we fall on the spectrum of the sample distribution of the mean difference (Mean = 0, SD = ?) |
|
critical value
|
We give a 5% chance on a two tailed test as to whether were getting this wrong. If this value passes this critical value it means it a reasonable enough difference away from the (mean of the sample distribution of the mean difference (0)) to say that there was a difference between the two groups.
|
|
T-table
|
using the degrees of freedom you find out what the critical value is.
|
|
Confidence Intervals
|
Sample mean + t-crit (standard error)
Dif of Sample means - t crit (standard error) You take these two values, and if the mean of the sample distribution of the mean difference (0) is not within those values, you can then reject the null. |
|
DATA (SAS)
|
name of data set
|
|
INPUT (SAS)
|
gives variable names
|
|
CARDS; put in data ; (SAS)
|
gives data on card sheets next to each other
|
|
PROC MEANS DATA = data_02;
VAR cog1 cog2 diff; RUN; (SAS) |
looks at the means of your variables
|
|
PROC TTEST DATA = data_02;
PAIRED cog1 * cog2; RUN; (SAS) |
dependent samples t-test
mean difference and appropriate standard error immediately |
|
PROC TTEST DATA = data_02;
VAR diff; RUN; (SAS) |
One Sample t-test
compares difference to 0 |
|
Pr > t
(SAS) |
significance value, probability of observing value greater than t-value
(if its greater than .05 we reject null) |
|
$ (SAS, AND R)
|
tells SAS its a categorical variable
|
|
PROC MEANS DATA = ex2;
CLASS group; VAR score; RUN; (SAS) |
calculates means for both groups in independent samples t-test
|
|
PROC TTEST DATA = ex2;
CLASS group; VAR score; RUN; (SAS) |
tells you grouping variable and outcome variable score
tells you the results of independent samples ttest |
|
cog1 < -- c(data) (R)
|
arranges data as column
|
|
diff (cog1-cog2) (R)
|
gives you difference
|
|
summary diff (R)
|
summary of dif
|
|
sd diff (R)
|
standard d of dif
|
|
t.test(diff, mu=0) (R)
|
one sample t-test
|
|
t.test(cog2,cog1, paired=true) (R)
|
dependent samples t-test
|
|
t.test(score~group) (R)
|
Independent sample t-test
|