Survey Design 101

SURVEY DESIGN 101Gay Hylton

Institutional Research and Planning

WHAT YOU’RE TRYING TO FIND OUTOften there is the temptation to skip on preparation in order to move to the field too rapidly. This temptation should be avoided.

Ghislaine Delaine“The Social Dimensions of

Adjustment Integrated Survey”

INFORMATION WANTED: What information needs to be obtained to

meet the objectives of the survey Strategize with stakeholders Questions that need to be answered Focus groups Has the information been gathered in the past What are the outcomes desired Will basic demographical data be required/confidential How will the information be obtained?

Electronic survey Mail Telephone Interview Personal Interview Paper

INFORMATION WANTED: CONTINUED Determine what correlations of the data you

want to see. What will the answers to the questions allow you to

determine If X then Y

WHO YOU ASKOne common misconception is that the adequacy of a sample depends heavily on the fraction of the population included in a sample.

Floyd J. FowlerSurvey Research Methods

CHOOSING THE POPULATION All or representative sample http://www.surveysystem.com/sscalc.htm Is a comparative sample needed Response rate

# of responses/ # surveyed If survey needs to reflect the entire student

body population of HSU, complete a survey request form at http://www.humboldt.edu/irp/survey.html

HOW YOU ASK THE QUESTION“The goal is to have differences in answers reflect differences in where people stand on the issues, rather than differences on their interpretations of the questions.”

Floyd Fowler,Improving Survey Questionnaires:

Design and Evaluation

BASICS Be aware of your own biases Use language that survey participants will

understand Develop neutral questions Ask enough to cover topic adequately while

keeping the survey as short as possible Pay attention to the order of questions Provide exhaustive range of response

categories Write clear unbiased instructions

BASICS - CONTINUED Order

Matters Start with easy - proceed to complex Keep respondents interested Vary the question type

Simple is better Avoid technical jargons or concepts Use the same definitions and scales throughout the

form Watch for “double-barreled” questions Be specific

TYPES OF QUESTIONS Multiple Choice

Allow multiple responses? Ranking Scale Open Ended

SPECIFICS Multiple Choice

Lists Should be exhaustive while not being too long Categories should be mutually exclusive Allow respondents to provide multiple answers when

relevant When appropriate – use Other

Scales Odd or Even Likert or Numeric Provide clear, concise instructions on scale meaning Order matters – Positive to Negative vs Negative to

Positive Use the same scale throughout the survey

SPECIFICS - CONTINUED Open-Ended

Allows for spontaneous responses Use when you don’t know the answer Analyzing the responses can be difficult and time-

consuming

Question Logic Will all participants answer all questions?

HOW AND WHEN TO ASK“The questionnaire is only one element of a well-done survey.”

Don A. DillmanMail and Internet Surveys

– The Tailored Design Method

BEFORE YOU SURVEY Design your communication to participants

Personal or authoritarian Confidentiality? Explain the nature and reasons for the survey How it benefits them Include timeframe to answer survey

BEFORE YOU SURVEY - CONTINUED TEST

Pilot the survey with colleagues or a small sample of the population

Include all communications they will receive with the survey instrument

Helps ensures the reliability and the validity of your survey

For electronic surveys check for the flow of logic questions

TRAIN (If survey is to be administered via telephone or personal interviewer)

All interviewers need to conduct the survey in the same manner, using the same language

Design response sheets that include instructions to interviewers

INCREASING RESPONSE RATES“Sending questionnaires out is one thing; getting them back is quite another.”

Bill GillhamDeveloping a Questionnaire

TIMING MATTERS What are your target population’s habits Ask again but don’t ask too often Choose a different time to ask for each

additional request

COMMUNICATION Say please and thank you for your help Show positive regard – provide the reasons

for the survey and what the results will achieve

Asking for advice – “We need your input on this matter.”

Subject line of email – does it look like spam Does it need to come from an authoritative

source or a personal approach

TOOLSAll tools have intrinsic politics and technology is the tool of now.

Godfrey Reggio

FREE OR NOT Software

Survey Monkey Zoomerang Surveygizmo And many more

Design help Library – reference section “Survey Kit” , ebooks and

books on survey design IRP website has some links to reference material

QUESTIONS?“Who questions much, shall learn much, and retain much.”

Francis Bacon