Questionnaires & quantative data (MR, unit 8)

QUESTIONNAIRE DESIGN - facilitate precise statistical analysis
either used to explore attitudes or seek factual information

Method:
Develop question topics derived from research objectives Select Q&A formats and wording
Determine sequence
Design layout
Pilot test (Wilson 2002)

FORMATS & WORDING

Precision

• Beware of:
• Ambiguity
  
• lack of Clarity - why am I asking this? what do I want to know?
  
• Conflation of multiple questions into 1
• making Unjustified assumptions

Open & closed questions
Closed - establish basic facts (predetermined answers or Y/N)
Open - own words but timely analysis. (offer ltd range of possible responses)

Leading questions
Mix positive & negative statements. If you suggest intrinsic preference, they'll try to answer as you want).

Formats

• Error of the central tendency - reluctant to be extreme, sit middle ground, only offer even no.
• Ranking - clearly indicate whether 1 is least or most important. Can confuse so use sparingly. Better to select one e.g. single most important contributor to...
• Scales - where factual info lies on continuum e.g. less than 3, 4-6, 7-10
• Select most relevant -"tick ANY that apply"
• The Likert Scale - respondents rate each on scale from strong disagreement to strong agreement (1-5) no absolute meaning, ltd statistical use
• The Semantic Differential Scale (+3 to -3, powerful to weak) pairs opposite attributes. Select position of object to degree which adjective describes. Subjectivity of language, some attributes bring others with them e.g. old cars = classic

SEQUENCE
Funnelling - progress from the general to the specific (broad context to underpinning data)
Halo effect - overall +/- feelings about issues influence responses to individual questions (distribute throughout rather than in sections)

LAYOUT

• lines, boxes, print size, small pictures
• clarity of tick boxes
• explain purpose - guarantee confidentiality (if possible)
• thank at end

LENGTH (short is always better)
Cold surveys: no prep or incentive
1-2 sides, 15-20 questions (presentation is key , user-friendly)
Postal questionnaires: respondents briefed & prepared but doing in own time
6-8 sides, 30-40 questions max (clear instructions, provide examples)

PILOT TESTS - VITAL!!!!