Speaking Notes

PADM 5502

September 24, 2007

Dr. Neubauer

WHERE WE ARE:

• The first installment of the research prototype is due October 1. 
• Our midterm exam will be October 8.  It will include questions related to all the things we are discussing in class. This maps approximately to the first 7 or 8 chapters of our textbook.

Chapter 5 -- Sampling REVIEW

1) A ______________________ derived from a random sample is hopefully a good

estimate of a ____________________________ in the population.

2) What is required in order to really have a random sample?

3) Say you select undergraduate students at an institution this way. You put all the classes being taught into a hat and randomly draw our five of them. You then get a lists of students in each of those five courses. Then from each list you randomly draw 10 students. This approach produces a ____________________ sample. Why is it that every undergraduate does not have an equal, known chance of being selected?

4) You do interviews with people my asking the people you interview to suggest

others to interview. This approach produces a _________________________ sample.

5) You create a sample that "looks like" the population in terms of major attributes such as gender and age distribution by selecting people to ask to participate on the bases

of those major attributes. This is called a ___________________________ sample.

6) Write a hypothesis in the way taught in class that anticipates a relationship between political party affiliation and attitudes toward a local tax increase to finance a new public park.

7) Write a hypothesis in the way taught in class that anticipates a relationship between an after school youth recreation program and number of dependent children.

8) Write a question at the interval level of measurement that operationalizes annual family income.

9) Write a question at the ordinal level of measurement that operationalizes annual family income.

10) Write a Likert-type "question" intended to operationalize computer anxiety.

11) Consider the following research design.

Figure 1: Research Model

11a. What appears to be the dependent variable?

11b. What appear to be the intervening variables?

11c: What appear to be the independent variables.

11d. How many hypotheses does this model suggest?

11e. Which variables appear to be the relatively more difficult to measure?

11f. When multiple questions are included in the survey in an attempt to measure one variable, why should one of the questions have face validity?

Chapter 6 -- Contacting and Talking to Subjects

Surveys

Interviews

Focus Groups

In a survey, everyone gets the same SURVEY INSTRUMENT. This approach can be ANONYMOUS.

In GUIDED INTERVIEWS the PI (or someone else) actually talks to individuals. There is a list of questions but the interviewer is able to follow the lead of the person being interviewed and make adjustments along the way. This approach can be CONFIDENTIAL.

In a FOCUS GROUP the PI (or someone else) guides a conversation involving MULTIPLE SUBJECTS (persons) at the same time. . There is a list of questions but the interviewer is able to follow the lead of the person being interviewed and make adjustments along the way. The members of the group are likely to know one another. The researcher cannot promise that participation is anonymous or confidential.

There are many options regarding collection of SURVEY DATA.

in person

telephone

e-mail

a tool such as SurveyMonkey

even postal mail

Sometimes a small amount of MONEY is sent to encourage people to participate. This is very costly. It may (or may not) increase RESPONSE RATE. Low response rates are a big problem because people who respond may well be systematically different from people who do not respond.

Researcher frequently make MULTIPLE FOLLOW UP efforts. But in order to do so they really need a way to keep track of who has responded and who has not.

THINGS THAT CAN AFFECT RESPONSE RATE

• Whether or not the topic of the survey is of interest to the respondent.
• The means by which the subject was asked to particpate.
• THE LENGTH OF THE SURVEY.
• The perceived LEGITIMACY of the research/organization.
• The time constraints of the subject.
• THE SKILL OF THE INTERVIEWER.

PROBLEMS WITH SURVEY RESEARCH INSTRUMENTS

You may want to do a PILOT STUDY to make sure people apparently understand the questions.

But if people do not understand correctly you may not know it.

You can observe which questions subjects refuse to answer.

You may be able to better figure out how long it takes to complete.

THE COST OF VARIOUS METHODS

• Postal mail is not cheap if few people respond. It can be very expensive if you put money in every envelope.
• You probably get better data from PERSONAL INTERVIEWS but the cost of the time and travel of the person who collects the data can be very high.
• Bottom line, it is getting harder to get people to participate in research using traditional methods of data collection.
• Think about the kinds of people who want to participate and the nature of the data you need and try to pick a method that will be effective and not too costly.