Multiple Regression Analysis

Multiple Regression Analysis

Regression analysis:

Introduction:

Regression analysis entails finding out the relationship that exist between economic variables, regression involves the use of the linear classical model which gives directions to the process of regression, simple regression model entails the process of regressing only two variables where one variable is a dependent variable and the other variable is a independent variable. Multiple regression models entails the regression of more than two variables, in this case we have one dependent variable and more than one independent variables.

The regression process:

When we consider the stock market data below we can regress the variables depending on the relationship we specify, in regression we follow the following steps:

Model specification

This involves writing the model using theory or available literature, you also specify the signs

Estimation of the model:

This entails gathering data and then using the classical model to estimate the specified model

Evaluation of estimates:

This involves checking whether the estimates of the model are theoretically meaningful.

Evaluation of forecasting power:

This involves checking whether the estimated model has predictive power; this involves either changing the data size or data source.

The data:

Can you perform a multiple regression analysis with the data that has been collected without changing that data or adding to it? Why or Why not?

We can use multiple regression model on the data because we will have first to specify our model, if we estimate this model and find out that the estimated parameters are not statistically significant then there is need to add data, for the stock market data, we need to find out the dependent variable and the independent variable, however if we have a large stochastic variable then we will have to add data, if we have autocorrelation then we will have to add another independent variable.

Therefore for this data we can apply multiple regression model which when we estimate we will be in a position to check if the model is theoretically meaningful, we have to consider what determines the stock price by checking the volume traded and also the opening stock price, other factors will also influence the closing stock price which will depend on the model specification where one may add another or other independent variables.

Can you perform a multiple regression analysis with the data that has been collected and additional data derived from that collection? Why or Why not?

It is possible for us to regress a model using additional data regarding the data, this will increase the sample size, in statistics it always advisable to use a large sample, a large sample ensures that the estimated model is free of errors resulting from as very small sample, therefore adding data to the data is possible and recommended in order to increase the sample size which will ensure that we achieve a more accurate model that we estimate.

For the data it is also advisable that we add some more independent variable, the stock market price is too complex to be determined by two variables, there are other factors that will result into a change in the stock market price, in the file on regression the closing stock price depends on the opening and the stock volume traded, there are other factors that should be considered and this include the prices of other stock in the market and their volume traded and other factors which one may think of.

References:

Barnes and Noble (2007) the Barnes and Noble company, Retrieved September 28th ,

Schmidt P.(1996) Econometrics: introduction to econometrics, McGraw Hill publishers, New York



Source by Charles Kelly