5. Measurement & Scaling

8/9/2019 5. Measurement & Scaling

http://slidepdf.com/reader/full/5-measurement-scaling 1/33

Fundamentals and Comparative Scaling



Measurement means assigning numbers or othersymbols to characteristics of objects according tocertain pre-specified rules.

◦ One-to-one correspondence between thenumbers and the characteristics beingmeasured.

◦ The rules for assigning numbers should be

standardized and applied uniformly.◦ Rules must not change over objects or time.



Scaling involves creating a continuum uponwhich measured objects are located.

e.g. an attitude scale from 1 to 100, where 1 =

Extremely Unfavorable, and 100 = ExtremelyFavorable.

Each respondent is assigned a number from 1 to 100

Scaling is the process of placing therespondents on a continuum with respect totheir attitude toward something/object.



A ranking scale -- indicate the relative extent to whichthe objects possess some characteristic.

Relative position of a object on some characteristic --but do not measure how much more or less.

Operations/Statistics allowed -- counting, percentages,mode, calculation of centiles, e.g., percentile, quartile,median, chi square.



Numerically equal distances on the scale -- representequal values in the characteristic.

Permits comparison of the differences betweenobjects.

Location of zero point is not fixed. Zero point andunits of measurement are arbitrary.

Any positive linear transformation of the form y = a +bx will preserve the properties of the scale.

Statistical techniques: all of those applicable tonominal and ordinal data

Arithmetic mean, standard deviation, correlation, t test, z test,and other statistics commonly used in marketing research.



Possesses all the properties of the nominal,ordinal, and interval scales.

It has an absolute zero point.

It is meaningful to compute ratios of scale values.

Only proportionate transformations of the form y= bx, where b is a positive constant, are allowed.

All statistical techniques are applicable

Ex. Height, weight, age, money, sales, cost,market share, no. of customers, annual revenue



S ca le Ba sicCharacter ist ic

C o m m o nE x a m p l e s

Market ingE x a m p l e s

N o m i n a l Num bers ident i

& c lass i fy obje

Numbering of

players

B rand nos. , s t

types

Percentages,

mode

Chi-squa

binomial

Or d ina l Nos. indicate th

relative positionof objects but n

the magnitude

differences

between them

Quality ranking

rankings of teain a tournam en

Preference

rankings, m arkposi t ion, s ocia

c lass

Percent i le,

median

Rank-ord

correlatioFr iedman

A N O V A

Ratio Zero point is fix

rat ios of s c ale

values c an be

compared

Length, weight Age , sa les ,

incom e, cos ts

Geometr ic

mean, harmon

mean

Coefficie

variation

P e rm issible S tatist iDescriptive Infere

Interval Differencesbetween objec t

Temperature(Fahrenheit)

At t i tudes,opinions

Range, mean,standard

Correlatiotests ,



LikertSemantic

Differential

Stapel

Scaling Techniques

NoncomparativeScales

Comparative

Scales

PairedCompariso

n

RankOrder

Constant Sum

Q-Sortand Other

Procedures

ContinuousRating Scales

ItemizedRating

Scales



Comparative scales -- direct comparison of objects.◦ Comparative scale data must be interpreted in relative

terms◦ have only ordinal or rank order properties.

Non-comparative scales -- each object isscaled independently of the others◦ Data are generally assumed to be interval or ratio

scaled.



Advantages

• Small differences between stimulus objects can bedetected

• Reduce halo or carryover effects from one judgment toanother

• Same known reference points for all respondents

• Easily understood and can be applied

Disadvantages

•

Ordinal nature of the data• Inability to generalize beyond the stimulus objects

scaled



A respondent is presented with two objects and asked toselect one according to some criterion

The data obtained are ordinal in nature

Paired comparison scaling is the most widely-usedcomparative scaling technique

With n brands, [n(n - 1) /2] paired comparisons are required

Assumption of transitivity -- convert paired comparison datato a rank order

Taste testing - Coca



Instructions: We are going to present you with ten pairs of shampoobrands. For each pair, please indicate which one of the two brands of shampoo you would prefer for personal use.

Clinic allclear

Pantene Sunsilk H&S Garnier

Clinic all clear 0 0 1 0

Pantene 1 0 1 0

Sunsilk 1 1 1 1

H&S 0 0 0 0

Garnier 1 1 0 1

No. of TimesPreferred

3 2 0 4 1

A 1 in a particular box means that the brand in that column was preferred over the

brand in the corresponding row. A 0 means that the row brand was preferred overthe column brand.



Respondents are presented with several objectssimultaneously and asked to order or rank themaccording to some criterion.

Rank order scaling also results in ordinal data.

Only (n - 1) scaling decisions need be made.



Preference for Toothpaste Brands

Brand Rank Order

1. Colgate _________

2. Dabur _________

3. Vicco _________

4. Neem _________

5. Anchor _________

6. Ajanta _________

7. Close Up _________

8. Pepsodent _________9. Himalaya _________

10. Cibaca _________



Respondents allocate a constant sum of units,such as 100 points to attributes of a product to

reflect their importance.

If an attribute is unimportant, the respondentassigns it zero points.

If an attribute is twice as important as someother attribute, it receives twice as many points.

The sum of all the points is 100 – Constant sum.



Please allocate 100 points among the eight attributes of bathing soap in a waythat your allocation reflects the relative importance to each attribute. The morepoints an attribute receives, the more important the attribute is. If an attributeis not at all important, assign it zero points. If an attribute is twice as importantas some other attribute, it should receive twice as many points.

Attributes Segment I Segment II Segment III

Mildness 8 2 4

Lather 2 4 17

Color 3 9 7

Price 53 17 9

Fragrance 9 0 19

Packaging 7 5 9

Moisturizing 5 3 20

Cleaning Power 13 60 15



Respondents evaluate only one object at a time

Consist of continuous and itemized rating scales.

Itemized rating scales

◦ Likert Scale

◦ Semantic Differential Scale

◦ Stapel Scale



Form of the continuous scale

Q. How would you rate Samsung Mobile Phones?

Version 1

Version 2

Version 3

Probablythe worst

Probablythe best

-----------------------I------------------------------ -----

Probablythe worst

Probablythe best

-----------------------I------------------------------ -----0 10 20 30 40 50 60 70 80 90 100

Probablythe worst

Probablythe best

-----------------------I------------------------------ -----0 10 20 30 40 50 60 70 80 90 100

Very bad Neither good Verygoodnor bad



Scale has a number or brief description associated witheach category.

The categories are ordered in terms of scale position, and

the respondents are required to select the specifiedcategory that best describes the object being rated.

The commonly used itemized rating scales are the Likert,semantic differential, and Stapel scales.



Likert scale requires the respondents to indicate a degree of agreement or disagreement with each of a series of statements aboutthe stimulus objects.

Analysis can be conducted on an item-by-item basis (profile analysis),or a total (summated) score can be calculated.

Statements SD D NA ND A SA

Sony sells high quality products

Sony has poor after sales service

I like to shop Sony products



Semantic differential is a seven-point rating scale with end pointsassociated with bipolar labels that have semantic meaning.

SONY IS:

Powerful --:--:--:--:-X-:--:--: Weak

Unreliable --:--:--:--:--:-X-:--: Reliable

Modern --:--:--:--:--:--:-X-: Old-fashioned

The negative adjective or phrase sometimes appears at the left sideof the scale and sometimes at the right.

Individual items on a semantic differential scale may be scored on

either a -3 to +3 or a 1 to 7 scale.



Stapel scale is a unipolar rating scale with tencategories numbered from -5 to +5, without a neutralpoint (zero). This scale is usually presented vertically.

SONY

+5 +5

+4 +4

+3 +3

+2 +2

+1 +1

HIGH QUALITY POOR SERVICE

-1 -1

-2 -2

-3 -3

-4 -4-5 -5

The data obtained by using a Stapel scale can beanalyzed in the same way as semantic differential data.



Scale Basic Characteristics

Examples Advantages Disadvantages

ContinuousRatingScale

Place a mark on acontinuous line

Reaction toTVcommercials

Easy to construct Scoring can becumbersomeunlesscomputerized

Itemized Rating Scales

Likert Scale Degrees of agreement on a 1(strongly disagree)to 5 (strongly agree)scale

Measurementof attitudes

Easy to construct,administer, andunderstand

Moretime-consuming

SemanticDifferential

Seven-point scalewith bipolar labels

Brand,product, andcompanyimages

Versatile Controversy asto whether thedata are interval

StapelScale

Unipolar ten-pointscale,-5 to +5,without a neutralpoint (zero)

Measurementof attitudesand images

Easy to construct,administer overtelephone

Confusing anddifficult to apply



Number of categories Balanced vs. unbalanced

Odd/even no. of categories

Forced vs. non-forced Verbal description

Physical form or configuration



1) Number of categories Although there is no single, optimal number,traditional guidelines suggest that thereshould be between five and nine categories

2) Balanced vs. In general, the scale should be balanced tounbalanced obtain objective data

3) Odd/even no. If a neutral or indifferent scale response is

of categories possible for at least some respondents,an odd number of categories should be used

4) Forced vs. In situations where the respondents arenon-forced expected to have no opinion, the accuracy of

the data may be improved by a non-forced scale

5) Verbal description An argument can be made for labeling all ormany scale categories. The category descriptionsshould be located as close to the responsecategories as possible

6) Physical form A number of options should be tried and thebest selected



ABC detergent is:ABC detergent is:

1 V Harsh -- -- -- -- -- -- -- V Gentle

2 V Harsh 1 2 3 4 5 6 7 V Gentle

3 ______ ______ ______ ______ ______ ______ ______

V Harsh Harsh Somewhat

Harsh

Neither HarshNor Gentle

Somewhat

Gentle

Gentle VGentle

4 -3 -2 -1 0 +1 +2 +3

V Harsh Neither harshNor Gentle

VGentle



Discriminant

Nomological

Convergent

Test/Retest

Alternative Forms

InternalConsistenc

y

Content

CriterionConstruct

Generalizabilit

y

Reliability Validity

ScaleEvaluation



Reliability can be defined as the extent to whichmeasures are free from random error, then the measureis perfectly reliable.

In test-retest reliability, respondents areadministered identical sets of scale items at twodifferent times and the degree of similarity between thetwo measurements is determined.

In alternative-forms reliability, two equivalent formsof the scale are constructed and the same respondents

are measured at two different times, with a differentform being used each time.



Internal consistency reliability determines the extent towhich different parts of a summated scale are consistent inwhat they indicate about the characteristic being measured.

In split-half reliability, the items on the scale are dividedinto two halves and the resulting half scores are correlated.

The coefficient alpha, or Cronbach's alpha, is the averageof all possible split-half coefficients resulting from differentways of splitting the scale items. This coefficient variesfrom 0 to 1, and a value of 0.6 or less generally indicatesunsatisfactory internal consistency reliability.

5. Measurement & Scaling

Documents

Transcript of 5. Measurement & Scaling