Print / Share

Student Research Forum

Jeff Williams: "Forecasting Area Home Sale Prices"

 Forecasting Area Homes Research Poster

Research Category

Empirical Study

Research Purpose

The purpose of my research is to use a multiple linear regression model to predict home values in a particular area.  Having learned more about multiple linear regression models in ISDS 702, I wanted to apply it in a practical manner to further my understanding and perhaps deepen my appreciation of its use.

Research Methodology

Data was collected on seventy homes in the Golden Meadows subdivision in Bossier City.  This data was entered into an Excel computer program. I designated sales price as the dependent variable. The following were designated as independent variables:  home square footage, pool or no, home age, fireplace of no, and garage or no.  Then I used the programs to round several regression models and conducted F-tests and T-tests in order to reach a better forecasting model.

Research Conclusion

After conducting several F-tests and T-tests, I established a significant regression equation employing only the significant variables. 


Predicted Sales Price = 79,904.65 + 58.46 (Square Feet) +17,855.08(Pool) - 857.49 (Year Old)

The results show that the established regression equation has proven to be useful in forecasting area home sale prices.  After I shared my output with a real estate broker that specializes in the Golden Meadows housing market, he affirmed my conclusion. In addition, this project has expanded my knowledge and deepened my appreciation for the practical application of regression analysis.

Social Implications

The established forecasting model can help real estate brokers make a better predictive value for each potential home sale.  Future research may include additional independent variables and/or conducting a cross section comparison.