I. Hypotheses

A. Descriptive hypotheses best answered by this method

B. If you don't consider your hypotheses before writing your survey, you may be

1. Overwhelmed with data

2. End up with data that doesn't address your concerns

C. Good to go through several different scenarios of outcome from survey to see whether different outcomes would indeed have different implications for:

1. Your hypotheses

2. What action you will take (if survey is to address applied issue)

II. After determining precisely what you want to find out, determine who you want to ask

A. Defining your population

B. Determining whether to use the population or to sample

C. Types of samples

1. Convenience samples

2. Quota samples

3. Random samples: Allow you to use inferential statistics to determine how closely your results reflect their population

4. Stratified random samples: The advantage of random samples, but with a smaller sample and/or greater accuracy

III. Questionnaire, Interview, or Telephone Survey?

A. Issues to consider

1. Cost

2. Response rate

3. Honesty of responses

4. Standardization

B. The case for the telephone survey

IV. Format issues:

A. Format of questions

1. Dichotomous versus continuous

2. Fixed versus open-ended

B. Format of survey

1. Structured

2. Semi-structured

3. Unstructured

C. Why a novice might be better off with fixed alternative questions and a structured survey:

1. Data is easily coded

2. Structure may reduce investigator bias: Data on hypothesis-confirming bias (Snyder, 1984, Snyder, 1981, Snyder and Cantor, 1979)

V. Rules for asking good questions

A. Use words a third-grader would understand

B. Use words that won't be misinterpreted

C. Avoid personal questions

D. Make sure your sample has the information you seek

E. Avoid leading questions

F. Avoid questions loaded with social-desirability

G. Avoid double-barreled questions

H. Keep questions short and concise

I. Avoid negations

J. Avoid irrelevant questions

K. Pretest the questions

VI. Analyzing survey data

A. Summarizing data

B. Summarizing interval data

C. Summarizing ordinal or nominal data

D. Using inferential statistics

1. Parameter estimation with interval data

2. Hypothesis testing with interval data

a. Relationships among more than two variables

b. More complicated procedures

E. Using inferential statistics with nominal data

1. Estimating overall percentages in population

2. Relationships between variables

Back to Chapter 8 Main Menu