Question

Assume you are a business manager that's considering the use of multiple regression analysis to gather data about what impacts consumer demand.

1. What must you do to make sure this tool is implemented wisely so it will
provide useful information?

2. What specific kind of data does multiple regression analysis provide, and
what makes such data change?

3. What is a basic time limitation of multiple regression analysis, and why is
there such a limitation?

Answer 1

1. The following points need to be taken into consideration in order to have useful prediction and information from multiple regression analysis
. The predictor may not be completely independent of each other and correlated with one another
. The multiple T-square value must be higher so that the model can be explained well by most of the raw data
. The P- value of the intercepts and slopes and the ANOVA F- value should be less than the type I error estimate to reject the null hypothesis that no relationship exists between the predictors and dependent variable

2. Multiple regression analysis provides the final output in terms of a linear equation of type
Y =a1X1+ a2X2 + ----------+ anXn+I
Where Y is the estimate of the dependents variable and ( X1,X2 ------ Xn ) are the predictor variables
ai' is the coefficient of 'Xi' I=1,2,-----n
'I' is the intercept
With know value of Xi we can predict the value of Y such estimate is subject to change if the raw data { the historical observation of (Xi,Y) change } or the confidence interval of the estimator change

3. Multiple regression is applicable generally on cross sectional data i,e the predictor Xi in the raw data must be collected in a single point of time. If the time gap is too large the variation of Y vs Xi may actually be due to the variation of both them independently in the time series and not due to the linear relationship between them so a multiply regression analysis os good when it is done on a short run basis and the historical data are actually not too