PSY-201: Quiz 5 - t-tests

In a t test for dependent means, the mean change score is 12 and the standard deviation of the distribution of means is 6. What is the t score?

2

In a t test for dependent means, 10 participants each are tested twice. This makes a total of 10 before scores and 10 after scores. What are the degrees of freedom?

9

Why does it make sense that the standard error of the difference between independent means can be estimated by summing the standard error estimated separately from each of the two groups? That is, why does it make sense that the denominator of the independent-samples t-test requires adding two things together?

Because we have two sources of sampling error.

True or False? For a hypothesis test with an independent-samples t, the larger the two sample variances are, the greater the likelihood that you will reject the null hypothesis.

False

True or False? You will be much more likely to detect a treatment effect when the estimated standard error is large.

False

True or False? Degrees of freedom for the independent-measures t statistic can be computed as df = n1 + n2 - 2.

True

True or False? For a two-tailed hypothesis with an independent-samples t statistic using two samples, each with n = 10 scores, the boundaries for the critical region using a = .05 are t = +/- 2.101.

True

True or False? For an independent-measures t statistic, the magnitude of the estimated standard error is directly related to the sample variances (the bigger the variances, the bigger the error.)

True

True or False? In an independent-samples t-test, the only way to obtain a value of zero for the independent measures t statistic is when M1 = M2.

True

True or False? The estimated standard error for the independent-measures t statistic combines the error from each of the two separate samples.

True

True or False? The larger the difference between the two sample means, the larger the value for the independent-measures t statistic. (Assume all other factors are held constant.)

True

True or False? The null hypothesis for the independent-samples t test states that there is no difference between the two population means.

True

How does the summation referred to above (in Question 1) impact the power of the independent-samples ttest?

We have lower statistical power (to reject the null hypothesis) as a result of the increased sampling error.

difference scores

calculate difference between first and second score - e. g. Difference = Before - After - Base subsequent analysis on difference scores

In the formula for related-samples t-tests, the scores analyzed are ___________ _________.

difference scores

t-tests

hypothesis-testing procedure in which the population variance is unknown; it compares t scores from a sample to a comparison distribution called a t distribution.

All else equal, relative to a(n) ________________ t-test, a(n) _______________ t-test will often confer greater statistical power to reject the null hypothesis.

independent-samples, related-samples

MD and SDD =

mean and standard deviation of difference scores, respectively.

The ___________ _______ is assumed to be 0 when you do a t test using difference scores.

population mean

In a study of the effects of a particular drug on creativity, 15 participants took a creativity test twice— once in a normal mental state, and once under the influence of the drug. You would use a t test for __________ _________.

related samples

related samples

the same participants give us data on two measures - e. g. before and after treatment, aggressive responses before video and aggressive responses after

Advantages of Related Samples

• Eliminate subject-to-subject variability • Control for extraneous variables • Need fewer subjects

Disadvantages of Related Samples

• Order effects • Subjects no longer naïve • Sometimes not logically possible

Two Major Assumptions of Related Samples

1. Both groups are sampled from populations with the same variance "homogeneity of variance" 2. Both groups are sampled from populations with normal distributions, or at least that sampling distributions are normal (large samples help here)