Attitude measurement
Attitude measurement
* Self-reports of attitudes are highly context
dependent and minor changes in question
wording, question format, and question order
can profoundly affect the obtained results
* Key components of the response process:
Comprehension ­ retrieval ­ judgment ­ formatting - editing
ˇ Question comprehension
* Information retrieval and judgment
 Question order effect
 Response order effect
* Response formatting
 Categorical response alternatives
 Rating scales
* Response editing
Question comprehension
1 9
4-4
How successful have you been in your life?
Information retrieval and judgment
 Question order effect
 Response order effect
Question order effect
General satisfaction Marriage satisfaction
General satisfactionMarriage satisfaction
1.
2.
1.
2.
r= .43 r= .67
"We now have two questions about your life. The first one pertains
to your marriage and the second to your life in general."
r= .18
"Aside from your marriage, which you already told us about, how
satisfied are you with your life in general"
r= .20
WHAT TO DO?
Response order effect
* Written vs. Read alternatives
* Written ­ primacy effect
* Read ­ recency effect
* Response order effects are more significant for
older and less educated respondents
Response formatting
* Categorical Response Alternatives
 Any opinion omitted from the list is unlikely to be reported
 Q: "What is the most important thing for children to prepare
them for life?"
 A: "To think for themselves"
* Rating Scales
 The effect of numeric values
 Context dependent (answer is compared with previous
answer)
61.5% 4.6% Advice: Provide "I don't know"
alternative
Response editing
* Respondents may hesitate to report their
attitude when they are concerned that their
answer may present them in a negative light.
* Social desirability
Attitude measurement tools
* Likert scale
* Semantic differential
* Error choice technique
Likert Scale
* An one-dimensional scale that consists of
attitude-related statements and asks individuals
to indicate their agreement/disagreement with
each statement along a numerical scale
Stan is the best teacher in the world.
Strongly
agree
Somewhat
agree
Neutral Somewhat
disagree
Somewhat
disagree
How to create the Likert scale?
1) Generate more questions than you need
2) Get a group of people to score the questionnaire
3) Sum the scores for all items
4) Calculate the intercorrelations between all pairs of
items
5) Reject questions that have a low correlation with
the overall score
6) For each item, calculate the t-value. Reject
questions with lower t-values (higher t-values
show questions with higher discrimination).
Semantic differential
* Osgood, Suci, Tannenbaum (1957)
* Respondents rate the attitude object on a number of (typically) 7point
bipolar scales that are end-anchored by evaluative adjective
pairs
* (connotative meaning of words (freedom, marriage); ideas
(capitalism); products (cell phones); brands (Volvo); etc...)
good
harmful
kind cruel
beneficial
dishonest
bad
honest
Barack Obama
Error-choice technique (Hammond, 1948)
* Attitudes can be indirectly measured by the bias
shown in answering a mixture of factual and nonfactual
questions concerning a given topic
* Respondents think it is a knowledge test
The average IQ of black people is:
A: 87% B: 113%
The average IQ of Japanese people is:
A: 85% B: 115%
Summary ­ Attitude measurement
* Comprehension ­ retrieval ­ judgment ­ formatting ­ editing
* Likert Scale
* Semantic Differential
* Error-Choice Technique