Writing Good Questions



  •   Identify the need for each question.

    Each question should align with your objectives. Make sure the results from each question are necessary either for reporting (summative evaluation) or for continued program development (formative evaluation). Consider the following:

    1. Do you need to ask the question?
    2. Does the question require an answer?
    3. Do survey respondents have an accurate, ready-made answer?
    4. Can respondents accurately recall past behavior?
    5. Will respondents be willing to reveal information?
    6. Will respondents be motivated to answer each question?

     

      Consider the demographic information you will need depending on the purpose of your survey.

    Consider how you will use this information for program-planning and decision-making. For example, if the implications of the findings may be different for men than for women, make sure that you collect gender information. On the flip side, if a demographic characteristic has no bearing on the interpretation of findings, it is better not to collect it. (A good rule of thumb is to only gather information that you plan to report and use for decision-making.)

    Demographic information should also be gathered in order to assess how closely the demographics of the study population match those of the larger population from which the survey sample was drawn. The closer the match between the sample and the population, the more support you have to generalize your findings to the larger population.

      Be as brief as possible.

    Questions should consist of no more than 20 words and three commas. Keep multi-syllable words to a minimum.

      Avoid double-barreled questions (asking two or more questions in one survey item).

    Edit your questions for the following words: and, or, but, with, except, if, and when. These are words that usually mean you have multiple questions in the same statement. For example:

    Instead of:

    1. I have used the information I received from coaching and professional development this year in my classroom practice.

     

    Use:

    1. I have used the information I received from coaching this year in my classroom practice.
    2. I have used the information I received from professional development this year in my classroom practice.

     

      Provide a time frame if it will clarify the question(s) or assist the reader in responding.

    Choose a time period that is appropriate when considering the importance of the question(s) and the respondents' ability to recall the information. For example, for most school-based surveys, you would ask respondents to recall information from the most recent school year. If this time frame applies to the entire survey or to a group of questions, it should be described in the survey introduction or in the instructions immediately preceding the series of questions to which it pertains.

      Keep questions clear and specific, yet do not ask for more detail than what can be easily recalled.

    For example:

    Instead of:
    How much did you spend on school supplies last year?

      $0 - $10
      $11 - $20
      $21 - $30
      over $30

     

    Use:
    How much do you estimate you spent on school supplies in the last year?

      $0 - $10
      $11 - $20
      $21 - $30
      over $30

     

    Most respondents would not be able to remember exactly how much they spent on school supplies. Responses would likewise be guesses rather than actual numbers, and many respondents could become frustrated trying to calculate in their heads how much they spent. If an approximate answer is all that is needed, simply rephrasing the question by adding the word "estimate" will make it much easier for the respondent.

      Clearly define potentially ambiguous words and phrases to ensure common understanding.

    Avoid terms that have several possible meanings or that may be interpreted differently by different respondents; this includes words like "recently" and "frequently" and other time-related terms, as well as terms like "family" and "household," which may mean different things to different people. Even asking people "when" they did something (assuming an open-ended format) could result in different responses depending on each reader's interpretation of the question. Try using language that is as precise and specific as possible. For example:

    Instead of:
    How frequently do you communicate with your child's teacher?

    Use:
    On average, how many times a week do you communicate with your child's teacher?

      Do not assume too much knowledge or use jargon or acronyms that respondents may not understand.

    For example:

    Instead of:
    Which of the following types of courses has your child taken? (check all that apply)

      AP
      IB
      PSEO
      CIS

     

    Use:
    Which of the following types of courses has your child taken? (check all that apply)

      Advanced Placement (AP)
      International Baccalaureate (IB)
      Post-Secondary Education Options (PSEO)
      College in the Schools (CIS)

     

      Make sure that questions are not too complex or difficult to understand or answer.

    For example:

    Instead of:
    Please rank the following factors from 1 to 15 as to their importance in selecting a school for your child:

      Transportation
      School start/end times
      School's general reputation
      Quality of teaching staff
      School building/facilities
      Availability and use of technology
      Existence of Discovery Club in the building
      Existence of early childhood programs in the building
      Familiarity with other families who send their children to school
      Test scores
      School leadership/principal
      Athletic offerings
      Arts programs
      Proximity to home
      Programs serving children with special needs (i.e., special education, ELL)

     

    Use:
    Please check the three most important factors in selecting a school for your child.

    Most respondents would have considerable difficulty answering a question as complex as the original one and/or would not be able to answer it with any degree of accuracy or precision. Analyzing the data from ranking questions is equally complex. A preferred method is to ask respondents to check the top factors or items on the list; the analysis will provide a ranking based on the number of votes each factor receives.

      Avoid double negatives.

    For example:

    Instead of:
    Are teachers in your school never unfair?

    Use:
    Are teachers in your school fair?

      Avoid biased or leading questions.

    Do not word questions in a way that suggests a preferred, correct, or expected answer. For example:

    Instead of:
    Do you limit your children's television viewing to two hours a day?

    Use:
    On average, how many hours of television do your children watch in a day?

      Use language and wording that is culturally sensitive and inoffensive.

    Ensure that the language/wording you use is not offensive to anyone and that it has the same basic semantic content for all respondents regardless of race, age, gender, ethnicity, sexual orientation, or religion. Independent representatives from diverse groups should examine the survey and identify items and procedures that may yield differential responses from one group relative to another. Consider a sensitivity review of the survey (sensitivity review information).

      Write items on sensitive topics in ways that respect the privacy of your respondents and increase their comfort level with taking the survey.

    Only ask for information that you actually need. For example, if you only need to know if a respondent's income is above or below the poverty line, do not list many categories of income level or require them to report their exact income. For example:

    Instead of:
    How much was your total family income in 2008? 

    $

     

    Use:
    Which of the following categories best describes your total family income in 2008?

      LESS THAN $5,000
      $5,000 to $9,999
      $10,000 to $19,999
      $20,000 OR MORE

     

    Back Back      Continue Next


  • For further information on conducting surveys, contact Marian Heinrichs(651)767-8116, Department of Research, Evaluation and Assessment.