Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more ➡
Standard view
Full view
of .
×
0 of .
Results for:
P. 1
Multivariate Analysis

# Multivariate Analysis

Ratings: 0|Views: 33|Likes:

### Availability:

See More
See less

05/01/2013

pdf

text

original

MULTIVARIATE ANALYSIS
Nature of Multivariate AnalysisMost business problems are multi-dimensional. MVA helps to solve complex problemsInvestigation of one variable
–
Univariate AnalyisInvestigation of 2 variables
–
Bivariate AnalysisInvestigate 3 or more variables
–
Multi-variate analysisEg; sales dependent on profits alone
–
Classification of MV techniques1.Dependence methods2.Interdependence methodsIf a multivariate technique attempts to explain or predict the dependent variables onthe basis of 2 or more independent then we are analyzing dependence. Multipleregressionanalysis, multiple discriminant analysis, multi-variate analysis of variance andcanonicalcorrelation analysis are all dependence methods.Analysis of InterdependenceThe goal of interdependence methods is to give meaning to a set of variables or to seek togroupthings together. No one variable or variable subset is to be predicted fromtheothers or explained by them. The most common of these methods are factor analysis,cluster analysis and multidimensional scaling. A manager might utilize thesetechniquesto identify profitable market segments or clusters. Can be used for classification of similar cities on the basis of population size, income distribution etc;As in other forms of data analysis, the nature of measurement scales willdeterminewhich MV technique is appropriate for the data. The exhibits below show theselectionof MV technique requires consideration of the types of methods for dependent andinterdependent variables. Non-metric- nominal and ordinal scalesMetric- Interval and ratio scalesExhibit 1
–
independent variable is metric

Classification of dependence Methods
ANALYSIS OF DEPENDENCE
Multiple Regression analysis is an extension of bivariate regression analysis, whichallows forthe simultaneous investigation of the effect of two or more independentvariables on a single interval-scaled dependent variable. In reality several factorsarelikely to affect such a dependent variable.An example of a multiple regression equation is
Y = a + B1X1 + B2X2 + B3X3+ …………..BnXn + e
Where B0= a constant, the value of Y when all X values =0Bi= slope of the regression surface, B represents the regression coefficientassociatedwith each XE= an error term, normally distributed about a mean of 0Let us look at a forecasting example. Suppose a toy manufacturer wishes to forecast sales bysales territory. It is thought that c
ompetitor’s sales, the presence or absence of
a
company’s salesperson in the territory (a binary variable) and grammar school enrollment
arethe independent variables that might explain the variation in the sales of a toy. Thedata isfit in and the results from the mathematical computations are as followsY = 102.18 + 387X1 + 115.2X2 + 6.73X3R2 = 0.845F value 14.6The regression equationindicates sales are positively related to X1 and X2 and X3The coefficients B show the effects on the dependent variables of unit increases inanyindependent variable. The value of B2 = 115.2 indicates that an increase of Rs 115,200 intoysales is expected with an additional unit of X2. Thus it appears that adding acompanysalesperson has a very positive effect on sales. Grammar school enrollments alsohelp predict sales. An increase in 1 enrollment of students ( 1000) indicates a sales increaseof Rs 6730. A 1 unit increase in competitor sales volume X1 does not add much to thetoyman
ufacturer’s
sales.The regression coefficient can either be stated in raw score units(actual X values) or asstandardized coefficients(X values in terms of their standard deviation.When regressioncoefficient are standardized they are called as beta weights B an their values

indicate therelative importance of the associated X values especially when predictors areunrelated. If B1= .60 and B2 = .20 then X1 has three times the influence on Y as X2In multipleregression the coefficients B1 and B2 etc are called coefficients of  partialregression because the independent variables are correlated with other inde pendentvariables. The correlation between Y and X1, with the correlation that X1 and X2 haveincommon with Y held constant is partial correlation. Because the partial correlation between sales and X1 has been adjusted for the effect produced by variation producedinX2(and other independent variables) , the coefficient of correlation coefficient obtainedfromthe bivariate regression will not be the same as the partial coefficient in themultipleregression coefficient. N multiple regression the coefficient B1 is defineda partialregression coefficient for which the other independent variables are held constant.Thecoefficient of multiple determination indicates the percentage of variation inYexplained by the variation in the independent variables. R2 = .845 tells us thatthevariation in the independent accounted for 64.5% of the variance in thedependentvariable. Adding more of the independent variables in the equation explains more of thevariation in Y.To test for statistical significance an F-test comparing the different sources of variation isnecessary. The F test allows for testing the relative magnitudes of the sum of squaresdueto the regression (SSe) and the error sum of squares (SSr) with their appropriate degreesof freedomF = (SSr)/k/ (SSe) / (n-k-1)K= nos of independent variables N= nos of respondents or observations.Refer F tables and testhypothesis at .05 level of significanceIn the eg F ratio = 14.6df = df for numerator =k =3df for denominator n-k-1 = 8Accept or reject H0 on the basis of comparison between calculated andtable valueA continuous interval-scaled dependent variable is required in multipleregression asin bivariate regression, interval scaling is also required for the independent varia bles.However dummy variables such as the binary variable in our example may be utilized.Adummy variable is one that has two or more distinct levels 0 and 1Multiple regression is usedas a descriptive tool in three types of situations1.It is often used to develop a self-weightingestimating equation by which to predictvalues for a criterion variable (DV) from thevalues of several predictor variables(IV)2.A descriptive application of multiple regcalls for controlling for confoundingvariables to better evaluate the contribution of other variables- control brand andstudy effect of pricealone3.To test and explain casual theories-referred to as Path analysis-reg is used todescribe an entire structure of linkages that have advanced fromcasual theories4.Used as an inference tool to test hypotheses and estimate populationvaluesLet us look at the following eg for SPSSLet us assume that we use multiple regression toarrive at key drivers of customer usagefor hybrid mail. Among the explanatory variables arecustomer perceptions of (1) costspeed valuation, (2) security, (3) reliability, (4) receiver technology, (5) Impact/emotionalvalue. Let us choose the first 3 variables all measured on a 5

## Activity (0)

### Showing

AllMost RecentReviewsAll NotesLikes