On New Year's Day, Jeff started weighing himself every morning. Two charts are given below. One has his weight for the first eight Thursdays of the year (January 1, 2009 was a Thursday and we'll count it as Day 1) and the other has his weight for the first 10 Saturdays of the year.
[table — see PDF]
[table — see PDF]
(a) Find the least squares line for the Thursday data and comment on its goodness of fit.
(b) Find the least squares line for the Saturday data and comment on its goodness of fit.
(c) Use Quadratic Regression to find a parabola which models the Saturday data and comment on its goodness of fit.
(d) Compare and contrast the predictions the three models make for my weight on January 1, 2010 (Day #366). Can any of these models be used to make a prediction of my weight 20 years from now? Explain your answer.
(e) Why is this a Civics lesson in honesty? Well, compare the two linear models you obtained above. One was a good fit and the other was not, yet both came from careful selections of real data. In presenting the tables to you, I have not lied about my weight, nor have you used any bad math to falsify the predictions. The word we're looking for here is 'disingenuous'. Look it up and then discuss the implications this type of data manipulation could have in a larger, more complex, politically motivated setting.