lecture 4 - questionnaires Flashcards
(20 cards)
what is the overall research journey?
research question - population -> sample -> questionnaire design + testing including pilot testing + reliable and valid measures -> data collection -> analyse data using appropriate analysis techniques
what are questionnaires and surveys?
- applicable to multiple study designs
- mail or postal questionnaire traditionally most common form (increasingly online)
- self or researcher administered
- enables standardised data collection
- common tool to collect data
- large or small samples
advantages of (self-completion) questionnaires
- cheaper
- quicker, familiar to most people
- absence of interviewer effect
- convenience
- go at your own pace
disadvantages of (self-completion) questionnaires
- cannot prompt
- difficult to ask many questions
- do not know who answers/completes
- not appropriate for some respondents - low literacy
- respondents may not take research seriously
- lower response rate
- greater risk of missing data
explain the different levels (types) and basic derivatives?
Meaningful coding of questions is important for appropriate quantitative analysis
Different ‘types/levels’ of data
Nominal/Categorical
Ordinal
Interval
Ratio/metric
Convey different amounts of information
Can be analysed/described
what are the types/levels of data?
- nominal/categorical
- ordinal
- interval
- ratio
explain categorical data
data can be counted/placed in a ranked order or pin Tinto distinct groups
- nominal
- ranked
what is normal data?
can’t be measured numerically, but can be distinguished by classifying into sets (categories) e.g. gender/type of sport
what is ranked/ordinal data?
can’t be measured numerically but can be put into ranked order e.g. podium finish/order of sports preference
what is numerical data?
data which can be measured as quantities
- interval
- ratio
what is interval data?
can state difference between two data values but NOT relative difference e.g. Celsius temperature scale
what is ratio data
can state difference between two data values AND relative difference e.g. WNBA player salaries
what are the types of questions in a questionnaire?
open v closed
respond as you wish versus select from fixed set of alternatives
what is an open question?
- use own language
- novel responses possible
- exploratory in general (but not always)
BUT - need coding - categorising and/or scores being attached - look for themes in data AFTER collection
- scores (usually) require mutually exclusive categories
what are closed questions?
(pre-coded!)
- easy to process responses - link to prior theory/concepts
- comparability (e.g. with other studies) clear
- meaning ‘clarified’ before questionnaire sent out
- pre-test/pilot study
other things to consider
- are the categories in a question mutually exclusive?
do they need to be or not?
what if you’re undertaking ‘exploration’?
what is the likert scale?
balanced response scale
- scale lengths 3,5, or 7 options often used i.e. an odd number
(even number scale lengths rarely seen)
- equal number positive and negative options
- data captured
- technically ordinal data
- some disciplines treat this data as being interval in nature
what are issues with questionnaire design?
- question ambiguity - important to be as specific as possible, although also important to consider previous literature (sometimes can cause conflict!)
- avoid long or double(-barrelled) questions
- implies length and clarity can be in tension - design important
- avoid negatives - “are you not…”
- avoid technical/theoretical terms
- balance closed and response questions to avoid bias
- avoid leading questions
what should you think about before you send questionnaire out?
number of responses and response rate
- response rates can often be low when using questionnaires
what are the ethical considerations of a questionnaire?
cover letter/paragraph(s), with the following information can help enhance the number of responses:
- aims of project and important of it
- confidentiality and/or anonymity
- research for academic purposes
- no right or wrong answers
- explain that you are not trying to deceive/trick people
- monetary sponsorship incentive
what should design features be to increase response rates?
- return postage
- follow-up non-response: letter, restate aims etc
- consider length but do not compress
- clear layout and instructions on responding
- consistent fonts etc (potentially make different for questions, instructions, responses)
