We are doing quantitative research on play habits at work and in people's free time.
This includes a series of frequency questions like "How often do you play xxxgame typexxx while at work"...at home.. etc.
I've done sthis before, but recently I've been looking for a broadly accepted frequency scale to use, but to no avail.
I don't want "very rarely, very frequently" because it's too abstract/impriecise. An open number format won't really fit either, as some participants may legitimately answer once or twice a year while others play multiple times a day.
I've arrived at something like this, but somehow it doesn't "feel right"
- Never
- Once or twice a year
- Once or twice a month
- Once or twice a week
- Many times a week
- Every day
- Many times a day
Does this look reasonable? How would you react to it as a survey taker? As a reviewer?
Are there any good standards (ideally with a reference) for this kind of question? Alternatively, have you published similar research with a frequency scale that worked well for you?
https://www.clemson.edu/centers-institutes/tourism/documents/sample-scales.pdf
Thanks Emre. That's great advice. We actually have quite a few "pilot" studies on hand using similar scales for behavior frequency. Maybe I should go back and check those before proceeding.
In the meantime, it would be wonderful if someone had a link to a nice similar scale already used in a paper in a good journal. Would make publication a bit easier I think.
What bothers me with the format is that it is not remotely "interval/ratio" like. "how many days..." or "how many hours a day do you..." can be analyzed as interval/ratio variables. Likert scales of at least 5 points are routinely analyzed under the premise that the distance between to adjecent values is equal. (ie the disance between 5 and 4 is the same as the distance between 1 and 2).
With your response categories you wont have the equal distance which might be a problem for some stats tests.
Perhaps a "during how many days the last month have you played XXGAMEXX at work?" (or 3/6 months perhaps depending on the gametype and frequency one would play it).
and you might want response categories to be open ie punch in a number or something like: 0, 1-5, 5-10 ,10-15, 15-20
As part of my PhD study (in architecture and urban studies), I collected data on children's (aged 7 to 15) actual play behaviour across 'time' in two ways: daily time-activities in hours and weekly play activities in days.
For the daily time-activities, I asked them about daily activities, including play activities, that they performed on the school day and the school holiday immediately before the survey/interview date. I provided a list of activities (with the option of including 'other' activities). Children were asked to write down the time of those activities (in ‘from hr' -‘to hr’ format) in the given boxes. For younger age group, I provided direct assistance.
For the weekly play activities, I conducted survey of children's respective parent(s). In the questionnaire, children's parents reported on their children’s outdoors play (of minimum 30 minutes, in number of days) in the 7 days immediately before the interview date. I provided the numbers for days (1,2,3,4,5,6,7) and asked them to 'Tick'.
I adopted a mixed methods research model.
My work is not published yet. You might consult recent works in the journal: 'Children's Geographies'. You might get some recent trends in similar studies.
I agree with the above mentioned comments. 'Pilot study' and 'equal-interval scale' are important.
Have you googled "behavioral frequency questions" (and behavioural for UK spelling)? The choice of the alternatives will depend on the behaviors and the choice of which to use will affect aspects of the respondents (for example, in the attached we show it affects what people believe are the norms for the population).
Article How response alternatives affect different kinds of behaviou...
@daniel: I did quite a bit of googling, but I didn't hit any gold. Thanks! Downloaded.
- Badges
- Science method
- Similar topics
- Mathematical Sciences
- Statistics
- Applied Statistics
- Survey Design