University of Alberta

• STAT 252

• MULTIPLE LINEAR REGRESSION In a multiple linear regression model, the mean of a response variable is a linear function of several explanatory variables. The model and associated inferential tools are valid as long as the assumptions of independence, normality and constant variance are not seriously violated. In this assignment, you will use the multiple linear regression tools available in SPSS to study the relationship between the percentage of body fat and age, weight, height and various body circumference measurements for 252 men. You will use the first 200 observations in the data set to develop an equation to predict the body fat percentage of a subject based on the above numerical characteristics. Then you will test the validity of the equation using the data for the remaining 52 observations. Moreover, you will verify the regression model assumptions with the appropriate diagnostic tools in SPSS. Estimating Percentage of Body Fat It is well known that body fat is an important factor for health. The higher your percentage of fat above the average levels, the higher your health risk for some weight-related illness, like heart disease, high blood pressure, and diabetes. Therefore, estimating the percentage of body fat to assess your health is highly recommended by a variety of popular health books. One of the accurate techniques to obtain the percentage of body fat is underwater weighing in which body volume is computed as the difference between body weight measured in air and weight measured during water submersion. In other words, body volume is equal to the loss of weight in water with the appropriate temperature correction for the water's density. As the technique of underwater weighing though very accurate is very inconvenient and costly, some new techniques have been developed that require some simple body circumference measurements (e.g. abdominal circumference) and skin-fold measurements obtained by using a caliper. The new techniques also use height and weight of an individual to estimate the body fat percentage.