Exam 14: Students T Test for Correlated and Independent Groups

Which of the following are assumptions underlying the use of the t test for independent groups?

Compare and contrast the independent groups design with the correlated groups design.

With the t test for correlated groups, the difference between the sample means always equals the mean of the difference scores.

An automotive engineer wishes to determine the fuel efficiency of a test engine compared to the standard production model. The following data (miles/gallon) were obtained under standard conditions. Assume independence between scores in each group.

Define t test for correlated groups.

As an owner of a computer component factory you want to buy a new machine to make integrated circuits. You can buy one of two types of machines. You obviously want to buy the better machine (we'll assume they are comparable in price and reliability). You rent both machines for 8 days and make sample runs on both machines. The following data are the number of defective components per 1000 on each day.

The use of the t test for independent groups assumes _________.

In a correlated t test, if the independent variable has no effect, the sample difference scores are a random sample from a population where the mean difference score ( µ D ) equals _________.

The t test for independent groups assumes homogeneity of variance.

For the independent groups design, increasing sample variance decreases power.

Define confidence-interval approach.

How does the variability of the sample data affect the power of the t test? Explain, using the appropriate equations.

A worker in a neighborhood clinic wishes to assess the impact of showing an educational film on patient compliance in taking an antihypertension medication. The diastolic blood pressure (BP) is used as the dependent variable. During the study the medication dosage was kept constant. Blood pressure was measured one week prior to the film, then the film was shown, and the blood pressure measured again 3 weeks later. The data are shown below:

A major advantage to using a two condition experiment (e.g. control and experimental groups) is _________.

Which of the following increase(s) the power of the t test?

When a result is statistically significant, that guarantees that the result is an important one.

If r between the scores of both groups is high and negative then the t test for correlated samples is not a powerful statistical test.

A professor of women's studies is interested in determining if stress affects the menstrual cycle. Ten women are randomly sampled for an experiment and randomly divided into two groups. One of the groups is subjected to high stress for two months while the other lives in a relatively stress-free environment. The professor measures the menstrual cycle (in days) of each woman during the second month. The following data are obtained.

A school psychologist is interested in determining if children with attention deficit disorder (ADD) learn better if English literature is read to them rather than having them read the material themselves. A random sample of 10 sixth graders with ADD is selected and divided into two groups of 5 each. One of the groups has a story read to them (Listening Group) and the other reads the story themselves (Reading Group). A quiz on the story is given after each group has finished reading or hearing the story. The following scores were obtained with 20 being a perfect score. Assume there is a real effect. The size of the effect, using equals _________.

A physical therapist wants to know if football players (guards in this experiment) recover full strength following injuries. A group of injury free guards were given a strength test as were a group of guards who had finished a rehabilitation program following an injury. The groups were matched in height and weight. The following strength ratings were recorded. What do you conclude using a = 0.01 2 tail ?