Multiple Regression Analysis sample essay

Before answering the question for this session long project, I have to note that a multiple regression model implies that there are more than one predictor variables. Otherwise, if the number of predictor variables is equal to one, regression in simple. In my case, there are truly three data samples. The question is whether I can utilize the values of the two samples in order to determine the average values of the third sample, and perform a multiple regression analysis. The definite answer to this question is “yes!” It is possible. The three categories of data are, as usual, stock prices of Coca-cola at New York Stock Exchange, the indexes of Dow Jones Industrial Averages, and the dates on which the observations were made. Now let us see how we can perform the analysis.

Generally, I can perform a multiple linear regression analysis for literally any three (or more) categories of data, provided there are more than one observations in each data set, and sample sizes are the same for all samples. For example, I can attempt to predict the length of hair of a woman, given the color of her nail polish, and the total years of schooling. What really astonishes, is that I would definitely find at least a minor correlation. From a statistical point of view, the two predictor variables would most probably have at least a little influence on the dependent variable. But the question is, who needs this prediction? Is it worth making? Is there any sense in making it at all? The same questions arise when I try to perform a multiple regression analysis for these three data categories. I shall take into account that I already performed a simple linear regression analysis, and concluded that there is a strong correlation between the two samples. Adding the dates as the second predictor variable would serve no use whatsoever, except only we can see the increase of both DJIA and COKE with the flow of time. In some situations for some people this information can be valuable, however it is a well-known fact that nothing is stable with stocks. Therefore, adding dates as a second independent variable makes no sense at all. The indexes and stock prices may both go down next month or even next week, and there still will be a correlation between them. However the dates are only a sequence of infinitely increasing integers, that is why putting them in the multiple linear regression model either as predictor variable, or as the dependent variable makes absolutely no sense whatsoever. But I added sales volume as the second predictor variable, the model would make much more sense. This multiple regression model would be meaningful and useful.

Statistically, one can find a correlation between the slope of rain and the number of newborn babies that day. This correlation may be very small, or it can even be significant. However, who needs this kind of prediction? Any regression model must be conducted with a useful purpose, so that a statistician (or anybody else involved) can make a logical inference about the events, and be able to make important and productive decisions based on this inference. Unfortunately, this multiple regression analysis with dates is pointless; the only use of it is that it perfectly demonstrates that not all regression models are useful and meaningful. The one with sales volumes, on the other hand, is truly useful and meaningful.
Order custom essay now!