3. Response Formats
Open ended
simple
probed (looking for more details)
Closed ended
dichotomous (two choices only)
Do you like this class Yes No
multiple category (need to decide how may are to
be picked)
Do you like this class? 1=not at all 2= somewhat 3=like 4=like very much 5=like a lot
Scaled response (a type of close ended)
Labeled
Do you like this class? 1=not at all 2= somewhat3=like 4=like very much 5=like a lot
Unlabeled
Do you like this class? 1 2 3 4 5
4.
5. Levels of Measurement
CATEGORICAL SCALES
Ordinal and Nominal Scales
to label objects
e.g., gender, occupation, yes/no, school year
To indicate rank order
e.g., preference, higher than/lower than, best
liked/least liked
Types of Analysis
counts, percentages
6. Levels of Measurement
METRIC SCALES
Interval and Ratio Scales
for rating
e.g., attitude scales, agree/disagree,
actual number
e.g., “how many times…”
Types of Analysis
counts, percentages, averages, correlations,
7. Comparing Metric and Categorical
Scales
What is your age? _____ years
Indicate the age range you belong to
< 18
18-25
25-35
36-45
46-55
56-65
> 65
8. Other Considerations
How many categories for rank order
How many scale items? (sensitivity)
Do you like this class? 1= not at all like 2 3 4 5=like very much
Do you like this class? 1= not at all like 2 3 4 5 6 7 8 9 10 11 12 13 14 15=like very much
Even or odd number of scale items
Balanced versus unbalanced scales
Please rate the service at this restaurant
1= very poor 2=poor 3=neutral 4=good 5=very good
Please rate the service at this restaurant
1= poor 2=fair 3=good 4=very good 5= excellent
9. Commonly Used Scales
Likert Scale
5 point agree/disagree scale with a set of
statements
measures intensity of feeling
10. Semantic Differential
bipolar adjectives as anchors
scale levels are usually not given
descriptors
typically 7 point scales are used
used for evaluation of objects, brands, etc.
11. Reverse Scaling Issues
Why do you need reverse scaling
Avoid halo effect
Check to see if respondent is paying attention
Consistency of responses
I consider Brand A to be a quality brand
1= strongly disagree 2=disagree 3=neutral 4=agree 5=strongly agree
Brand A is poorly made
1= strongly agree 2=agree 3=neutral 4=disagree 5=strongly disagree
12. Some other scales
graphic scale
Stapel scale
negative and positive numbers
14. How to choose a question
questionnaire administration method (data
collection mode)
ability of the respondent
what kind of analysis do you need (scale
level desired)
e.g., do you want to report a count, a
percentage, an average
Previous research studies
15. What makes a good scale?
1. Reliability
consistency of scale
2. Validity
how accurate is the scale
16. Testing for Reliability
administering same questions to same
respondents at two different times
by using call backs
respondents may remember earlier responses
if time gap is too long, conditions may change
Use reverse scaled items
Repeat some items
17. Testing for Validity
Face validity
does the question measure what it is supposed
to measure
this is a subjective evaluation