Exam 10: Hypothesis Testing, Effect Size, and and Confidence Intervals: Two-Sample Designs

In which of the following would an increase from a smaller number to a larger number produce a more powerful statistical test?

In the simple experiment described by your text, the groups were made equivalent with random assignment.

Identify each experiment below as an independent samples design or a paired samples design. -On the basis of a pretest on knowledge of foreign governments, each student in the 9:00 class was matched with a student in the 10:30 class. The 9:00 class used a "participant approach" to the study of politics and the 10:30 class used the "Heidleburg method." At the end of the term the same test was given to both classes.

In an independent samples design, the null hypothesis is that

The formula for power is 1 - $\mu$ .

Data Set 10-2: The IQs of a group of special education students were compared to the IQs of their siblings who were not in special education. The following data were obtained. Special Ed Not Special Ed 70 104 65 99 67 95 -The numerical value of the effect size index for Data Set 10-2 is

Data Set 10-7: In that long standing feud between the Hatfields and the McCoys, Old Doc Sawbones has been keeping score. During the last 16 years he noted in his feudian notebook a "degree of injury" measure each time he patched up, sewed up, or made a coroner's report on anyone in the two families. Summary statistics were: Hatfields McCoys 41 34 6 8 N 16 16 -Which of the following conclusions is proper for Data Set 10-7?

In an independent samples design the Hatfields had a mean score of 25 against a mean score of 26 for the McCoys. Low scores mean better performance. The researcher ran a two-tailed test with $\alpha$ at .05. A t value of 1.99 was found.

The denominator of a paired-samples t test is a standard error of a difference.

Data Set 10-8: Drugs that reduce emotionality make big money for drug companies. These drugs undergo extensive testing, some of which is done with animals. One common test of emotionality in rats is the number of fecal boluses that are dropped when an animal is left for 10 minutes in an open field (a 4' x 4' space). Greater emotionality is indicated by more boluses. Imagine that two drugs, aloe and vera were compared. Five rats were given aloe and five were given vera. The numbers below represent the number of boluses dropped in the open field. Aloe Vera 4 2 12 9 7 6 10 7 6 4 -The effect size index for Data Set 10-8 is

The t test was developed as a way to analyze flavors in taste experiments.

For each description, identify a) the independent variable, b) the dependent variable, and c) whether a paired or an independent-samples design is being described. -Brad gathered data on the effect of authority among college students. He had 24 students read and give a rating of quality on a scale of 1-7 to an abstract of a scientific article. For 12 students, the apparent author of the abstract was a person with a title of authority. For the other 12, the apparent author's title was not one of authority.a) Independent variable _______________b) Dependent variable ______________c) Design______________d) df ______________

Identify the design of each experiment below as paired samples or independent samples. Give the degrees of freedom in each case. -Germination rates from 10 packages of sugar peas from Burpees were compared to 10 packages from Starks.

Indicate whether each design below is independent samples or paired samples and where possible, give the degrees of freedom. -Thirty-two freshmen had applied for the sophomore honors course, but only 16 could be accepted. The college dean arranged the 32 students' GPA from highest to lowest. Among the top two a coin was flipped, which designated one to be assigned to the course. Among the second highest two, the procedure was repeated and so on until the last two. At graduation the dean compared the "winners" to the "losers" on overall GPA.

For a normally distributed set of scores, it is often best to use the design that has the most power. Which design has more power?

Researchers control the probability of a Type I error in their experiment by setting the value of

Which conclusion is not appropriate when using null hypothesis statistical testing (NHST)?

"There is no difference in the two populations"is a statement of

As you may know, there has been bad blood between the Montague family and the Capulet family for a good while. In this modern day, resolution can be achieved by using a psychological test. In the test of "propensity to fall in love,"the mean of the 6 Montagues was 54 and the mean of the 10 Capulets was 64. (Italian norms show a national average of 100.) When a statistician compared the families with a t test, a value of 2.13 was obtained. If you adopt an $\alpha$ level of .05 (two-tailed test), you should conclude that the Capulets are

The two kinds of t tests described in Chapter 10 were power samples and paired samples.