Week Seven Flashcards
When do you use a t test
When variables aren’t run on a continuum and thus cannot use correlation analysis
Variables must be radio or interval
What do t tests look at
Group differences: grouping
Only two groups
Considers all of the scores of each group, constructs a mean score of the groups, can construct a confidence interval around the mean score. And see if those two intervals overlap.
What is an independent (unrelated) t test?
Samples are independent
Compare the means of two groups
(Descriptive statistics & error bar)
Calculate the t statistics and effect size
What is the t statistic
The t statistic: a ratio of the between group variance and the within group variance.
The larger the between group variance (compared to within groups) the larger the t statistic.
What are degrees of freedom
The number of degrees of freedom is the number of values in the final calculation of a statistic that are free to vary
What is the formula for degrees of freedom? (On the exam!)
DF = N (first group sample number) - 1 + N (second group sample number) - 1
15 horror gamers - 1 = 14
15 fantasy gamers - 1 = 14
14 + 14 = 28
DF = 28
What is the t value and p value
T value: larger the t value, the more likely it is to be significant
P value: probability of obtaining t values due to sampling errors (test of null hypothesis)
If p is high, t is low; if t is low to 0 there is no difference statistically. P threshold is .5
Which row of the levins test do you interpret
The second row if the data is statistically significant so p is less than .5
If p value is not below .5 the group variance are the same so you interpret the top row of the t values
What factors affect the t value
The actual obtained difference
The magnitude of the sample variance
The sample size (N)
The significance level (a)
Whether the test is a one or two tailed test
Effect size (cohens d )
What is cohens d
The effect size: it indicates the degree that differences between the groups on the DV can be attributed to the IV
How to calculate d
(Group one mean - group two mean) = first half of equation
(Group one standard dev + group two standard dev) divided by 2 = second part of equation
Take group means and divided by mean as
Group Means/Mean SD
Only report d when results are significant
What are the effect sizes of d
- 2 = small effect, 85% overlap
- 5 = medium effect
- 8 = large effect
What are the assumptions of t-tests
Continuous data: independent groups (for independent samples t test)
Normally distributed - checked skewness or kurtosis
Test required normality of population and not necessarily the samples drawn from those
Homogenous variables
Samples are assumed to be drawn from the same population so they are assumed to have the same variance
Describe paired (related) samples t test
Paired samples (related samples, repeated samples, matched samples)
An experimental design in which the same participants is observed under more than one condition
What is a one sample t test
Used to compare the mean of a group with an known mean
