EXERCISES FOR EPSY 640

1 Select a data-based article from a journal in your area of interest; list the variables measured in the article and their level of measurement. Comment on the appropriateness of the level of measurement selected by the authors. Append the article to your exercise.

3a. Glass & Hopkins Chapter 6, Problems 1, 2, 5, 6, extra credit: 13

3b. For either the BASC or TEA datasets, select 3 variables; correlate them and graph using SPSS their scatterplots. Identify for one scatterplot one or more possible outliers; circle them on the plot and recalculate the correlation with the datapoint excluded. Compare the correlation values with and without the datapoint(s).

4a. . Glass & Hopkins Chapter 10, Problems 1, 2, 4, 5

4b.

1. For the following population sizes and expected differences d, compute the sample

size required that would locate the population mean. Assume p=.05 for error.

a. Population =200, d=.3SD ANS_________

b. Population =500, d= 1SD ANS_________

c. Population = 5000, d= .1SD ANS________

d. Population =1,000,000, d = .5SD ANS_______

2. Compute the standard error of the mean for the following conditions of population and sample size, and use the fpc. Assume the standard deviation of scores is 10 for each case:

a. Population = 500, sample=200 ANS___________

b. Population = 1000, sample = 400 ANS___________

c. Population = 50, sample=40 ANS___________

3. For each answer in 2 above, construct a 95% confidence interval around a mean of 50. Theoretically, which would you expect to have the smallest confidence interval? Why?

a. (_____________, ____________)

b. (_____________, ____________)

c. (_____________, ____________)

Smallest?_______

Why?_____________________________________________________________
__________________________________________________________________
__________________________________________________________________

4. For the following data, construct the estimate of the mean and the estimated standard error of the mean:

SUBPOPULATION N_I n_i X. s_i

A 77 40 10 5

B 229 90 11 6

C 738 100 12 7

How does this differ from the result in Table 4.3?

5. Decide for each case if it is a Stratified (S), Multistage (M) or Cluster (C) sampling procedure.

___ a. The 50 largest cities and all surrounding counties are used as the basis for sampling elementary schools.

___ b. US Census Region population size is used to select proportionately a total of 1000 small business owners taken from lists available from the US Dept. of Commerce.

___ c. All eligible school districts in Texas are sampled based on US Dept. of Education criteria for Title 1 using a 40% sample; of the districts sampled, half the secondary schools in each district is randomly sampled; for each school sampled, one full time science teacher is sampled, if there is one.

___ d. A sample of American Psychological Association members with licensure is sampled based on gender, region of the country where they reside, and years of practice (over 10, under 10)

6. For each situation decide which threat (Selection =S, Mortality=M) might apply to the hypothesis mentioned, if any (None=N).

___ a. The lowest achieving children in each 3^rd grade in the district's schools are chosen for a remediation program.

___ b. A marketing company is able to contact 20% of the visitors who signed the guest book at Big Bend National Park.

___ c. An Internet questionnaire is answered by 500 persons who accessed the site over a two week period.

___ d. A systematic sample of physicians listed in the Houston telephone directory is taken and all those with working numbers are contacted.

5a. . Glass & Hopkins Chapter 11, Problems 2 (assume s=4 instead of 3), 9, 10

5b. For either the TEA or BASC dataset, select a variable and obtain its mean and SD,

which you can assume are the population values. Select a subgroup of 100 from among the various other indicators (such as males in the BASC or low performing schools in the TEA set, and test that the mean for that group is equal to the population mean.

6. Glass & Hopkins Chapter 12, Problems 1 (assume s₁²=6.1, s₂²= 5.5), 4 (assume r=.7), #7

(assume n=27 for both groups)

7a. Glass & Hopkins Chapter 13, Problems 1 (assume True = 65, False=45), 5, 10 (add 3 to

each cell).

7b. For either the TEA or BASC dataset, select two variables to correlate with each other for

two distinct samples (such as boys and girls, or large and small schools). Test the

hypothesis that the correlations are equal to each other.

8a. Glass & Hopkins Chapter 8, Problem 6

9a. Glass & Hopkins Chapter 15, Problems 1, 4, 5, 12 (use SPSS-PC for power calculations)

9b. Glass & Hopkins Chapter 17, Problems 1, 2, 4