Questionnaires and Survey Design

1
Questionnaires and Survey Design
2
Not included in sampling frame
Sampling frame
Cannot be contacted
Not eligible for survey
SAMPLED POULATION
Refuse to respond
Incapable of responding
3
Considering survey methods

• There is no best method
• Each method has strengths and weaknesses
• Ask yourself
• 1.Who are the survey respondents?
• 2. Is the research question more amenable to one
  method than to another?

4
(No Transcript)
5
Other Considerations

1. Administrative factors such as cost, time,
   geographic distribution
2. Questionnaire issues such as branching,
   complexity of questions, sensitivity issues,
   rapport with respondent
3. Data quality issues such as bias, knowledge about
   refusals and non-contacts, control of replies

6
Mailed questionnaires

• Cost is relatively low
• Response rate may be poor and biased towards more
  educated and those with an interest in the topic
• No knowledge about non-response
• Long time between data collection and analysis
• Geographic distribution can be wide

7
Telephone surveys

• Most widely used type
• Numbers are selected randomly from a phone book
  or by generating random numbers
• (RDD (Random Digit Dialling)
• Problems include unlisted numbersLandline
  coverage is decreasing so increasing bias against
  the young (no directory for cell phones)
• Geographic distribution can be wide
• Data collection period can be short
• Questions must be short and simple

8
(No Transcript)
9
Face-to-face surveys

• Cost is high
• Data collection period is long
• Geographic distribution must be clustered
• Good rapport
• Take a long time
• Good control and quality of responses
• Sensitivity issues fair quality

10
(No Transcript)
11
Internet

• Low cost (no paper, no data entry costs)
• data collection is quick
• Geographic distribution may be wide
• Questionnaires may be complex because the skips
  are programmed in
• Pop-up instructions, videos, voice-overs,
  animation are available to make it more fun and
  dynamic
• If self-selected, results may be biased.