Scribd
Upload
39 views9 pages

Multidimensional Scaling - 20MBA55

Multidimensional scaling (MDS) is a statistical technique that represents similarities or dissimilarities between datasets visually in a low-dimensional space. It shows how variables are related to one another based on their distances on the chart. MDS has advantages like simplifying complex datasets and showing psychological variable relationships. However, it also has disadvantages like being subjective and not representing real numeric distances. The steps involved in MDS include calculating dissimilarities between points, optimally scaling the points in a low-dimensional space, and assessing the reliability and validity of the solution. MDS is useful for market research applications like determining brand positions and identifying consumer segments.

Uploaded by

Sagar Gupta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
39 views9 pages

Multidimensional Scaling - 20MBA55

Multidimensional scaling (MDS) is a statistical technique that represents similarities or dissimilarities between datasets visually in a low-dimensional space. It shows how variables are related to one another based on their distances on the chart. MDS has advantages like simplifying complex datasets and showing psychological variable relationships. However, it also has disadvantages like being subjective and not representing real numeric distances. The steps involved in MDS include calculating dissimilarities between points, optimally scaling the points in a low-dimensional space, and assessing the reliability and validity of the solution. MDS is useful for market research applications like determining brand positions and identifying consumer segments.

Uploaded by

Sagar Gupta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 9

TOPIC: MULTIDIMENSIONAL SCALING

NAME: SAGAR KUMAR GUPTA


ENROLMENT NO.: 20MBA55
SUB.: RESEARCH METHODOLOGY AND SPSS LAB (MBA301)
What is Multidimensional Scaling?

 Multidimensional scaling is a visual representation of distances or dissimilarities between


sets of objects. “Objects” can be colors, faces, map coordinates, political persuasion, or
any kind of real or conceptual stimuli (Kruskal and Wish, 1978). 
 Multi-dimensional scaling (MDS) is a statistical technique that allows researchers to find
and explore underlying themes, or dimensions, in order to explain similarities or
dissimilarities (i.e. distances) between investigated datasets
 Multidimensional scaling is a method of expressing information visually. Rather than
show raw numbers, a multidimensional scale chart will show the relationships between
variables; things that are similar will appear close together while things that are different
will appear far away from one another.
Advantages & Disadvantages of Multidimensional Scales:

 Relationship Modeling
Multidimensional scales show how things stand in relation to one another. For example, if you made a multidimensional scale of city distances in the United States, Chicago would be
closer to Detroit than it would be to Phoenix.
A- An advantage of this method is that you can look at a multidimensional scale and immediately assess how closely related different values are.
D - A disadvantage, though, is that this technique doesn't deal in real numbers — a multidimensional scale of Boston, New York and Los Angeles would look roughly similar to a
multidimensional scale of London, Dublin and Buenos Aires, even though the actual figures are profoundly different.

 Simplifying Tables
A - A multidimensional scale is best used in situations where there's a large amount of data organized in table form. By converting it to a multidimensional scale, you can immediately
assess relationships, which is essentially impossible in a table with 10,000 or more different figures — an amount that's entirely feasible.
D - The disadvantage of this is that a complex formula is necessary to convert raw figures into a multidimensional scale. Therefore, while it's easy to see the relationships between figures,
it takes a large amount of effort to create the table. This means that if you're going to use a multidimensional scale, you need to be certain that there's an actual demand for the information
it's presenting. Otherwise, you're using your time now for no reason other than to save someone else time in the future.

 Application
A - Multidimensional scaling is generally used in psychology, graphing subject responses to various stimuli. This method is used because researchers can show relationships of importance
— i.e., how much importance is placed on different variables. This can be extremely useful, as psychological data tends to be high volume and have many different aspects.
D - A disadvantage of this is that it adds another layer of subjectivity to psychological data, as modeling tabled data into a multidimensional scale requires some decision-making. Which
data will go into the scale? Which multipliers will be used to create relationship figures? This has an effect on the multidimensional scale's accuracy.
Steps involved in MDS
Notation

 Although MDS is commonly used as a measure of dissimilarity, MDS can technically


measure similarity as well. Dissimilarity between two points r and s is denoted δrs and similarity is
denoted srs. Small δrs indicates values that are close together and larger values indicate values that
are farther apart (i.e. are more dissimilar). On the other hand, similarity values are the opposite:
small srs indicates values that are farther apart; larger values suggest more similarity (i.e. values are
closer together). Similarity measures are easily converted from one to another by a monotone
decreasing transformation (Buja et. al, 2007). NCSS (n.d.) gives the following formula for the
transformation:
Where:
•drs = a dissimilarity
•srs = a similarity

Other notation you may come across:

•i and j = sometimes used in place of s and r to indicate primary and secondary points.
•Δ = the matrix (usually n x n) representing the dissimilarities.
•drs = the distance between point r and point s (not to be confused with the dissimilarity
notation drs in the above conversion equation).
Assess Reliabilities and Validity

 The input data, and consequently the MDS solutions, are invariably subject to sub
random variability. Hence, it is necessary that some assessment be made of the
relation and validity of MDS solutions. The following guidelines are suggested.
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 analyzed. For Kruskal’s stress
formula 1, the recommendations for evaluating stress values are as follows.
Importance of Multidimensional Scaling (MDS) for Marketing

 Multidimensional scaling is a powerful tool that helps market researchers determine


brand or product position, or find market segments.
 To determine the number and nature of dimensions consumers use to perceive different
brands.
 To position brands on these dimensions.
 To identify the position of consumer’s ideal brand.
Limitations of MDS

 It is assumed that the similarity of stimulus A to B is the same as the similarity of B to A


but this assumption is not necessarily true.

You might also like