I have run a linear regression for the first part of the below question. The sca

Question

I have run a linear regression for the first part of the below question. The scatterplot shows it is not a good predictive model. And the linear regression shows there is a relationship between BMI and Weekly minutes of exercise. How do I then answer the last part of the question? I have another variable AgeGrp A that is categorised into the two age groups (under 45 and over 45). How do I analyse the two continuous variables (BMI) and (WeeklyMinutesExercise), which are both continuous variables, with the categorical variable AgeGrpA?

Consider body mass index BMI (BMI) as a continuous variable. Use weekly minutes of moderate exercise hours (WeeklyMinutesModerateExercise) in minutes per week as a continuous variable to predict BMI. Do you think this will provide a good predictive model? Also, does the relationship between BMI and weekly minutes of moderate exercise hours differ between those people under the age of 45 and greater than or equal to 45 years old?

Residuals: 1 Median Max Min 3Q -15. 177 4.326 1.085 3.106 40.423 Coefficients Estimate Std. Error t value Pr (>lt I) (Intercept) 28.6767537 0.2374252 120.782 16 WeeklyMinutes ModerateExercise r 0.0016322 0.0004762 3.427 0.000636 Signif codes 0 0.001 0.01 0.05 0.1 1 Residual standard error: 6.448 on 932 degrees of freedom. (66 observations deleted due to missingness) Multiple R-squared 0.01245, Adjusted R-squared 0.01139 F-statistic: 11.75 on 1 and 932 DF, p-value 0.000636

Explanation / Answer

Your model would be

BMI=28.67-.0016322*Weekly exercised

Now from the model we can see that R squared value is 1.245% that means the model can explain just 1.245% of the variability. Hence the model is not a good fit.

But from the table we can see that there is a strong relationship between weekly exercised and BMI, this suggest that there exist a model which can explain the variability. You can add few more variable into the model and check the r squared value, it has to increase to certain extent.

We can conclude that although there exists a stong relationship between BMI and weekly exercised, but weekly exercised alone is not sufficient to establish a model.

Now you can group the data between two groups namely, below 45 and above 45. Try analysis of variance to explain whether there is any significant chnage in BMI due to exercised in these two group.

If you want to use a model than you can use a multiple regression model with a dummy variable taking two values , 0 , 1 and fit a regression line on this.

model would be:

BMI= coeff + beta1 * weekly exercised + beta2 * dummy variable

Related Questions

Navigate

• Browse (All)
• Browse I
• Subjects

---