Unit-III

Advanced Analytical Tools:

Meaning:-

“Conjoint analysis' is a survey-based statistical technique used in market research that helps
determine how people value different attributes (feature, function, benefits) that make up an
individual product or service.”

Conjoint analysis is the optimal market research approach for measuring the value that
consumers place on features of a product or service. This commonly used approach combines
real-life scenarios and statistical techniques with the modeling of actual market decisions.

The objective of conjoint analysis is to determine what combination of a limited number of

attributes is most influential on respondent choice or decision making. A controlled set of
potential products or services is shown to survey respondents and by analyzing how they make
preferences between these products, the implicit valuation of the individual elements making up
the product or service can be determined. These implicit valuations (utilities or part-worths) can
be used to create market models that estimate market share, revenue and even profitability of
new designs.

Important Terminology

• Factor: Independent variable the researcher manipulates that represents a specific attribute

• Level: Specific non-metric value describing a factor. Each factor must be represented by two
or more levels.

• Profile : Combination of all possible levels of factors. For ex: 3 factors with 2 levels each will
create (2x2x2) i.e. 8 profiles.

• Utility: An individual’s subjective preference judgment representing the holistic value or worth
of a specific object.
• Part-worth: Estimate from conjoint analysis of the overall preference or utility associated with
each level of each factor used to define the product or service

Advantages

 estimates psychological tradeoffs that consumers make when evaluating several attributes
together
 measures preferences at the individual level
 uncovers real or hidden drivers which may not be apparent to the respondent themselves
 realistic choice or shopping task
 able to use physical objects
 if appropriately designed, the ability to model interactions between attributes can be used
to develop needs based segmentation

Disadvantages
 designing conjoint studies can be complex

 with too many options, respondents resort to simplification strategies

 difficult to use for product positioning research because there is no procedure for
converting perceptions about actual features to perceptions about a reduced set of
underlying features

 respondents are unable to articulate attitudes toward new categories, or may feel forced to
think about issues they would otherwise not give much thought to

 poorly designed studies may over-value emotional/preference variables and undervalue

concrete variables

 does not take into account the number items per purchase so it can give a poor reading of
market share
Steps in Conjoint Analysis:-

1. Choose the product attributes

2. Choose the values or options for each attributes.
3. Choose the form in which the combinations of attributes are to be presented to the
respondents.
4. Decide how responses will be aggregated. There are three choices- Use individual
responses pool all responses in to a single utility function and define segments of
responses.

Conjoint analysis is the optimal market research approach for measuring the value that
consumers place on features of a product or service. This commonly used approach combines
real-life scenarios and statistical techniques with the modeling of actual market decisions.

Conjoint can help you determine pricing, product features, product configurations, bundling
packages, or all of the above. Conjoint is helpful because it simulates real-world buying
situations that ask respondents to trade one option for another.

For example, in a survey, the respondent is shown a list of features with associated prices. The
respondent then chooses what they want in their ideal product while keeping price as a factor in
their decision. For the researcher, key information can be gained by analyzing what was selected
and what was left out. If feature A for $100 was included in the menu question but feature B for
$100 was not, it can be assumed that this respondent prefers feature A over feature B.

Conjoint analysis is an advanced market research analysis method that attempts to understand
how people make complex choices. Every day, people make choices that require trade-offs—so
often that we may not even realize it. Simple decisions such as, “what type of laundry detergent
should I buy?” or “which flight should I book for my upcoming trip?” all contain multiple
elements that ultimately formulate a decision.

Subconsciously, some of us may be more price-sensitive while others are more feature-focused.
This is the core of conjoint—understanding which elements consumers consider to be important
or trivial.
Ever bought a house? As one of the more complex purchase decisions, homebuyers must
consider their preference of location, price, interest rates, proximity to work, modern
conveniences, entertainment, quality of schools, and among others.

Conjoint analysis is one of the most effective models in extracting consumer preferences during
the purchasing process into a quantitative measurement. It evaluates products or services in a
way no other method can.

Traditional rating surveys and analysis do not have the ability to place a value on the different
attributes that make up a product. Conjoint analysis extrapolates the respondent’s preference for
a quantitative measurement.

Conjoint Analysis enables businesses to mathematically analyze consumer or client behavior and
make decisions based on real insights from data! This allows them to better cater to consumer
needs and develop business strategies that provide a competitive edge. This is because the
fulfillment of customer wishes in a profitable way requires companies to fully understand which
aspects of their product and service are most valued by the customer.

Conjoint analysis is considered to be the best survey method for achieving this purpose. It
consists of creating, distributing and analyzing surveys among customers with the purpose of
modeling their purchasing decision based on response analysis.

 Types of Conjoint Analysis

There are two main types of conjoint analysis:

 Choice-based Conjoint (CBC) Analysis

 Adaptive Conjoint Analysis (ACA).

Choice-based Conjoint (CBC) Analysis: This type of conjoint analysis is the most popular
because it asks consumers to imitate the purchasing behavior in the real market: which products
they would choose, given certain criteria on price and features.

Adaptive Conjoint Analysis (ACA): This type of conjoint analysis is used more often in
scenarios where the number of attributes/features exceeds what can be done in a choice based
scenario. ACA is great for product design and segmentation research, but not for the price.
Conjoint Deliver Realistic and Actionable Insights. Although Conjoint analysis requires more
involvement in survey design and analysis, the benefits of realistic and actionable insights than a
standard survey alone are worth the additional planning effort. Price is important to most folks
shopping for a laptop. But how much more is the majority willing to pay for longer battery life
for their laptop if it means a heavier and bulkier design? How much less in value is a smaller
screen size compared to a slightly larger one? Find this out now before making a huge
investment.

The other thing to note is that conjoint is just a piece of the insights pie. Capture the full story
with a cohesive pricing, consumer preference, branding, or go-to-market strategy using other
question types and delivery methodologies to stretch the project to its full potential.
 Factor Analysis

Factor analysis is a statistical method used to describe variability among observed,

correlated variables in terms of a potentially lower number of unobserved variables
called factors.

Factor analysis is a technique that is used to reduce a large number of variables into fewer
numbers of factors. This technique extracts maximum common variance from all variables and
puts them into a common score. As an index of all variables, we can use this score for further
analysis.
Types of factoring:

There are different types of methods used to extract the factor from the data set:

1. Principal component analysis: This is the most common method used by researchers. PCA
starts extracting the maximum variance and puts them into the first factor. After that, it removes
that variance explained by the first factors and then starts extracting maximum variance for the
second factor. This process goes to the last factor.

2. Common factor analysis: The second most preferred method by researchers; it extracts the
common variance and puts them into factors. This method does not include the unique variance
of all variables. This method is used in SEM.

3. Image factoring: This method is based on correlation matrix. OLS Regression method is
used to predict the factor in image factoring.

4. Maximum likelihood method: This method also works on correlation metric but it uses
maximum likelihood method to factor.

5. Other methods of factor analysis: Alfa factoring outweighs least squares. Weight square is
another regression based method which is used for factoring.

Factor loading:
Factor loading is basically the correlation coefficient for the variable and factor. Factor loading
shows the variance explained by the variable on that particular factor. In the SEM approach, as a
rule of thumb, 0.7 or higher factor loading represents that the factor extracts sufficient variance
from that variable.
Eigen values: Eigen values is also called characteristic roots. Eigen values shows variance
explained by that particular factor out of the total variance. From the commonality column, we
can know how much variance is explained by the first factor out of the total variance. For
example, if our first factor explains 68% variance out of the total, this means that 32% variance
will be explained by the other factor.
Factor score: The factor score is also called the component score. This score is of all row and
columns, which can be used as an index of all variables and can be used for further analysis. We
can standardize this score by multiplying a common term. With this factor score, whatever
analysis we will do, we will assume that all variables will behave as factor scores and will move.

Criteria for determining the number of factors: According to the Kaiser Criterion, Eigen
values is a good criteria for determining a factor. If Eigen values is greater than one, we should
consider that a factor and if Eigen values is less than one, then we should not consider that a
factor. According to the variance extraction rule, it should be more than 0.7. If variance is less
than 0.7, then we should not consider that a factor.

Rotation method: Rotation method makes it more reliable to understand the output. Eigen
values do not affect the rotation method, but the rotation method affects the Eigenvalues or
percentage of variance extracted. There are a number of rotation methods available: (1) No
rotation method, (2) Varimax rotation method, (3) Quartimax rotation method, (4) Direct oblimin
rotation method, and (5) Promax rotation method. Each of these can be easily selected in SPSS,
and we can compare our variance explained by those particular methods.

Assumptions:

1. No outlier: Assume that there are no outliers in data.

2. Adequate sample size: The case must be greater than the factor.
3. No perfect multicollinearity: Factor analysis is an interdependency technique. There
should not be perfect multicollinearity between the variables.
 4. Homoscedasticity: Since factor analysis is a linear function of measured variables, it
does not require homoscedasticity between the variables.
5. Linearity: Factor analysis is also based on linearity assumption. Non-linear variables
can also be used. After transfer, however, it changes into linear variable.
6. Interval Data: Interval data are assumed.

Key concepts and terms:

Exploratory factor analysis: Assumes that any indicator or variable may be associated with any
factor. This is the most common factor analysis used by researchers and it is not based on any
prior theory.

Confirmatory factor analysis (CFA): Used to determine the factor and factor loading of
measured variables, and to confirm what is expected on the basic or pre-established theory. CFA
assumes that each factor is associated with a specified subset of measured variables. It
commonly uses two approaches:

What are factor loadings?

The relationship of each variable to the underlying factor is expressed by the so-called factor
loading. Here is an example of the output of a simple factor analysis looking at indicators of
wealth, with just six variables and two resulting factors.

Variables Factor 1 Factor 2

Income 0.65 0.11
Education 0.59 0.25
Occupation 0.48 0.19
House value 0.38 0.60
Number of public parks in 0.13 0.57
neighborhood
Number of violent crimes per year in 0.23 0.55
neighborhood
The variable with the strongest association to the underlying latent variable. Factor 1, is income,
with a factor loading of 0.65.

Since factor loadings can be interpreted like standardized regression coefficients, one could also
say that the variable income has a correlation of 0.65 with Factor 1. This would be considered a
strong association for a factor analysis in most research fields.

Two other variables, education and occupation, are also associated with Factor 1. Based on the
variables loading highly onto Factor 1, we could call it “Individual socioeconomic status.”

House value, number of public parks, and number of violent crimes per year, however, have high
factor loadings on the other factor, Factor 2. They seem to indicate the overall wealth within the
neighborhood, so we may want to call Factor 2 “Neighborhood socioeconomic status.”

Notice that the variable house value also is marginally important in Factor 1 (loading = 0.38).
This makes sense, since the value of a person’s house should be associated with his or her
income.
 Cluster Analysis:

Cluster analysis is a multivariate method which aims to classify a sample of subjects (or objects)
on the basis of a set of measured variables into a number of different groups such that similar
subjects are placed in the same group.

In marketing, it may be useful to identify distinct groups of potential customers so that, for
example, advertising can be appropriately targeted.

The goal of performing a cluster analysis is to sort different objects or data points into groups in
a manner that the degree of association between two objects is high if they belong to the same
group, and low if they belong to different groups.

Cluster analysis differs from many other statistical methods due to the fact that it’s mostly used
when researchers do not have an assumed principle or fact that they are using as the foundation
of their research.

• Cluster: a collection of data objects – Similar to one another within the same cluster –
Dissimilar to the objects in other clusters

• Cluster analysis – Grouping a set of data objects into clusters

Examples of Clustering Applications

• Marketing: Help marketers discover distinct groups in their customer bases, and then use this
knowledge to develop targeted marketing programs

• Land use: Identification of areas of similar land use in an earth observation database

• Insurance: Identifying groups of motor insurance policy holders with a high average claim cost
• City-planning: Identifying groups of houses according to their house type, value, and
geographical location

• Earth-quake studies: Observed earth quake epicenters should be clustered along continent
faults

Common Applications of Cluster Analysis

Marketing

Marketers commonly use cluster analysis to develop market segments, which allow for better
positioning of products and messaging. company to better position itself, explore new markets,
and development products that specific clusters find relevant and valuable.

Insurance

Insurance companies often leverage cluster analysis if there are a high number of claims in a
given region. This enables them to learn exactly what is driving this increase in claims.

Putting Clustering into Context

It’s easy to overthink cluster analysis, but our brains naturally cluster data on a regular basis in
order to simplify the world around us. Whether we realize it or not, we deal with clustering in
practically every aspect of our day-to-day lives.

For example, a group of friends sitting at the same table in a restaurant can be considered a
cluster.

In grocery stores, goods of a similar nature are grouped together in order to make shopping more
convenient and efficient.

This list of events during which we use clustering in our everyday lives could go on forever, but
perhaps it makes more sense to consider a more classic, archetypal example.

In biology, humans belong to the following clusters: primates, mammals, amniotes, vertebrates,
and animals. In this example, note that as we move down the chain of clusters, humans show less
and less similarities to the other members of the group. Humans have more in common with
primates than they do with other mammals, and more in common with mammals than they do
with all animals in general.

The Benefits of Cluster Analysis

Clustering allows researchers to identify and define patterns between data elements.
Revealing these patterns between data points helps to distinguish and outline structures which
might not have been apparent before, but which give significant meaning to the data once they
are discovered.

Once a clearly defined structure emerges from the dataset at hand, informed decision-making
becomes much easier.

The Different Types of Cluster Analysis

There are three primary methods used to perform cluster analysis:

Hierarchical Cluster

This is the most common method of clustering. It creates a series of models with cluster
solutions from 1 (all cases in one cluster) to n (each case is an individual cluster). This approach
also works with variables instead of cases. Hierarchical clustering can group variables together in
a manner similar to factor analysis.

Finally, hierarchical cluster analysis can handle nominal, ordinal, and scale data. But, remember
not to mix different levels of measurement into your study.

K-Means Cluster

This method is used to quickly cluster large datasets. Here, researchers define the number of
clusters prior to performing the actual study. This approach is useful when testing different
models with a different assumed number of clusters.

Two-Step Cluster

This method uses a cluster algorithm to identify groupings by performing pre-clustering first, and
then performing hierarchical methods. Two-step clustering is best for handling larger datasets
that would otherwise take too long a time to calculate with strictly hierarchical methods.

Essentially, two-step cluster analysis is a combination of hierarchical and k-means cluster

analysis. It can handle both scale and ordinal data, and it automatically selects the number of
clusters.
What Does The Clustering Process Look Like?

Step #1: Build and Distribute a Survey

Your survey should be designed to include multiple measures of propensity to purchase and the
preferences for the product at hand. It should be distributed to your population of interest, and
your sample size should be large enough to inform statistically-based decisions.

Step #2: Analyze Response Data

It’s considered best practice to perform a factor analysis on your survey to minimize the factors
being clustered. If after your factor analysis it’s concluded that a handful of questions are
measuring the same thing, you should combine these questions prior to performing your cluster
analysis.

After reducing your data by factoring, perform the cluster analysis and decide how many clusters
seem appropriate, and record those cluster assignments. You’ll now be able to view the means of
all of your factors across clusters.

Step #3: Take Informed Action!

Comb through your data to identify differences in the means of factors, and name your clusters
based on these differences. These differences between clusters are then able to inform your
marketing, allowing you to target precise groups of customers with the right message, at the right
time, in the right manner.

Applications for Cluster Analysis

Market Segmentation: Companies can’t connect with all of their customers, but by dividing
markets into groups of consumers with similar needs and wants, they can position themselves to
appeal to these unique segments. Consumers may be clustered on the basis of benefits sought
from the purchase of a product. Each cluster would consist of consumers who are relatively
homogeneous in terms of the benefits they seek, thus allowing companies to deploy targeted
marketing campaigns that promote the most alluring benefits and products to this consumer
segment.
Understanding Buyer Behaviors: Cluster analysis can be used to identify homogeneous groups
of buyers. Then, the buying behavior of each group can be examined separately on measures
such as favorite stores, brand loyalty, price willing to pay, frequency of purchase, etc.
Identifying New Product Opportunities: By clustering brands and products, competitive sets
within the market can be determined. Brands in the same cluster complete more fiercely with
each other than with brands in other clusters. A company can examine its current offerings
compared to those of its competitors to identify potential new product opportunities.
Selecting Test Markets: By grouping cities into homogeneous clusters, it is possible to select
comparable cities to test various marketing strategies.
Data Reduction: A researcher may be faced with a large number of observations that can be
meaningless unless they are classified into meaningful groups. Cluster analysis can help by
reducing the information from an entire population of sample to information about specific
groups.
Hypothesis Generations: Cluster Analysis is also useful when a researcher wishes to develop
hypotheses concerning the nature of the data or to examine previously stated hypotheses.
Steps for Cluster Analysis
Formulate the problem – Select the variables on which the clustering will be based. The
variables should describe the similarity between objects in terms that are relevant to the research
problem. The variables should be selected based on past research, theory, the hypotheses being
tested, or the judgment of the researcher.
Select a distance measure – An appropriate measure of distance needs to be selected to
determine how similar or dissimilar the objects being clustered should be. The most commonly
used measure is Euclidean distance.
Select a clustering procedure – Several clustering procedures have been developed and the one
most appropriate for the problem at hand should be chosen.
Decide on the number of clusters – The number of clusters can be based on theoretical,
conceptual, or practical considerations.
Interpret and profile clusters – This involves examining cluster centroids. The centroids
represent the mean values of the objects contained in the cluster on each of the variables.
Asses the validity of clustering – Some methods to validate the data quality include using
different methods of clustering and comparing the results or clustering on a smaller set of
variables (randomly deleted) and comparing the results with the entire set of variables.
 1. Multidimensional scaling

Multidimensional scaling (MDS) is a means of visualizing the level of similarity of individual

cases of a dataset. MDS is used to translate "information about the pairwise 'distances' among a
set of n objects or individuals" into a configuration of n points mapped into an abstract Cartesian
space.

From a non-technical point of view, the purpose of multidimensional scaling (MDS) is to

provide a visual representation of the pattern of proximities (i.e., similarities or distances) among
a set of objects. For example, given a matrix of perceived similarities between various brands of
air fresheners, MDS plots the brands on a map such that those brands that are perceived to be
very similar to each other are placed near each other on the map, and those brands that are pe

 When to Use MDS

Let’s say you were given a list of city locations, and were asked to create a map that included
distances between cities. The procedure would be relatively straightforward, involving nothing
more complicated than taking a ruler and measuring the distance between each city. However,
what if you were given only the distances between the cities (i.e. their similarities) — and not
their locations? You could still create a map — but it would involve a fair amount of geometry,
and some logical deductions. Kruskal & Wish (1978) — the authors of one of the first
multidimensional scaling books — state that this type of logic problem is ideal for
multidimensional scaling. You’re basically given a set of differences, and the goal is to create a
map that will also tell you what the original distances where and where they were located.
  A Simple Example

Multidimensional scaling uses a square, symmetric matrix for input. The matrix shows
relationships between items. For a simple example, let’s say you had a set of cities in Florida and
their distances:

The scaling produces a graph like the one below.

The very simple example above shows cities and distances, which are easy to visualize as a map.
However, multidimensional scaling can work on “theoretically” mapped data as well. For
example, Kruskal and Wish (1978) outlined how the method could be used to uncover the
answers to a variety of questions about people’s viewpoints on political candidates. This could
be achieved by reducing the data and issues (say, partisanship and ideology) to a two-
dimensional map.

Basic steps:

1. Assign a number of points to coordinates in n-dimensional space. N-dimensional

space could be 2-dimensional, 3-dimensional, or higher spaces (at least, theoretically,
because 4-dimensional spaces and above are difficult to model). The orientation of the
coordinate axes is arbitrary and is mostly the researcher’s choice. For maps like the one
in the simple example above, axes that represent north/south and east/west .

2. Calculate Euclidean distances for all pairs of points. The Euclidean distance is the “as
the crow flies” straight-line distance between two points x and y in Euclidean space. It’s
calculated using the Pythagorean theorem (c2 = a2 + b2), although it becomes somewhat
more complicated for n-dimensional space (see “Euclidean Distance in n-dimensional
space“). This results in the similarity matrix.

3. Compare the similarity matrix with the original input matrix by evaluating the stress
function. Stress is a goodness-of-fit measure, based on differences between predicted and
actual distances. In his original 1964 MDS paper, Kruskal wrote that fits close to zero are
excellent, while anything over .2 should be considered “poor”. More recent authors
suggest evaluating stress based on the quality of the distance matrix and how many
objects are in that matrix.

4. Adjust coordinates, if necessary, to minimize stress.

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 associations. However,
while FA requires metric data, MDS can handle both metric and non metric data.
MDS is more useful than FA if you are able to create a 2D map, as you’ll be able to visually
confirm your findings.

Perceptual Mapping

What is a perceptual Mapping?

A perceptual map is of the visual technique designed to show how the average target
market consumer understands the positioning of the competing products in the marketplace. In
other words, it is a tool that attempts to map the consumer’s perceptions and understandings in a
diagram.

The word ‘perceptual’ comes from the word ‘perception’, which basically refers to the
consumers’ understanding of the competing products and their associated attributes. (Note: In
some textbooks, perceptual maps are referred to as positioning maps, but for our purposes they
essentially have the same meaning.)

The most common presentation format for a perceptual map is to use two determinant attributes
as the X and Y axes of a graph, however there are also other formats that are sometimes used.

 “Perceptual maps measure the way products are positioned in the minds of consumers
and show these perceptions on a graph whose axes are formed by product
attributes.” (Kardes, Cronley, & Cline, 2011).

 “A perceptual map represents customer perceptions and preferences spatially by means

of a visual display”(Ferrell & Hartline, 2008).

You will note that both definitions highlight that the information is visually presented on some
form of graph or display. And that a perceptual map is designed to examine consumer
perceptions and understanding, primarily of products and their associated positioning.
What Does a Perceptual Map Look Like?

A typical perceptual map is a two-dimensional graph with a vertical axis (Y) and a horizontal
axis (X). Each axis consists of a pair of opposite attributes at each end.

For example, if the map’s purpose is to investigate how consumers feel about breakfast cereals,
the Y axis might have “great taste” at one end, and “poor taste” at the other. The X axis might
have “low nutritional value” at one end, and “high nutritional value” at the other end.

Each cereal brand is then plotted on the perceptual map based on how consumers perceive the
brand’s products relative to those attributes.

In order for a perceptual map to be as useful as possible, both the X and Y axes need to measure
attributes that consumers actually consider while making purchasing decisions in the real world.

Again, the best way to begin to understand how customers and potential customers view the
products and brands in a market is to survey them!

Types of Perceptual Maps

There are multiple types of perceptual maps.

Simple maps like the one above represent just two variables, but perceptual maps can also be
created using sophisticated multivariate factor analysis and statistical software.

Both simple and complex perceptual maps have their purposes.

The data required for simple maps can be collected more conveniently, which allows for quick
insights. Alternatively, more complex maps consisting of three or more variables can uncover
even deeper insights into the dynamics of a particular market, which allows for the identification
of future growth opportunities.

How To Build a Perceptual Map

For the sake of simplicity, let’s continue with our breakfast cereal example.

Step #1: Select Your Determinant Attributes

Determinant attributes are the variables that the customer factors into their decision to purchase
the cereal or not. These attributes will always be important to customers, as they help to
differentiate competitive offerings in the market.The best determinant attributes for your study
will vary depending on your topic of interest.

If you just want to learn the basic dynamics of a market, it’s best practice to consider the
attributes that are most important to the average consumer’s experience. This will entail some
generalization, but that’s fine while performing exploratory research. Remember that in this
example we’re using health and taste as our determinant variables.

Step #2: Build and Distribute a Survey

In order to establish a significant dataset to work with, we’d need to build and distribute a survey
to an audience that is relevant to our study.

Building and administering a survey in SurveyGizmo is simple and straightforward, which will
allow you to get to the actual perceptual mapping stage of your study faster. In this case, we’d
want to include questions that address the tastes and nutritional values of cereal products
provided by competitors in the cereal market. We could then leverage panel services to identify
an audience of routine cereal eaters.

Step #3: Create Scores for The Brands in The Market

It’s time to start analyzing your survey response data. Pay particular attention to the differences
in how respondents felt about cereals being offered by competing brands.

This data will often come in the form of qualitative data from some questions that are open
ended in your survey. In these cases, it’s essential to create numeric scales such as Likert Scales,
from which you can analyze long form responses and then assign the response a numerical value.

While considering our selected determinant attributes of health and taste, we would then need to
provide each brand with a comprehensive score. This score represents how the average target
consumer feels about the brand’s product offering.

Step #4: Create Your Perceptual Map

Once you’ve scored each brand, it’s time to build your perceptual map and plot the brands
accordingly.

Staying true to the structuring discussed earlier in this post, our X axis will represent nutritional
content, and our Y axis will represent taste. Once we plot our brands accordingly based on their
customer perspective scores, our perceptual map will be complete, and we’ll have an easily
digestible visualization to work with.

This theoretical perceptual map would look something like this:

 Discriminant Analysis

Discriminant Analysis is a statistical tool with an objective to assess the adequacy of a

classification, given the group memberships; or to assign objects to one group among a number
of groups. For any kind of Discriminant Analysis, some group assignments should be known
beforehand.

General Purpose

Discriminant function analysis is used to determine which variables discriminate between two or
more naturally occurring groups. For example, an educational researcher may want to investigate
which variables discriminate between high school graduates who decide (1) to go to college, (2)
to attend a trade or professional school, or (3) to seek no further training or education. For that
purpose the researcher could collect data on numerous variables prior to students' graduation.
After graduation, most students will naturally fall into one of the three categories. Discriminant
Analysis could then be used to determine which variable(s) are the best predictors of students'
subsequent educational choice.

A medical researcher may record different variables relating to patients' backgrounds in order to
learn which variables best predict whether a patient is likely to recover completely (group 1),
partially (group 2), or not at all (group 3). A biologist could record different characteristics of
similar types (groups) of flowers, and then perform a Discriminant function analysis to determine
the set of characteristics that allows for the best discrimination between the types.

Computational Approach

Computationally, discriminant function analysis is very similar to analysis of variance

(ANOVA). Let us consider a simple example. Suppose we measure height in a random sample of
50 males and 50 females. Females are, on the average, not as tall as males, and this difference
will be reflected in the difference in means (for the variable Height). Therefore, variable height
allows us to discriminate between males and females with a better than chance probability: if a
person is tall, then he is likely to be a male, if a person is short, then she is likely to be a female.

We can generalize this reasoning to groups and variables that are less "trivial." For example,
suppose we have two groups of high school graduates: Those who choose to attend college after
graduation and those who do not. We could have measured students' stated intention to continue
on to college one year prior to graduation. If the means for the two groups (those who actually
went to college and those who did not) are different, then we can say that intention to attend
college as stated one year prior to graduation allows us to discriminate between those who are
and are not college bound (and this information may be used by career counselors to provide the
appropriate guidance to the respective students).

To summarize the discussion so far, the basic idea underlying discriminant function analysis is to
determine whether groups differ with regard to the mean of a variable, and then to use that
variable to predict group membership (e.g., of new cases).

Analysis of Variance. Stated in this manner, the discriminant function problem can be rephrased
as a one-way analysis of variance (ANOVA) problem. Specifically, one can ask whether or not
two or more groups are significantly differentfrom each other with respect to the mean of a
particular variable. To learn more about how one can test for the statistical significance of
differences between means in different groups you may want to read the Overview section to
ANOVA/MANOVA. However, it should be clear that, if the means for a variable are
significantly different in different groups, then we can say that this variable discriminates
between the groups.

In the case of a single variable, the final significance test of whether or not a variable
discriminates between groups is the F test. As described in Elementary Concepts and ANOVA
/MANOVA, F is essentially computed as the ratio of the between-groups variance in the data
over the pooled (average) within-group variance. If the between-group variance is significantly
larger then there must be significant differences between means.

Multiple Variables. Usually, one includes several variables in a study in order to see which
one(s) contribute to the discrimination between groups. In that case, we have a matrix of total
variances and covariance’s; likewise, we have a matrix of pooled within-group variances and
covariance’s. We can compare those two matrices via multivariate F tests in order to determined
whether or not there are any significant differences (with regard to all variables) between groups.
This procedure is identical to multivariate analysis of variance or MANOVA. As in MANOVA,
one could first perform the multivariate test, and, if statistically significant, proceed to see which
of the variables have significantly different means across the groups. Thus, even though the
computations with multiple variables are more complex, the principal reasoning still applies,
namely, that we are looking for variables that discriminate between groups, as evident in
observed mean differences.

Stepwise Discriminant Analysis

Probably the most common application of Discriminant function analysis is to include many
measures in the study, in order to determine the ones that discriminate between groups. For
example, an educational researcher interested in predicting high school graduates' choices for
further education would probably include as many measures of personality, achievement
motivation, academic performance, etc. as possible in order to learn which one(s) offer the best
prediction.

Model. Put another way, we want to build a "model" of how we can best predict to which group
a case belongs. In the following discussion we will use the term "in the model" in order to refer
to variables that are included in the prediction of group membership, and we will refer to
variables as being "not in the model" if they are not included.

Forward stepwise analysis. In stepwise Discriminant function analysis, a model of

discrimination is built step-by-step. Specifically, at each step all variables are reviewed and
evaluated to determine which one will contribute most to the discrimination between groups.
That variable will then be included in the model, and the process starts again.

Backward stepwise analysis. One can also step backwards; in that case all variables are
included in the model and then, at each step, the variable that contributes least to the prediction
of group membership is eliminated. Thus, as the result of a successful Discriminant function
analysis, one would only keep the "important" variables in the model, that is, those variables that
contribute the most to the discrimination between groups.

F to enter, F to remove. The stepwise procedure is "guided" by the respective F to enter

and F to remove values. The F value for a variable indicates its statistical significance in the
discrimination between groups, that is, it is a measure of the extent to which a variable makes a
unique contribution to the prediction of group membership. If you are familiar with
stepwise multiple regression procedures, then you may interpret the F to enter/remove values in
the same way as in stepwise regression.

Capitalizing on chance. A common misinterpretation of the results of stepwise Discriminant

analysis is to take statistical significance levels at face value. By nature, the stepwise procedures
will capitalize on chance because they "pick and choose" the variables to be included in the
model so as to yield maximum discrimination. Thus, when using the stepwise approach the
researcher should be aware that the significance levels do not reflect the true alpha error rate,
that is, the probability of erroneously rejecting H0 (the null hypothesis that there is no
discrimination between groups).

Interpreting a Two-Group Discriminant Function

In the two-group case, discriminant function analysis can also be thought of as (and is analogous
to) multiple regression (see Multiple Regression; the two-group discriminant analysis is also
called Fisher linear discriminant analysis after Fisher, 1936; computationally all of these
approaches are analogous). If we code the two groups in the analysis as 1and 2, and use that
variable as the dependent variable in a multiple regression analysis, then we would get results
that are analogous to those we would obtain via Discriminant Analysis. In general, in the two-
group case we fit a linear equation of the type:

Group = a + b1*x1 + b2*x2 + ... + bm*xm

where a is a constant and b1 through bm are regression coefficients. The interpretation of the
results of a two-group problem is straightforward and closely follows the logic of multiple
regression: Those variables with the largest (standardized) regression coefficients are the ones
that contribute most to the prediction of group membership.