SPRING Questionnaires Flashcards
what are questionnaires
survery method wehn limited resources, need lots of data/pps., gathering data about knoeldge/beliefs/attitudes, want to measure specific contracuts and must protect pps privacy
how might questionnaires be used in experimental design
gather descriptive and analytic data
about individual but numerical ie scales
how do you design a questionnaire
identify RQ/hyp - define study goal identify info needed for logistics of study and confounding factors write qs organise into questionnaire format pilot and revise qs
1.define the goal of the study
define construct
ensure questionnaire taps into desired cosntruct
problems with existing questionnaire
revert back to contruct own
- goal to measurement
clarify what you want to measure and how you measure it
ie is it a behaviour?attitude?
concepts you can measure with questionnaires
attributes (facts of ppl)
behaviours
beleifs
attitudes
what are response scales
yes/no or yes/no/dont know
yes/no may be problematic - force an answer where might not be one
ALSO may response where socially desireable
what is a likert scale
v strong agree - v strong disagree
5 midpoint - no opinion
encourage across all scale
what is a semantic difference scale
enoy - hate
forces an answer where may not be one - no defined midpoint
close qs
what is a labelled response scale
excellent - good - fair - poor
no set numerical value
subjective to interpretation by reader
if using scales what data can you collect
catergorical or ordinal
need at least interval for a parametric test (gaps equal)
what is a continuous visual analogue scale
position on a continuous line between two extremities
what problems might be encountered when asking about periodical behaviours
time frame used cannot be too ambiguous
cannot be too immediate or old (forget)
cant use ambiguous words ie “usually” as interpretation of what usual time frame is is dependent on the person
describe social desirability response bias
people want to make a good impression so answer in socially desirable way
how might social desirability be avoided in qs
avoid phrases that insigate desirability ie “it is commonly believed” “research has shown” “doctors say” .. etc
define acquiescence
tendency to agree with every statement
how can you avoid acquiesence
reverse framing of the question
rules for question writing - DOs
use simple language short qs "dont know" / "na" relevent definitions write sevral qs for pilot
rules for question writing - DONTs
dont use double barrelled qs, leading qs, double negatives, amiguous words, qs that create opinios or ask unanswerable qs
how can the order of qs negatively influence questionnaire data
initial questions may influence how later respond
ie instigate a social desireablilty bias
how to correctly order questions
filler qs - exclude some pps
funnell approach - ask broad qs, introduce key concepts gradually and encourage completion by putting easiest first
where should demographic questions be put
end of questionnaire
qs that exert effort are early and more important qs first - caputre immediate emotions
what should you judge during a pilot test
discrimination
validity
length of questionnaire
problems with wording
what is discrimination during pilot testing
should be able to discriminate between two samples/groups
People with the same score should be equal to each other along the measured construct
People with different scores should be different to the other groups along the measured construct.
The degree of difference between people should equal the difference score.
discrimination and question contamination
two diff samples score same on construct - qs not measure same construct
types of validity
content validity
criterion validity
factorial validity
what is content validity
all items must relate to the construct
what is criterion validity
a measure of how well one variable or set of variables predicts an outcome based on information from other variables
compare one method to another known standard measure of same construct
what is factorial valdity
the degree to which covariance among judged traits resembles the actual or true covariation of observable behaviors underlying these traits.
ie are the questions themselves correlated?
pros of questionnaires
standardised format quick and cost effective lots of responses easy to analyse anonymity no experimenter bias
cons of questionnaires
low response rate cant probe responses respondent may not be who intended cant correct misunderstandings difficult to establish causality
methods of questionnaire administration
personal
mail
telephone
internet
pros of personal q administration
spontaneous response
high response rate
emplain abiguity or misunderstandings
cons of personal q administration
time consuming
expernisive
not anonymous
social bias
pros of mail q administration
low cost and time
demographic
anonymous
cons of mail q administration
high no response rate
incomplete answers
not get desired sample
pros of telephone q administration
low cost and time
anonymous
spontaneous
cons of telephone q administration
no cisual aids
- can help misunderstandings
pros of internet q administration
low cost and time spontaneous and convenient felxible desgin anonymous demographic
cons of internet q administration
no response high
incomplete answers
misunderstandings
