MULTI-DIMENSIONAL

SCALING
B Y- P R A C H I G U P TA
AKSHIT SHARMA
S H I VA M B A J A J
D A A M A N B H A G AT

1
 MULTIDIMENSIONAL SCALING
(MDS)
• Multidimensional scaling (MDS) is a class of procedures for representing
perception and preference of respondents spatially by means of a visual display.

• Perceived or psychological relationships among stimuli are represented as

geometric relationships among points in a multidimensional space.

• These geometric representations are often call spatial maps. The axes of the
spatial map are assumed to denote the psychological bases or underlying
dimensions respondents use to form perceptions and preference for stimuli.

2
What is MDS?
Multidimensional scaling (MDS) is a technique that creates a map
displaying the relative positions of a number of objects, given only a
table of the distances between them.
 why is it called Multidimensional scaling ?
multidimensional because researcher is not limited to two
dimensional data or graph and scaling comes from psychometrics,
where object are assigned numbers according to trochim rule, 2006
 

3
The map may consist of one, two, three, or even more
dimensions. The program calculates either the metric or the
non-metric solution.
. The table of distances is known as the proximity matrix. It
arises either directly from experiments or indirectly as a
correlation matrix.
The program offers two general methods for solving the
MDS problem-;
1. Metric, or Classical, Multidimensional Scaling (CMDS)
2. Non-Metric Multidimensional Scaling (NMMDS)

4
USE OF MDS
 The number and nature of dimensions consumers use to perceive different brands in
market place.
 Positioning of current brands on various dimensions
 Image measurement.
 Market segmentation.
 New product development.
 Assessing advertising effectiveness.
 Pricing analysis.
Channel decisions.

5
TERMS ASSOCIATED WITH MDS
 Similarity Judgements: Similarity judgements are ratings on all possible pairs of brands or
other stimuli in terms of their similarity using a Likert type scale.

 Preference ranking: Preference ranking are rank ordering of the brands or other stimuli from
the most preferred to the least preferred. They are normally obtained from the respondents.

 Stress: This is a lack-of-fit measure; higher values of stress indicate poorer fits. It is not
necessary that an MDS map have zero stress in order to be useful. A certain amount of
distortion is tolerable. Different people have different standards regarding the amount of
stress to tolerate. The rule of thumb that we use is that anything under 0.1 is excellent and
anything over 0.15 is unacceptable.

6
 Spatial map: Perceived relationships among brands or other stimuli
are represented as geometric relationships among points in a
multidimensional space called a spatial map. The axes of the spatial
map denotes the underlying dimensions respondents used to form
perceptions.

 Coordinates: Coordinates indicate the positioning of a brand or a

stimulus in a spatial map.

7
Multidimensional scaling vs. Factor Analysis

 Both Multidimensional scaling (MDS) and factor analysis (FA)

uncover hidden patterns or relationships in data.
Both techniques require a matrix of measures of association.
However, while factor analysis requires metric data,
Multidimensional scaling can handle both metric and non metric
data.
Multidimensional scaling is more useful than factor analysis if you
are able to create a 2D map, as you’ll be able to visually confirm your
findings.

8
CONDUCTING
MULTIDIMENSIONAL SCALING

9
 FORMULATE THE PROBLEM
• Formulating the problem requires that the researcher specify the purpose for which
the MDS results would be used and select the brand or other stimuli to be included in
analysis.
• The number of brands or stimuli selected and the specific brands included
determine the nature of resulting dimensions and configurations.
• At a minimum, eight brands should be included so as to obtain a well defined spatial
map. Including more than 25 brands is likely to cumbersome and may result in
respondent fatigue.
• The choice of the number and specific brands or stimuli to be included should be
based on the statement of the marketing research problem, theory, and the
judgement of the researcher.

10
OBTAIN INPUT DATA

11
PRECEPTION DATA
Direct Approaches: In direct approaches to gathering perception data, the respondents
are asked to judge how similar or dissimilar the various brands or stimuli are using their
own criteria. Respondents are often required to rate all possible pair of brands or
stimuli in terms of similarity on a Likert scale. These data are referred to as similarity
judgments. For example, similarity judgment on all possible pairs of car brands may be
obtained in the following manner.
VERY VERY
DISSIMILAR SIMILAR
Toyota vs. Ford 1 2 3 4 5 6 7
Toyota vs. Jaguar 1 2 3 4 5 6 7
Toyota vs. Hyundai 1 2 3 4 5 6 7
-
-
-

12
 Similarity Rating of Car Brands
The number of pairs to be evaluate is n(n-1)/2, where n is the number of brands or stimuli.
TOYOTA FORD JAGUAR HYUNDAI NISSAN HONDA MERCEDES PORSCHE AUDI
TOYOTA 0
FORD 5 0
JAGUAR 6 7 0
HYUNDAI 4 6 6 0
NISSAN 3 3 4 5 0
HONDA 2 3 4 4 5 0
MEERCEDES 2 2 2 3 5 5 0
PORSCHE 2 2 2 2 6 5 6 0
AUDI 1 2 4 2 6 6 7 6 0

13
Derived Approaches: Derived approaches to collect perception data are attribute-based
approaches requiring the respondent to rate the brand or stimuli on the identified
attributes using semantic differential or Likert scale example, the different brands of car
may be rated on attributes such as these. Sometimes an ideal brand is also included in
the stimulus set. The respondent asked to evaluate their hypothetical ideal brand on the
same set of attribute.
HYUNDAI

EXPENSIVE ____ ____ ____ ____ ____ ____ ____ ____ CHEAP

SPORTY ____ ____ ____ ____ ____ ____ ____ ____ CONSERVATIVE

GOOD DESIGN ____ ____ ____ ____ ____ ____ ____ ____ BAD DESIGN

14
OBTAINING DATA
DIRECT vs. DERIVED APPROACH
The direct approach has following advantages and disadvantages:

Advantage

• The researcher dos not have to identify a set of salient attributes.

DISADVANTAGES

• The disadvantages are that the criteria are influenced by the brands or stimuli being
evaluated.

15
Furthermore, it may be difficult to label the dimensions of the spatial map.

The attribute-based approach has the following advantages and disadvantages:

Advantages

• It is easy to identify respondents with homogeneous perceptions.

• The respondents can be clustered based on the attribute ratings.
• IT is easier to label the dimensions

Disadvantages.

• A disadvantage is that the researcher must identify all the salient attributes, a difficult task.
• The spatial map obtained depends upon the attributes identified.

16
CONCLUSION FOR USING
APPROCHES
It may be best to use both these
approaches in a complementary way.
Direct similarity judgements may be used
for obtaining the spatial map, and attribute
ratings may be used as an aid to
interpreting the dimensions.

17
 PREFERENCE DATA
• Preference data order the brand or stimuli in terms of respondent preference for some property.
• A common way in which such data are obtained is preference rating. Respondent are required to
rank the brand from the most preferred to the least preferred.
• Alternatively, respondent may be required to make paired comparison and indicate which brand
in a pair they prefer.
• The configuration derived from preference data may differ greatly from that obtained from
similarity data. Two brands may be perceived as different in similarity map yet similar in a
preference map and vice versa.
• From example, Mercedes and BMW may be perceived by a group of respondents as very different
brands and thus appear far apart on a perception map. However these two brands may be about
equally preferred and appear close together on a preference map.

18
SELECT MDS PROCEDURE
Selection of a specific MDS procedure depends upon:
 Whether perception or preference data are being scaled, or whether the
analysis requires both kinds of data.
 The nature of the input data is also a determining factor.

• Non-metric MDS: This assumes that input data are ordinal, but they result in
the metric output. These procedures find, in a given dimensionality, a spatial
map whose rank orders of estimated distances brands and stimuli best
preserve or reproduce the input rank orders.

19
 • Metric MDS: Metric MDS methods assume that input data are metric.
Because the output is also metric, a stronger relationship between the
output and input data is maintained, and the metric (interval or ratio)
qualities of the input data are preserved.
 The metric and non-metric methods produce similar results.
 Another factor influencing the selection of a procedure is whether the MDS
analysis will be conducted at the individual respondent level or at an aggregate
level.

20
• Individual level analysis: The data is analysed separately for each
respondent, resulting in a spatial map for each respondent. Although
individual level analysis is useful from a research perspective, it is not
appealing from managerial standpoint. Marketing strategies are typically
formulated at the segment or aggregate level, rather than at the individual
level.

• Aggregate-level analysis: Some assumption must be made in aggregating

individual data. Typically it is assumed that all respondents use the same
dimensions to evaluate the brand or stimuli, but that different respondents
weight these common dimensions differentially.

21
DECIDE ON NO. OF
DIMENSIONS
 A prior knowledge: Theory or past research may suggest a particular number
of dimensions.
 Interpretability of the spatial map: Generally, it is difficult to interpret
configurations or maps derived in more than three dimensions.
Ease of use: It is generally easier to work with two-dimensional maps than
with those involving more dimensions.
 Elbow criterion: A plot of stress versus dimensionality should be examined.

22
23
LABEL THE DIMENSIONS AND
INTERPRET THE
CONFIGURATION
1. Even if direct similarity judgments are obtained, ratings of the brands or researcher supplied
attributes may still be collected. The axes may then be labelled for the attributes with which
they are most closely aligned.
2. After providing direct similarities or preference data, the respondents may be asked to
indicate the criteria the used in making their evaluations. These criteria may then be subjectively
related to the spatial map to label the dimensions.
3. If possible, the respondents can be shown their spatial maps and asked to label the
dimensions by inspecting the configurations.
4. If objective characteristics of the brands are available, these could be used as an aid in
interpreting the subjective dimensions of spatial maps.

24
A Spatial Map of Car Brands

25
26
ASSESS RELIABILITY AND
VALIDITY
1. The index of fit, or R -square, should be examined. This is squared correlation
index that indices the proportional of variance of the optimally scaled data that
can be accounted for the MDS procedure. Thus, it indicates how well the MDS
model fits the input data. Although higher values of R-square are desirable,
values of 0.60 or better are considered acceptable.
2. Stress values are also indicative of the quality of MDS solutions. Whereas R-
square is a measure of goodness of fit, stress measures badness of fit, or the
proportion of variance of the optimally scaled data that is not accounted for the
MDS model. Stress values vary with the type of MDS procedure and the data
being analysed.

27
 Stress (%) Goodness of fit
20 Poor
10 Fair
5 Good
2.5 Excellent
0 Perfect

3. If an aggregate-level analysis has been done, the original data should be split
into two or more parts. MDS analysis should be conducted separately on each
part and the results compared.

28
SCALING PREFERENCE DATA
• In internal analysis of preference, a spatial map representing both brands or
stimuli and respondent points is solely derived from preference data.
• In external analysis of preference, the ideal points or vectors based on
preference data are fitted in a spatial map derived from perception.
• The representation of both brands and respondents as points in the same
space, by using internal or external analysis, is referred to as unfolding.

29
LIMITATIONS OF MDS
1. It assumed that the similarity of stimulus A to B is the same as the similarity
of stimulus B to A.
2. MDS assumes that the distance between two stimuli is some function of their
partial similarities on each of several perceptual dimensions.
3. When a spatial map is obtained, it is assumed that interpoint distances are
ratio scaled and that the axes of the map are multidimensional interval scale.
4. A limitation of MDS is that dimension interpretation relating physical changes
in brands or stimuli to changes in the perceptual map is difficult at best.

30
SOURCES CITED
o https://www.du.ac.in/du/uploads/departments/Operational Research/24042020_Lect-10 Mu
ltidimensional Scaling.pdf

oMultidimensional Scaling (analytictech.com)

oMultidimensional Scaling (MDS) for Marketing (michaelpawlicki.com)

oUses of Multidimensional Scaling (citeman.com)

oMultidimensional Scaling (netdna-ssl.com)

31
32