RESEARCH METHODS, FINAL EXAMINATION (Fall 2003) (SAMPLE 1)

1. (10 points) A state department head is concerned over implementing a union agreement that mandates that years of service will be the solely criterion for determining public employees’ salaries. She is the supervisor of twenty employees, who have the following years of service and salaries:

EMPLOYEE	YEARS OF SERVICE	SALARY
A	17	$43,000
B	2	$13,000
C	11	$32,000
D	17	$45,000
E	2	$15,000
F	12	$34,000
G	7	$24,000
H	14	$38,000
I	4	$17,000
J	15	$39,000
K	8	$26,000
L	15	$41,000
M	5	$20,000
N	16	$42,000
O	4	$19,000
P	10	$29,000
Q	20	$50,000
R	10	$31,000
S	18	$46,000
T	19	$48,000

Plot each of these points (employees) on graph paper, depicting years of service as the independent variable and salary as the dependent variable. Draw the best fitting straight line through these points, also known as the regression line.

What is the Y-intercept of this regression line?

Assume that the b value or slope is 2,000. What does that value mean? Specifically, each year of service is worth how many dollars in salary?

What is the starting salary? Specifically, what would be the salary for someone having zero years of experience?

If the department head hired someone with thirty years of experience at a comparable organization, what would you predict that that person should be paid?

2. (15 points) The director of the state department of public health is concerned over the negative health effects of daily drinking of alcoholic beverages. He wishes to conduct a radio advertising campaign warning against excessive drinking, but lacks information about what demographic groups are highest in drinking. He contracts out to a state university to conduct a public opinion poll that seeks to identify what demographic groups are related to drinking alcholic beverages. (Note: The following multiple regression results are real world data drawn from the 1984 Mississippi Poll.)

The dependent variable is frequency of drinking alcholic beverages, and it ranges in coding from 1 for never drink to 6 for drink every day.

The independent variables and their codes are:

Age, coded from the youngest age of 18 to the highest age of 89.

Education, coded from a low of 1 for 4^th grade or less to a high of 7 for graduate work.

Income, coded from a low of 1 for under $5,000 to a high of 7 for over $30,000.

Sex, coded as 1 for male and 2 for female.

Race, coded as 1 for white and 2 for black.

The adjusted R squared for the multiple regression results was 20%.

Predictor	Beta	Significance Level
Age	-.24	.001
Education	.09	.039
Income	.07	.050
Sex	-.29	.001
Race	.01	.840

How good a job are these five predictors doing in explaining frequency of drinking alcholic beverages? That is, what percentage of the variance in drinking is explained by these five predictors?

List all of the predictors that are statistically significant at the .05 level or lower.

What predictor is most important in explaining frequency of drinking? What predictor is second in importance? Which is third in importance? What is fourth in importance?

Which of the following of each pairs of categories is higher in frequency of drinking alcholic beverages?

The young, or the old?

High school dropouts, or the college educated?

Lower income people, or higher income people?

Men, or women?

3. (10 points) Draw up diagrams that show the Classical Experimental Design, and briefly explain how it can be used to test the effectiveness of a stimulus. In your diagrams, include the following terms, and use them as frequently as required: post-test; pre-test; stimulus; control group; experimental group. Also, mention comparisons of changes in test scores.

4. (10 points) In the Classical Experimental Design, what are ten sources of internal invalidity? List these sources using either the correct terminology used in class, or by explaining what the source is and how it operates.

5. (5 points) What is the definition of external invalidity problems? Now, list four sources of external invalidity. List these sources using either the correct terminology used in class, or by explaining what the source is and how it operates.

6. (5 points) Regarding the two types of quasi-experimental designs discussed in class, which type (if either) has a control group?

Time Series Design
Control Series Design
Both
None of the above

7. (5 points) Regarding the two types of quasi-experimental designs discussed in class, which type (if either) has multiple pre-tests (or a time series before the stimulus) and multiple post-tests (or a time series after the stimulus)?

Time Series Design
Control Series Design
Both

D. None of the above

8. (5 points) What are the four components of Informed Consent? Use the proper terminology, or explain how the component operates.

9. (5 points) What are the key differences between Anonymity and Confidentiality?

10. (10 points) What are three problems with obtrusive measures, and what are seven types of unobtrusive measures, as discussed in class? Use the proper terminology, or explain what these problems or types are and how they operate.

11. (20 points) Circle the correct response for each of the five readings. Each sub-question is worth four points.

What methodology was employed in the Justifications for Ratings of Old and Young Job Applicants article, authored by Finkelstein, Higgins, and Clancy?

Multiple regression
Experimental design
Quasi-experimental design
Content analysis
None of the above

What methodology was employed in the Gang Suppression Through Saturation Patrol, Aggressive Curfew, and Truancy Enforcement article, authored by Fritsch, Caeti, and Taylor?

Multiple regression
Experimental design
Quasi-experimental design
Content analysis
None of the above

C. Which of the following conclusions was reached in the Participative Management and Job Satisfaction article, authored by Soonhee Kim?

Participative management that incorporates effective supervisory communications has no impact on employees’ job satisfaction.
Participative management that incorporates effective supervisory communications reduces employees’ job satisfaction, as it increases their work load and stress levels.
Participative management that incorporates effective supervisory communications reduces employees’ job satisfaction, as it increases the number of hours that they work per week
Participative management that incorporates effective supervisory communications can enhance employees’ job satisfaction.

D. In the article on The Sources of Ethical Decision Making for Individuals in the Public Sector, authored by Montgomery Van Wart, the Code of Ethics of what professional association was examined?

1. American Political Science Association

2. American Society for Public Administration

3. American Association of Opinion Researchers

4. National Association of Governors

5. National Association of Public Managers

In the article authored by Lawrence M. Mead, what is the acronym for the employment program that they studied, which was intended to move welfare recipients into work or training?

JOBS
CETA
AFDC
WORK
OPPORTUNITY

1. (10 points) The Department of Corrections in a southern state is concerned about the growing size of its juvenile detention (prison) population. The agency head hypothesizes that the absence of early intervention programs (such as wilderness camps, midnight basketball, summer jobs programs for at-risk youth) contributes to a greater tendency of teenagers to commit crimes that are serious enough to warrant imprisonment. She hires a researcher to test this hypothesis. The researcher conducts the following study of the percentage of the male teenage population that is in prison in each state (dependent variable), and the percentage of a state's juveniles that has ever participated in early intervention programs (independent variable).

COUNTY	PERCENT OF JUVENILE MALE TEEN POPULATION THAT IS IN PRISON (DEPENDENT VAR)	PERCENT OF STATE'S JUVENILE POPULATION THAT HAS PARTICIPATED IN EARLY INTERVENTION PROGRAMS (INDEPENDENT VARIABLE)
A	5	40
B	9	33
C	14	21
D	20	9
E	33	2

The Department also conducted one public opinion poll of 1000 juvenile males in state C, and another opinion poll of 1000 juvenile males in state D, asking teenagers whether they had ever participated in an early intervention program, and whether they had ever been in prison or detention. They made the following findings, from which you should construct two tables depicting the possible relationships between early intervention programs and imprisoned juveniles. Remember, the hypothesis is that the absence of early intervention programs (independent variable) is contributing to serious juvenile crime that requires imprisonment (dependent variable).

STATE C. Of the 1000 male teens sampled: 29 participated in early intervention programs and had served in prison; 181 participated in early intervention programs and had never served in prison; 111 had not participated in early intervention programs, and had served in prison; 679 had not participated in early intervention programs, and had never served in prison.

STATE D. Of the 1000 male teens sampled: 18 participated in early intervention programs and had served in prison; 72 participated in early intervention programs and had never served in prison; 162 had not participated in early intervention programs, and had served in prison; 648 had not participated in early intervention programs, and had never served in prison.

If the department had not conducted the public opinion polls of the two states, what conclusion do you think they would have drawn on the basis of the aggregate data about the effect or non-effect of early intervention programs on male juvenile prison rates. Explain.

What do the two opinion polls demonstrate about the possible relationship between early intervention programs and the male juvenile prison rates? In each state, is there a relationship between early intervention programs and the prison rate? If so, what is the direction and magnitude of the relationships?

If the department had relied solely on aggregate data, do you think they would have made the ecological fallacy? Explain and defend your answer, and include a definition of what the ecological fallacy is?

2. (10 points) The state Department of Public Safety in a western state wants to encourage automobile drivers to buckle their seatbelts. In the past three years, the percentage of highway fatalities in which the victims had not been wearing their seatbelts had increased. It plans to run an aggressive television ad campaign that seeks to educate viewers about the dangers of not buckling up by graphically depicting injuries and deaths from automobile accidents where occupants had not buckled their seat belts. Ads would show broken windshields and blood-covered car hoods, re-enactments of overturning cars throwing unsecured passengers out of windows, and interviews with quadriplegics whose injuries were traceable to not wearing seat belts. The agency head is interested in scientifically determining how effective the proposed ads are in changing people's perceptions of the costs of driving without seatbelts, and their behavior in terms of driving with or without seatbelts. If you were the research advisor to the agency head, what panel design would you propose to test the effectiveness of the upcoming ad campaign, and discuss fully all aspects of that design (such as specific wording of questionnaire items, and how and when respondents would be sampled from the state's entire population)? From the discussion in class regarding the problems of panel designs in general, what problems might face your specific panel design, and explain fully?

3. (10 points) A school district superintendent is concerned about the growing incidents of illegal drug use by students. She pays for the production of a series of films and invited guest speakers, designed to increase students' perceptions of the illegality and harmful effects of taking illegal drugs. Films might graphically depict the corpse of a heroin overdose victim, and a day-in-the-life of a homeless drug addict. Speakers might include prisoners sentenced to life terms for murder, committed in the course of an armed robbery to obtain money for crack cocaine, who bring pictures depicting their small cells that lack radio and televisions. The superintendent is interested in whether these films and speakers have desirable effects on the attitudes and behavior of school children towards taking illegal drugs.

Devise a classical experimental design for this project, and discuss all aspects of this design in detail. What hypothetical results would demonstrate the effectiveness of this project, and what results would demonstrate ineffectiveness?

How specifically would this experimental design help you to eliminate possible sources of internal and external invalidity, and explain? Mention as many sources as you can recall, and give examples.

4. (10 points) You are employed by the state of Alabama as a policy analyst in 1999. In 1995 the state begins to require that prisoners work in chain gangs, picking up litter along the state's highways. Proponents of visible chain gangs argue that they will discourage prisoners from committing future crimes after being released from prison, and that non-incarcerated citizens seeing them will also be discouraged from committing crimes. Therefore, the crime rate in the state should be reduced. Opponents of visible chain gangs argue that they are a gimmick to show voters that legislators are being "tough on crime," and that they fail to deal with the real sources of crime such as poverty, limited job opportunities, and social injustice. Opponents argue that the crime rate will be unaffected by the chain gang law.

Devise the best possible quasi-experimental design in order to test the positive and negative effects of the chain gang law, and discuss all aspects of this design. How would this design accomplish the task of testing the effects of chain gangs, and ruling out other possible explanations for changes in the crime rate, and explain?

Draw a number of graphs including all of the relevant data points that demonstrate four different possible outcomes of the annual tests (such as beneficial or ineffectual results involving the chain gang law). How could you tell that the chain gang law was helpful or ineffective in each case, and defend your answer?

5. (10 points) How should one demonstrate a concern for ethics when conducting a telephone survey on the sensitive issue of domestic violence (specifically, "wife beating")? In your answer, address the different aspects of "informed consent". What are the differences between anonymity and confidentiality, and explain with examples? How can you guarantee confidentiality in your telephone survey, and explain fully?

6. (10 points) The agency head of a state Highway Patrol Department is concerned over whether white state troopers have negative attitudes toward African-Americans in general, which may result in discriminatory treatment towards some African-American civilians.

Devise at least five unobtrusive indicators designed to measure possible negative trooper attitudes toward African-Americans in general, and discuss each of the five clearly and in detail.

What possible validity problems exist with each of these indicators, and elaborate fully?

7. (5 points) In the Nachmias chapter on Observational Methods, the authors discuss the types of behavior that social scientists can observe, and do so with reference to the variables "frustration" and "aggression." What are the four types of behavior that they discuss, and briefly explain what each one involves?

8. (5 points) In the Nachmias chapter on Observational Methods, the authors discuss the sources of bias in laboratory experiments. What are the three sources of bias, and briefly explain what each of them entails?

9. (10 points) In the Nachmias chapter on Qualitative Research, the authors discuss the seven stages in the practice of field research. What are these seven stages, and briefly explain what each of them entails?

10. (10 points) Four outside readings illustrated the techniques of aggregate data, experimental designs, panel designs, and quasi-experimental designs. By authors' last names they were:

b) Huddy and Terkildsen's "Gender Stereotypes and the Perception of Male and Female Candidates";

c) Jennings' "Residues of a Movement: The Aging of the American Protest Generation";

d) Tatalovich's "Role of the Mississippi Electorate in a Changing Political System".

Explain fully how each of these four articles employed one of the four research designs mentioned above. What was each article's most important findings, and discuss?

11. (15 points) Social scientists are interested in what kinds of people are more supportive and more opposed to an African-American runnning for President of the United States. In the 1994 Mississippi Poll, MSU researchers asked 620 Mississippians who they would vote for if Colin Powell were the Republican nominee in 1996 and Clinton were the Democratic nominee. Some might hypothesize that race, party identification, ideology, and reaction to Clinton's job performance would be related to how people would vote. Perhaps Powell would be most supported by Republicans, conservatives, those rating Clinton as poor, and blacks. An analysis of the actual poll results leads to the following multiple regression equation:

X4 is party identification (coded: 1 = Democrat; 2 = Independent; 3 = Republican)

A) What percentage of the variance in the presidential vote is being explaining by these four independent variables, and how did you determine this?

B) What independent variable is most important in the presidential vote, and what variable is least important, and how did you determine this?

C) How do you think that a moderate black who rated Clinton as poor and who was a Republican would vote? Calculate that person's Y score and show your work.

D) How do think that a liberal white who rated Clinton as excellent and who was a Democrat would vote? Calculate that person's Y score and show your work.

E) In words, what do these actual poll results tell you about what really influences how people vote in Mississippi? What are the directions of the non-zero relationships between each predictor and the vote? Does Powell's race hurt him among whites, and explain?

12. (10 points) Under one state's school choice plan, a school district contracts with a private company to operate its school system. The private company decides to award teacher pay raises based on the performance of the students. In English classes, for example, the higher the scores of the students on a school-wide writing examination for those in a particular class, the higher the pay raise received by the teacher of that class. A legislative investigating committee chairwoman believes that the company is disregarding performance in the allocating of pay raises, and is simply rewarding friends of the company's owner. She obtains information from the 18 mathematics teachers in the district regarding the average scores of their students on a group mathematics test, and the dollar amounts that the teachers received in raises. This hypothetical information follows (each line is a separate teacher):

Graph the relationship between the students' average test score in each class and the pay raise for the teacher of that class. What are the Y-intercept and unstandardized regression coefficient values, and what do they tell you about these data points? If a teacher taught a class where all of the students received 100 grades, what would you predict that her pay raise would be, and show how you devised a regression equation to predict her pay raise?