Practical Data Management and Analysis for Public Health Assignment 9

(1C) Run a linear regression model with the natural logarithm of pesticide concentration as the dependent variable and time since exposure as the independent variable.

Part B. For this Part of the assignment our aim is to describe how height varies in relation to age across a broader age range than we have had to grapple with before: from birth to age 20. Across such a wide range the relationship may not be well approximated by a straight line. We will therefore use linear regression to fit a curvilinear model to the data, specifically treating height as a quadratic function of age.

(B.1) Using the dataset DataAssign9B, create a scatterplot with height on the vertical axis and age on the horizontal axis.

(B.2) Use SPSS to create a mean-centered version of the variable AGE. To do this, you will need first to determine the mean value of AGE in this dataset, and then run a COMPUTE command in which you subtract the mean value of age from each person’s age.

(B.3) Use SPSS to create a new variable that is the square of the mean-centered age variable.

(B.4) Set up and run a linear regression model with height as the dependent variable and two independent variables: the mean-centered age variable and the square of the mean-centered age variable.

Part C. For this part of the assignment our objective is to determine whether the effect of enrollment in a weight loss program on actual weight lost varies in relation to individual’s motivation to lose weight. Plausibly, being enrolled in a weight loss program may lead to substantial weight loss for people who are motivated to lose weight; but for people who are not motivated to lose weight, enrollment in the program may make little difference. Our dataset contains the three variables of interest: amount of weight lost by each respondent (WL), the enrollment status of each respondent (ENROLL), and the baseline motivation level of each respondent (MOTIVATE). Using these data, we will fit a linear regression model that includes an interaction term between program enrollment and motivation.

(C.1) Using the dataset DataAssign9C, run a simple linear regression model using ENROLL to predict WL.

(C.2). Use SPSS to create a mean-centered version of MOTIVATE. To do this, you will first need to determine the mean value of MOTIVATE in this dataset, and then run a COMPUTE command in which you subtract the mean value of that variable from each person’s individual score.

(C.3) Next, create a new variable that is the interaction (or product) of ENROLL and your meancentered motivation scale score variable.

(C.4) Run a linear regression model in which weight loss is predicted by ENROLL, your meancentered motivation scale score variable, and the interaction between these two.

(C.5) Determine the version of the regression model that pertains at 1.5 standard deviations below the mean value of MOTIVATE. To do this, you’ll need to (a) find the standard deviation of MOTIVATE and multiply it by -1.5; (b) substitute that number for your centered motivation scale score variable in the regression model; and (c) do some algebra to simplify the resulting expression in the form WL = b0 + b1ENROLL.

(C.6) Determine the version of the regression model that pertains at 1.5 standard deviations above the mean value of MOTIVATE. To do this, you’ll need to (a) find the standard deviation of MOTIVATE and multiply it by +1.5; (b) substitute that number for your centered motivation scale score variable in the regression model; and (c) do some algebra to simplify the resulting expression in the form WL = a0 + a1ENROLL.

Part D. For the last part of this assignment, our aim is to determine how the effect of motivation to eat a low cholesterol diet on daily cholesterol intake depends upon self-efficacy. The thinking behind this is that self-efficacy is necessary to translate motivation into action. In other words, when people have very little self-efficacy to adhere to a low cholesterol diet, it may not matter very much how motivated they are. On the other hand, when people have high self-efficacy, they are able to translate their motivation into action. To explore this we will run a linear regression model in which motivation, self-efficacy, and the interaction between them all predict cholesterol consumption.

(D.1) Using the dataset DataAssign9D, run a simple linear regression model using MOTIVATE to predict DIETCHOL.

(D.2). Use SPSS to create MOTIVCENT, a mean-centered version of MOTIVATE. To do this, you will first need to determine the mean value of MOTIVATE in this dataset, and then run a COMPUTE command in which you subtract the mean value of that variable from each person’s individual score.

(D.3). Use SPSS to create SECENT, a mean-centered version of SELFEFF. To do this, you will first need to determine the mean value of SELFEFF in this dataset, and then run a COMPUTE command in which you subtract the mean value of that variable from each person’s individual score.

(D.4) Next, create a new variable that is the interaction (or product) of your mean-centered motivation and self-efficacy scale scores.

(D.5) Run a linear regression model in which DIETCHOL is predicted by your mean-centered motivation scale score variable, your mean-centered self-efficacy scale score variable, and the interaction between these two.

(D.6) Determine the version of the regression model that pertains at 1.5 standard deviations below the mean value of SELFEFF. To do this, you’ll need to (a) find the standard deviation of SELFEFF and multiply it by -1.5; (b) substitute that number for your centered self-efficacy scale score variable in the regression model; and (c) do some algebra to simplify the resulting expression in the form DIETCHOL = b0 + b1MOTIVCENT.

(D.7) Determine the version of the regression model that pertains at 1.5 standard deviations above the mean value of SELFEFF. To do this, you’ll need to (a) find the standard deviation of SELFEFF and multiply it by +1.5; (b) substitute that number for your centered self-efficacy scale score variable in the regression model; and (c) do some algebra to simplify the resulting expression in the form DIETCHOL = a0 + a1MOTIVCENT.

(D.8) Use MS Excel to create a graph that displays DIETCHOL = b0 + b1MOTIVCENT and DIETCHOL = a0 + a1MOTIVCENT across the range -2 to 2 for MOTIVCENT.

Part E. Write a report that summarizes the results you obtained in Parts A, B, C, and D. For Part A, what do the scatterplot and regression results suggest about change over time in the concentration of pesticide in the blood of exposed individuals? For Part B, what do scatterplot and regression results suggest about the shape of the relationship between age and height? For Part C, what do you conclude about how the effect of program enrollment on weight loss depends (or does not depend) upon motivation to lose weight? And for Part D, what do you conclude about how the effect of motivation on cholesterol intake depends upon self-efficacy?

Submission

In the Assignments section, please upload the following items to the Practice Assignment 9 page 24 hours before live session 9:

1. Your syntax file for carrying out the analyses requested in Parts A, B, C, and D above;

2. Your written report as described in Part E.