measuring attitudes and expectancies lect 5 Flashcards
how is choice of data collection influenced
by the theory under scrutiny
-draws on previous research and looks at the methods used
explicit measures
-requires people to introspect
-take time to think about their views thoughts beliefs etc
-believed that introspections provide valid insight to how a person feels so measured using self report questionnaires
what are the issues to avoid in explicit research
-ambiguous qs: respondent may interpret diff to how you intended so therefore, not relevant to the subject of interest, hard to compare between people
-leading qs: encourages a certain response, response bias, responses are not authentic and do not reflect what people really think
-vague qs: where specific construct is not well defined pp may answer related not specific qs you are interested in
designing a questionnaire
-what is the issue/theory/topic
-dont try to cover too much: leads to long questionnaire so pp lose motivation
-keep them short and concise
-pilot studies helpful to establish what to include/exclude
question content
-wording of qs is important
-think about the respondents (do they have the necessary knowledge, target the survey to the people who have the info you need)
-ask a qs in a simple form
-use simple language and be specific
-avoid technical terms and slang
-void ambiguity as this adds noise to data
-keep in mind periodical qs (longer the time frame leads to more general estimates e.g how often do you exercise in a week/month)
how do you go about providing a scale for topic of interest
1.nominal scales (categorical)
2.likert scales (likert)
3.semantic differential scale (osgood 1957)
nominal scale
puts person in category, can get complicated with sub categories
likert scales
method of measuring attitudes in a way to allow for analysis, tried to capture degree of strength to which people might endorse a qs or concept
likert scales use a continuum where pp make ratings e.g 1=not at all to 5=very much, value given to each response category
semantic differential scale
-uses continuum with anchored adjectives
-respondents place mark with respect to adjective end points to indicate opinion strength
what is an issue encountered with a survey
desirability bias
-certain topics are sensitive/ value laiden/ embarrassing for pp
-pp may not want to be forthcoming so modify behaviour or not answer truthfully
-finding out motives for these behaviours is very hard
some solutions to social desirability bias
1.Marlowe-Crowne social desirability scale: asks pp their sensitivity to social desirability issues and gain measure of this to use as a covariate to control for this issue
2.’lie’ scale: including qs where most pp will answer differently to how they would actually behave
3.’bogus pipeline’: pretend lie detector, encourages pp to be truthful
4.implicit measure
what is the main implicit measure
-dominance of the explicit
-popular models are treated as explicit e.g theory of planned behaviour, theory of reasoned action, self efficacy etc
-BUT to what extent are people able to accurately introspect
-are there non conscious influences on peoples responding that are being overlooked
-if there are then the measure is not capturing a full picture as to why people behave in the way they do
nisbett and wilson research for dominance of the explicit
-set up stall in shopping centre
-asked passers by to select item they liked and why
-found most people selected items on the R of the display (R hand bias)
-when asked why people dont mention fact it is on R side
-pp unaware of R side bias when making judgements
-pp unaware so unable to make accurate introspections
behaving out of habit
Wood et al
-43% of peoples daily activities can be classed as habitual
-this means self report measures of motivation may overlook important processes related to unintended habitual behaviour
(processes outside of awareness/ implicit may be inaccessible to introspection)
implicit/ automatic attitudes
-evaluation/attitudes can be unintended or occur outside of conscious awareness (people do not always know they have made an evaluation/decision)
-evidence suggests that automatic attitudes can explain spontaneous behaviours
automatic attitudes and health
Swanson: found people who smoke have more pos implicit attitudes towards smoking than non smokers
Craeynest: children with obesity have more pos implicit attitude toward food than controls of a healthy weight
Calitri: regular exercisers had more implicit pos attitudes towards exercise than sedentary indiv
measuring implicit attitudes
-implicit association test IAT
-sequential priming
IAT
grennwald et al
-responding in attitude consistent way is faster than when attitude is inconsistent
-pp tasked with 2 classifications w. same keys on response panel
Z key for pos, N key for neg words
-keys used to assign objects to categories (pos/neg)
-where object key evaluation is congruent, responding will be faster
-e.g if someone likes flowers and dislikes insects it will be faster if they have to press Z key for daisy and N key for cockroach than vice versa
-pp given d score for difference in responding beween 2 blocks
considerations of the IAT
+very flexible (De Hauwer), can measure many types of association not just attitudes/valance
+ has relatively large effect sizes so is reliable
+Teige - Mocigemba: good internal reliability
-test retest reliability is weaker (implicit attitudes are more unstable rather than dispositional, might be subject to contextual factors that happen to be salient at particular time
what is a weakness of the IAT
karpinski and steinman 2006
-requires 2 categories to work
-may be some concepts where a viable comparison isnt obvious or appropriate
-doesnt tell you absolute positivity or neg of a given attitude
-it is a measure of one attitude relative to another
e.g if someone likes sweets more than salad on IAT it doesnt mean they dont like salas
how to measure implicit attitudes
sequential priming
-pp respond quickly to target stimulus on computer
-target primed with another stimulus which may facilitate or inhibit responding
-primes are presented very briefly so unavailable to conscious mind
-if prime and target share same valance, responding is faster
-actually what is being tested here is the prime
considerations of sequential priming
+you do not need a comparison category
-reliability can be low, reducing power
technical considerations
-type of monitor being used: constructing an image on a screen takes time, luckily implicit measures are usually on screen for good amount of time so doesnt matter
-for subliminal stimuli: 10-20 ms, image built up over time, screen may never had had the chance to present image so pp never saw it
-important to make sure the technology is capable
-keyboard or response device: inputs are cued behind each other so if computer is doing other things, response may seem delayed, use response pad for more precise timings
-software: gorilla, e prime, opensesame