Scribd
Upload
323 views

Overview of Multivariate Methods

This document provides an overview of multivariate analysis techniques. It discusses how multivariate analysis can be used to analyze multiple measurements and enable better decision making. Any analysis of more than two variables can be considered multivariate analysis. Common multivariate techniques include principal components analysis, multiple regression, discriminant analysis, and cluster analysis. Measurement scales, validity, reliability, and statistical power and error are also important concepts for multivariate analysis. A six-step approach is outlined for successful multivariate model building and analysis.

Uploaded by

Mohammad Dabat
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
323 views

Overview of Multivariate Methods

This document provides an overview of multivariate analysis techniques. It discusses how multivariate analysis can be used to analyze multiple measurements and enable better decision making. Any analysis of more than two variables can be considered multivariate analysis. Common multivariate techniques include principal components analysis, multiple regression, discriminant analysis, and cluster analysis. Measurement scales, validity, reliability, and statistical power and error are also important concepts for multivariate analysis. A six-step approach is outlined for successful multivariate model building and analysis.

Uploaded by

Mohammad Dabat
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 18

Applied Statistics in

Business Administration

 Professor Mohammad Al-Nouaimi

 Prepared by :
Jihad Saadeh
Hamza Al-Abed

29-03-2020
Overview of Multivariate
Methods

 Multivariate analysis techniques enable


researchers to improve their decision
making.
 Multivariate analysis is used to analyze
multiple measurements under
investigation.
Overview of Multivariate
Methods

 Any analysis of more than two variables


can be considered for multivariate analysis.
 Some researchers use multivariate to
examine the relationship between the
means of the desired variables
Overview of Multivariate
Methods

 Many multivariate techniques are


extensions of univariate analysis
(analysis of single-variable distributions)
and bivariate analysis (cross-
classification, correlation, analysis of
variance, and simple regression used to
analyze two variables).
The Variate

 The Variate: multivariate analysis consists


of building blocks called the variates, A
linear combination of variables with
empirically determined weights.
 The variables = specified by the researcher
 weights = determined by the multivariate
technique to meet a specific Objective.
The Variate

 A variate of n weighted variables (X1 to Xn)


can be stated mathematically as:
1. Variate valueW1X1+W2X2 +W3X3 . . .
+WnXn
2. Xn = observed variable
3. Wn =weight determined by the multivariate
technique.
Measurement Scales

 Measurement is essential while identifying


variation.
 According to the characteristics they represent.
Data is classified into two categories :
1. nonmetric (qualitative)
2. metric (quantitative)
NONMETRIC
MEASUREMENT SCALES
 Nonmetric : Describe differences by
indicating the presence or absence of a
specific property.

1.Nominal Scales: assigns numbers as a way


identify subjects objectives

2. Ordinal Scales: subjects are compared to


each other in terms of a “greater than” or
“less than” relationship.
METRIC MEASUREMENT
SCALES

 Metric data: difference between subjects is


identified as an amount or a certain degree.

 The two metric measurement scales are


1. Interval scales : use an arbitrary zero point
2. ratio scales : include an absolute zero point
THE IMPORTANCE OF
CHOOSING THE BEST
MEASUREMENT SCALE

 It is very important to select the appropriate


measurement scale of each variable being
examined as it is critical when choosing the most
applicable multivariate technique to be used.
Measurement Error and
Multivariate
Measurement

 Measurement error: describes to what


degree the observed values are not
represented of the true values.
 It is a must to assume that the multivariate
technique used to have some degree of
measurement error.
VALIDITY AND
RELIABILITY

 Validity: identifies to what extent a measure


accurately represents what it is supposed
to.

 Reliability: it is the opposite of


measurement error. The more the reliability
the lower the measurement error.
Types of Statistical Error
and Statistical Power
 Type 1 error : known as alpha is the
probability of rejecting a true null
hypothesis. It is described as a rining
fire alram with no fire existing.
 Type 2 error, or beta (β). The Type 2
error is the probability of not rejecting a
false null hypothesis. It is described as a
fire alarm failing to ring while there is
fire.
Impacts on Statistical
Power
 Power is determined by three factors:

1) Effect size helps to determine if the observed


relationship is meaningful , logical or not.

2) Alpha: it is recommended to select the


appropriate alpha level and compare it to the risk
associated with a false conclusion.

3) Sample size: the bigger the sample size the


more powerful is the statistical test.
A CLASSIFICATION OF
MULTIVARIATE TECHNIQUES
 A dependence technique : when one or a set of
variables is identified as the dependent variable to
be predicted by other variables known as
independent variables.

 Interdependence technique : the variables


cannot be classified as either dependent or
independent.
10 TYPES OF
MULTIVARIATE

TECHNIQUES
1.Principal components and common factor analysis
 2. Multiple regression and multiple correlation
 3. Multiple discriminant analysis and logistic regression
 4. Canonical correlation analysis
 5. Multivariate analysis of variance and covariance
 6. Conjoint analysis
 7. Cluster analysis
 8. Perceptual mapping
 9. Correspondence analysis
 10. Structural equation modeling and confirmatory
factor analysis
Model Building : Six-step
approach to a successful
multivariate analysis
 1. Define the research problem, objectives, and
multivariate technique to be used.
 2. Develop the analysis plan.
 3. Evaluate the assumptions.
 4. Estimate the multivariate model and evaluate fit.
 5. Interpret the variates.
 6. Validate the multivariate model.
 Thank you

You might also like