Vous êtes sur la page 1sur 25

Multidimensional scaling and conjoint analysis Dr James Abdey

Applied Marketing (Market Research Methods) Topic 12: Multidimensional scaling and conjoint analysis
Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Overview

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem

Multidimensional scaling allows the perceptions and preferences of consumers to be clearly represented in a spatial map Conjoint analysis helps to determine the relative importance of attributes that consumers use in choosing products

Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Multidimensional scaling
Multidimensional scaling (MDS) is a class of procedures for representing perceptions and preferences 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 called spatial maps. The axes of the spatial map denote the dimensions respondents use to form perceptions and preferences for stimuli

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Formulate the problem

Multidimensional scaling and conjoint analysis Dr James Abdey

Specify the purpose for which the MDS results would be used Select the brands or other stimuli to be included in the analysis. The number of brands or stimuli selected normally varies between 10 and 20 The choice of the number and specic brands or stimuli to be included should be based on the statement of the market research problem, theory and the judgement of the researcher

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Example: Similarity judgements on pairs of mineral water brands


Perception 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, i.e. similarity judgements Very dissimilar 1 1 . . . 1 Very similar 7 7 . . . 7

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Evian vs. Perrier Volvic vs. Perrier . . . Evian vs. Volvic

2 2 . . . 2

3 3 . . . 3

4 4 . . . 4

5 5 . . . 5

6 6 . . . 6

The number of pairs to be evaluated is n(n 1)/2, where n is the number of stimuli

Obtain input data


Perception data: derived approaches. Derived approaches to collecting perception data are attribute-based approaches requiring the respondents to rate the brands or stimuli on the identied attributes using semantic differential or Likert scales, for example: Tastes pure . . . Pleasant tasting Not taste pure . . . Unpleasant tasting

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

If attribute ratings are obtained, a similarity measure (such as Euclidean distance) is derived for each pair of brands

Obtain input data direct vs. derived approaches

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling

The direct approach has the following advantages and disadvantages:


The researcher does not have to identify a set of salient attributes The disadvantages are that the criteria are inuenced by the brands or stimuli being evaluated Furthermore, it may be difcult to label the dimensions of the spatial map

Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Obtain input data direct vs. derived approaches


The attribute-based approach has the following advantages and disadvantages:
It is easy to identify respondents with homogeneous perceptions The respondents can be clustered based on the attribute ratings It is also easier to label the dimensions A disadvantage is that the researcher must identify all the salient attributes, a difcult task

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Preference data
Preference data order the brands or stimuli in terms of respondents preference for some property A common way in which such data are obtained is through preference rankings Alternatively, respondents may be required to make paired comparisons and indicate which brand in a pair they prefer Another method is to obtain preference ratings for the various brands

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Decide on the number of dimensions


A priori knowledge: Theory or past research may suggest a particular number of dimensions Interpretability of the spatial map: Generally, it is difcult to interpret congurations or maps derived in more than three dimensions Ease of use: It is generally easier to work with two-dimensional maps or congurations than with those involving more dimensions Statistical approaches: For the sophisticated user, statistical approaches are also available for determining the dimensionality

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

A spatial map
Each point in the spatial map represents a brand, allowing similarities of perceptions or perceptions to be determined
Spatial map of brands
x 1.5

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli

x 1.0 x x

Select a conjoint analysis procedure

Dimension 2

0.5

0.0

0.5

1.0

x x 1.5 1.0 0.5 0.0 0.5 1.0 1.5

2.0

Dimension 1

Basic concepts in conjoint analysis


Conjoint analysis attempts to determine the relative importance consumers attach to salient attributes and the utilities they attach to the levels of attributes The respondents are presented with stimuli that consist of combinations of attribute levels and asked to evaluate these stimuli in terms of their desirability Conjoint procedures attempt to assign values to the levels of each attribute so that the resulting values or utilities attached to the stimuli match, as closely as possible, the input evaluations provided by the respondents

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Formulate the problem


Identify the attributes and attribute levels to be used in constructing the stimuli The attributes selected should be salient in inuencing consumer preference and choice and should be actionable A typical conjoint analysis study involves six or seven attributes At least three levels should be used, unless the attribute naturally occurs in binary form (two levels), such as convertible and non-convertible cars The researcher should take into account the attribute levels prevalent in the marketplace and the objectives of the study

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Construct the stimuli

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview

In the pairwise approach, also called two-factor evaluations, the respondents evaluate two attributes at a time until all the possible pairs of attributes have been evaluated In the full-prole approach, also called multiple-factor evaluations, full or complete proles of brands are constructed for all the attributes. Typically, each prole is described on a separate index card

Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Mobile phone attributes

Multidimensional scaling and conjoint analysis Dr James Abdey

Assume three mobile phone attributes of interest: battery life, screen size and weight Assume each attribute has three different levels, coded 1, 2 and 3:
Battery life: Long, medium, short Screen size: Large, medium, small Weight: Heavy, medium, light

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Mobile phone proles and their ratings


Prole number 1 2 3 4 5 6 7 8 9 Attribute levels Screen size Weight 1 2 3 1 2 3 1 2 3 1 2 3 2 3 1 3 1 2

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data

Battery life 1 1 1 2 2 2 3 3 3

Preference rating 9 7 5 6 5 6 5 7 6

Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Select a conjoint analysis procedure


The basic conjoint analysis model may be represented by the following formula:
m ki

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data

U (X ) =
i =1 j =1

ij xij

Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli

where
U (X ) = overall utility of an alternative ij = the part-worth contribution or utility associated with the j -th level (j = 1, 2, . . . , ki ) of the i -th attribute (i = 1, 2, . . . , m) xij = 1 if the j -th level of the i -th attribute is present, 0 otherwise ki = number of levels of attribute i m = number of attributes

Select a conjoint analysis procedure

Select a conjoint analysis procedure


The importance of an attribute, Ii , is dened in terms of the range of the part-worths, ij , across the levels of that attribute: Ii = max(ij ) min(ij ) for each i

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis

The attributes importance is normalised to ascertain its importance relative to other attributes, Wi : Wi = Ii
m i = 1 Ii m

Formulate the problem Construct the stimuli Select a conjoint analysis procedure

such that
i =1

Wi = 1

The simplest estimation procedure, and one which is gaining in popularity, is dummy variable regression. If an attribute has ki levels, it is coded in terms of ki 1 dummy variables

Select a conjoint analysis procedure


The estimated model may be represented as:

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview

U = b0 + b1 X1 + b2 X2 + b3 X3 + b4 X4 + b5 X5 + b6 X6 where
X1 , X2 = dummy variables representing battery life X3 , X4 = dummy variables representing screen size X5 , X6 = dummy variables representing weight

Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

For battery life the attribute levels were coded as follows: X1 X2 Level 1 1 0 Level 2 0 1 Level 3 0 0

Mobile phone data coded for dummy variable regression


Preference ratings Y 9 7 5 6 5 6 5 7 6 Attribute levels Battery life Screen size X1 X2 X3 X4 1 0 1 0 1 0 0 1 1 0 0 0 0 1 1 0 0 1 0 1 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem

Weight X5 X6 1 0 0 1 0 0 0 1 0 0 1 0 0 0 1 0 0 1

Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Select a conjoint analysis procedure


The levels of the other attributes were coded similarly. The parameters were estimated as follows: b0 b1 b2 b3 = 4.222 = 1.000 = 0.333 = 1.000 b4 = 0.667 b5 = 2.333 b6 = 1.333

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Given the dummy variable coding, in which level 3 is the base level, the coefcients may be related to the part-worths: 11 13 = b1 12 13 = b2

Select a conjoint analysis procedure


To solve for the part-worths, an additional constraint is necessary 11 + 12 + 13 = 0 These equations for the rst attribute, battery life, are: 11 13 = 1.000 12 13 = 0.333 11 + 12 + 13 = 0 Solving these equations, we get: 11 = 0.778 12 = 0.556 13 = 0.222

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Select a conjoint analysis procedure


The part-worths for other attributes can be estimated similarly For screen size we have: 21 23 = b3 22 23 = b4 21 + 22 + 23 = 0 For the third attribute, weight, we have: 31 33 = b5 32 33 = b6 31 + 32 + 33 = 0

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

Select a conjoint analysis procedure


The relative importance weights were calculated based on ranges of part-worths, as follows: Sum of ranges of part-worths = (0.778 (0.556)) + (0.445 (0.556)) +(1.111 (1.222)) = 4.668 Relative importance of battery life Relative importance of screen size Relative importance of weight = = =
1.334 4.668 1.001 4.668 2.333 4.668

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure

= 0.286 = 0.214 = 0.500

Assumptions and limitations of conjoint analysis


Conjoint analysis assumes that the important attributes of a product can be identied It assumes that consumers evaluate the choice alternatives in terms of these attributes and make trade-offs The trade-off model may not be a good representation of the choice process Another limitation is that data collection may be complex, particularly if a large number of attributes are involved and the model must be estimated at the individual level

Multidimensional scaling and conjoint analysis Dr James Abdey

Overview Multidimensional scaling Formulate the problem Obtain input data Decide on the number of dimensions Conjoint analysis Formulate the problem Construct the stimuli Select a conjoint analysis procedure