Questionnaire Checklist

       Following is a checklist to go over before finalizing a survey instrument. It is based on experience in URBS 492 "Research Methods" and is intended to guard against common errors beginning survey researchers commonly make. This checklist is written for a mail survey, but the format for phone or face-to-face surveys would be very similar. Suggested is one approach which will meet the needs of most of the surveys will do for class projects or simple surveys for organizations. For much more detail and suggestions on more sophisticated questionnaire construction see:

    Donald Dillman, Mail and Telephone Surveys: The Total Design Methods (New York: Wiley, 1978)
    Jean Converse and Stanley Presser, Survey Questions: Handcrafting the Standardized Questionnaire (Newbury Park: Sage, 1986)
    Floyd Fowler, Improving Survey Questions (Thousand Oaks, CA: Sage, 1995)
      Priscilla Salant and Don Dillman, How to Conduct Your Own Survey (New York: Wiley, 1994)

1. Margins: Are the margins set as follows: Top, Left, and Bottom 1"; Right 2"?

2. Title: Does your survey have a short, descriptive title no longer than one line  in 18 pt bold type, centered,
    at the top of your survey two double spaced skips above the beginning of the introduction?

3. Introduction: Is there a short introduction to the survey below the title (no more than 150 words)?

4. Edge Coding Line: Did you draw a heavy black vertical line from along the right hand margin of your questionaire using a
    felt pen. Does it start it at the bottom of the introduction and   run to the bottom of the page for the first page and the entire
    length of the subsequent pages?

5. Page numbering: Are all of the pages of the survey instrument numbered except for the first page?

6. Survey and Coder IDs: Did you put on separate lines at the top of the survey instrument right aligned with the edge
    coding line the words:

              Office Use Only  

Survey ID

Coder

      These should be on separate lines and double spaced. For a phone or face-to-face survey instrument you should
      also include the word Interviewer.

7. Question and response numbering and format: Did you

• Number each question with a Q in front of the number: Q1, Q2, Q3 etc.

• Indent response response categories

• Leave enough space for respondents to answer the question in the space provided on the questionnaire itself. One third to half a page or more is common.

• Place each response category on a separate line

• Have one or more engaging and easy to answer questions at the beginning

• Put sensitive questions and demographic questions at the end

• Have brief transitions between different sections of the survey: (e.g. "And now we have a few questions about how frequently you, yourself, use San Francisco Parks...")

• Double space the responses

• Put the number of underscore marks required to code the question in the space to the right of the edge coding line, centered at the middle of the question + response categories. You do not need to repeat the question number

• Number (not letter) responses

• Right align responses

• Connect the response choices with about 20 dots, like this:
  
  Q 47 What is your sex?
  
                    Circle one

                    Male..........................1

                    Female.......................2

• Set up questions where there is a common response as matrices
  like this:

          Q23 How would you rate MUNI performance on the following

                                                                                           Circle One

                                                                                Excellent   Good Fair Poor

a. Driver courtesy                                        1           2        3      4

b. Well planned routes                                 1           2        3       4

c. Reliability                                                 1           2      3          4

8. Question logic: Did you check to see that all questions are:

• Clear

• That the respondent will be able to answer them

• Are unambiguous

• That questions which apply to only some respondents are handled with a skip instruction asking those to whom they do not apply to skip to the next question which does apply to them?

• Are not leading or biased

• Are not "double barreled", i.e. are not asking two or more questions? (If they are they should be divided into multiple questions).

9. Response categories: Did you check to see that all response categories are:

• That you have enough response categories to give you useful information and that they are the right ones

• Exhaustive. I.e. that every possible response is covered. (When in doubt put the words "Other Specify" and leave a blank line for the response

• Mutually exclusive. That there is no overlap between categories

• That they call for data at an appropriate level: nominal, ordinal, inteval, or ratio level. (Remember, you can recode data to collapse categories, but not the other way around)

10. Pretest: Did you pre-test the survey instrument. This is the best possible way to improve it. If you have no time at all 
      pre-test the questionnaire once on yourself. Could YOU answer each of the questions as it is asked. If you have more time
      a pretest on a small convenience sample (roommates, friends) will greatly improve questions. For surveys used in class
      projects or work alway pretest with a small number of people from the universe that will be surveyed.

11. Other

• Is there a thank you at the end

• If a mail survey are instructions about returning it clear

• Did you leave enough space on the questionnaire. (Don't cramp it)

• Did you spell check the questionnaire and proof it for other errors