Measurements & Scaling

Question 1

scale characteristics

Answer 1

description (descriptors for each value)
order (sizes/positions of descriptors)
distance (b/w descriptors)
origin (beginning of scale)

Question 2

measurement vs scaling

Answer 2

measurement = actual assignment of a number
- ex: from1-100

scaling = creating a continuum upon which measured objects are located

Question 3

primary scales of measurement

Answer 3

nominal
ordinal
interval
ratio

Question 4

scaling techniques

Answer 4

1. comparative scales (direct comparison of stimulus objects)
   - paired comparison, rank order, constant sum
2. non-comparative scales (objects scaled independently)
   - continuous rating scales
   - itemized rating scales

Question 5

paired comparison

Answer 5

respondent is presented w/ 2 objects and asked to select one

• data obtained is ordinal

possible to convert paired comparison to rank order (assumption of transitivity)

Question 6

rank order

Answer 6

respondent is presented w/ several objects and asked to rank them according to some criterion

• data obtained is ordinal

Question 7

constant sum scaling

Answer 7

respondents allocate a constant sum of units (ex 100) to attributes of a product to reflect importance

Question 8

continuous rating scale

Answer 8

respondents place a mark on a position in a scale that runs from one extreme criterion to the other

ex: 1 completely unsatisfied – 10 completely satisfied

Question 9

itemized rating scale

Answer 9

scale has a number or brief description associated to each category
- Likert
- semantic differential
- Stapel

Question 10

Likert scale

Answer 10

indicate degree of agreement or disagreement to series of statements

• 1-5 scale
• profile analysis or total score calculation

Question 11

semantic differential scale

Answer 11

7 point rating scale w/ endpoints associated with bipolar labels w/ semantic meaning

• controls respondents with very positive/negative attitudes
• profile analysis

Question 12

Stapel scale

Answer 12

vertical unipolar rating scale w/ 10 categories numbered -5 to +5 (w/ no neutral point (zero))

• can be analyzed in the same way as semantic differential data (profile analysis)

Question 13

nominal

Answer 13

numbers only serve as labels for identification
- do not reflect characteristics
- statistics: frequency (percentages + mode)

Question 14

ordinal

Answer 14

ranking scale, n assigned to objects
- can determine whether or not an object has more or less of a characteristic but not how much
- statistics: centiles (percentile + median)

Question 15

interval

Answer 15

numerical equal distances on scale
- allows for comparison b/w objects
- statistics: frequency + percentile + arithmetic mean, st deviation…

Question 16

ratio

Answer 16

• possesses all properties of nominal, ordinal + interval scales
• meaningful to compute ratios (fixed zero point)
• all statistical techniques can be applied