Monday September 22, 2013
PSYC11H3 – Social Psych Lab
QUESTIONNAIRE AND SURVEY DEVELOPMENT
- The most familiar method for social psychology
- Advantages of using surveys
o The most common method used in social sciences
o Quick and inexpensive, exp if using the internet. Lots of software
programs, no money on paper.
o Easy to get large samples
o It can be as simple or detailed as u want it to be: open ended, MC, it
can be as rich as you want it to be or it can be specific. It can be two
ends and one continuum.
o Thres less social desirability, theres no pressure of the experimenter.
Less pressure to sway responses as participant.
o Don’t need as much training in terms of staff or research assistants.
Cost effective. Can reach a lot of ppl really quickly
Easy to analyze: all in spreadsheet, ready for u to analyze. Don’t
have to transcribe or code anything or train assistants etc
Familiar to most ppl: doesn’t make ppl apprechensive.
Everyones taken some kind of quiz at some point so not as
Reduces bias – not like asking them in an interview, and start
taking notes when respond, “okay…” could alter responses
Less intrusive: to participate, can do it in own time and space.
Less intrusive to even researcher to not sched time in their
day, go to lab etc. less intrusive for everyone.
o There may be bias in way people respond: on same scale on same
o Social desireability: what is socially acceptable in society- ppl may
bias if ppl think they may be judged
o The larger the questionnaire, the more missing data there will be.
o The more complicated it may be, so there will be bad data
o The longer it is, the worser it is ie; not accurate
o Bias sample: ppl who will respond r those who are very invested, and
perhaps those very invested in certain position.
Low response rate: easy for ppl to overlook and dismiss.
Inability to probe responses: no qualifying info if theres a
forced choice. How do we know that’s where they really
wanted to lie, what if it was in between.
1 Cannot guarantee that responses r accurate: cld be tired
Saying no not stressed, but really are
Cant guarantee the responses are real.
Kids doing it
Limit samples to only literate folks
Even internet as well.
o General Considerations
Clarify study goals
Survery should be developed to directly address study
o At the design stage. Examine how each and every
question will be analayzed. Figure out ur
hypothesis, then make a plan on how to test it.
Need a plan for statistical analysis. Clarify study
goals and plan before even starting
More responses if short
Keep it as concise as u can, the more conven, the better ur data
will be. The cleaner, the better. Bcs we spend a lot of time
cleaning and re-cleaning our data. Make it as clean as possible
on the outset.
o Example question, emotionally bound. English language is not very
straightforward. U don’t know what the question is asking. Its very
dichotomous, agree or disagree.
o Example 2: thinking about the health care law….”
Westions are too long
It can be more basic.
o Is evey poll a legitimate poll? No its not. Esp like fox news, trying to
draw an audeience so things are really exaggerated. We don’t know
what they were asked, how complicated it was, what questions were
asked etc. so is it scientifically precise? No, fox news isn’t.
o Question Wording
Strive for objectivity.
Adjectives and advers, like totally, always, never etc.
If double baralled, don’t know if they were answering
this or answering that.
o Ex. 3 How do you feel about building…” poor : using word unused.
Better: the first question dsnt say there is another option.
o “How often do you punish your toddler?” don’t know what the scale is.
Using word punishment more punitive than discipline. What is
punishment. How often- what does that mean? Vague: what does
punishment mean and how often?
Better: this is more clear, more objective.
o QUESTION RESPONSES
2 o Reverse scale items. Ppl can become habitually, but if range
and change them, the less habituation ull get the more scale u use.
“don’t know” – like polling: someone undecided vs neautral:
don’t care versus, don’t know.
If ppl are bored, theyre lazy so placting options at the end
increases their selection. Putting at the end makes it salient to
them and more likely to respond I don’t know.
o Marital status?
o Writing ability: terrible sounds too bad/extreme.
There are many diff kinds of writing styles, what is it referring
Mediocre sounds bad too. But there shld be two on good scale
and two on the bad end… need to balance it out.