Questionnaires Flashcards
What are we measuring?
Psychological constructs
What are the problems with what we are trying to measure?
They are psychological constructs which don’t actually exist - so they aren’t directly measurable
Examples of psychological constructs
empathy nosiness anxiousness greediness selflessness mental imagery ability attention memory
How do we measure the immeasurable?
Observe behaviour - put people in situations and see how they behave
Brain scans - see what area is activated when doing tasks, but no evidence that these exist, becomes circular
Brain lesions - lack a quality, see what corresponds to that issue but could be involved in other things
measure aspects of the quality - not perfect measurements, build a case for why you have chosen to measure something
How do we know these qualities/abilities can be measured?
How do we know they exist as traits?
have psychological reality
have particular characteristics
What is a construct?
The underlying concept that we can’t measure
What are items?
The things that you put in a questionnaire - thoughts, attitudes, feelings or behaviours
How do we arrive at the diagnosis?
By observing the combinations of different symptoms
How can we define and measure a construct?
By finding out what characteristics/items produce similar ratings/response - patterns indicate the degree to which they have a construct
How are factors measured?
By responses to questions / items
What does polarity mean?
Direction to a particular question - higher and lower mean different things
What makes a good questionnaire?
Discrimination
Validity
Reliability
What is discrimination?
Answers on a questionnaire corresponds to how much people have the characteristic - do people with different scores on the questionnaires actually differ in the construct being measured
whether the q can successfully tell people apart:
people with same scores should be equal to each other in the measured construct
people with different scores should be different to each other in the measured construct
degree of difference = difference in scores
What is validity?
Does the questionnaire measure what it intends too?
What is reliability?
Does the questionnaire produce the same results when completed under the same conditions?
questionnaires should produce the same results under the same conditions - each persons scores should be the same at different points in time
What are the types of validity?
Content
Criterion
Factorial
What is content validity?
Items relate to the construct of interest Are comprehensive (cover the entire construct) Do not overlap - ask same thing
How do you ensure for content validity?
Conduct thorough background research
Gather suggestions on items for a variety of sources
Ask experts in the field for feedback
Write good items
What is criterion validity?
Items do in fact measure the construct of interest in a meaningful way
How do you ensure for criterion validity?
Test predictions of the questionnaire
Correlate questionnaire with other established measures
Statistical testing
What is factorial validity?
Responses on certain items will stick together or correlate with each other - these groups are called a factor - does the factor structure you hypothesised correspond to reality
How do you test for factorial validity?
Factor analysis
Steps of creating a questionnaire
Choose a construct to measure
Decide on a response scale
Generate items (questions) to measure this construct
Collect the data
Analyse the data to discover what factors the items group in, whether this structure is reliable and whether the questionnaire is valid
Assess and write up the results
What is self-esteem?
One’s overall sense of worthiness as a person
Rosenberg scale
What is self-efficacy?
An optimistic sense of personal competence - accounting for motivation and accomplishments in human beings
What are the ways of quantifying responses?
Open and closed questions
open - choose answer, good as allows exploration of themes, doesn’t restrict answers but very difficult to quantify, can’t do statistical analysis
closed - fixed answer, content can only be certain choices. Make sure you have covered all of the options - need an other option
What are thy types of numeric response scales?
Ranking scales - assign them a rank of importance of doing something (1=most important)
Rating scales - rate how much you like/dislike doing something (1=not at all)
Semantic differentiations, used for attitude/emotions - rank where you fall on a scale with a number, pleasant to unpleasant for an activity
Likert scales - 5/7 points, include a midpoint (strongly agree/disagree). Avoid categorical questions or likert scales with an obvious right answer - leads to skewed data. offer a statement - answer how much this relates to them
What are rating scale effects?
The way you ask a question can dramatically effect the answer you get eg. how successful have you been in life:
0 - 10 - 34% gave values between 0-5
-5-5 - 13% gave values between -5-0 - people dont want to say they have been unsuccessful
What should you ensure with a rating scale?
It is kept consistent across your questionnaire - the way it is structured, biases you towards a response
What are the problems which can arise with wording effects?
Ambiguity - most serious threat to reliability, e.g. frequently may mean different things, better giving proper categories ‘less than once’
Leading questions - imply there is a right answer
Double negatives - are you against a ban
Double barrelled questions - cant respond to each thing individally
Acquiescence bias - tendency to agree or say yes, reverse the statement
Neutrally and clearly stating an idea is difficult to impossible
