Model Summaries Worksheet
Every workbook produced by RegressIt contains not only the data analysis and regression worksheets but also a model summaries worksheet that keeps an audit trail of all regression models fitted so far and allows side-by-side comparison of models fitted to the same dependent variable, suitable for framing. Here one model was fitted to CASES_18PK and three models were fitted to CASES_18PK_LN (natural log of cases of 18-packs sold). For each model, the regression summary statistics and the coefficients and P-values of the variables are shown. The third model is a multiple regression model obtained by adding the other two (logged) price variables as predictors, and the fourth model adds a time index variable (Week) to correct for a slight trend in the errors. Their coefficients are all significantly different from zero, and their inclusion yields a substantial reduction in the standard error of the regression (from 0.356 to 0.244 in log units). The positive coefficients of the other two logged price variables reveal the presence of cross-price elasticities: consumers tend to buy more 18-packs when the prices of 12-packs or 30-packs are increased, other things being equal. More details of this analysis can be found in the RegressIt User Manual.
Click here to proceed to the next page: Forecasting.
Every workbook produced by RegressIt contains not only the data analysis and regression worksheets but also a model summaries worksheet that keeps an audit trail of all regression models fitted so far and allows side-by-side comparison of models fitted to the same dependent variable, suitable for framing. Here one model was fitted to CASES_18PK and three models were fitted to CASES_18PK_LN (natural log of cases of 18-packs sold). For each model, the regression summary statistics and the coefficients and P-values of the variables are shown. The third model is a multiple regression model obtained by adding the other two (logged) price variables as predictors, and the fourth model adds a time index variable (Week) to correct for a slight trend in the errors. Their coefficients are all significantly different from zero, and their inclusion yields a substantial reduction in the standard error of the regression (from 0.356 to 0.244 in log units). The positive coefficients of the other two logged price variables reveal the presence of cross-price elasticities: consumers tend to buy more 18-packs when the prices of 12-packs or 30-packs are increased, other things being equal. More details of this analysis can be found in the RegressIt User Manual.
Click here to proceed to the next page: Forecasting.
