Optional extra on LinReg and crickets

As a follow-up to Wednesday's linear regression problem about cricket chirp rates, I think the following extra activity will interest some of you. And even if you only have a few minutes to spend, I think you'll get something out of it.



On this web page read at least the first two sections about crickets and linear modeling.



You'll see a linear model similar to the class problem, but note these changes: the x and y variables are reversed; the chirp rates are per-minute rather than per-second; and the data is different enough that it's probably from a different species of cricket from the class problem.



A few cool things about the story on the linked web page:

• One researcher (Dolbert) had proposed a certain linear equation as the model, but then other researchers (Bessey) used recorded data and linear regression to get a more accurate equation.
  
• The page includes an applet where you can control the slope and intercept using sliders, to make your own best guess of the best fit line. This is pretty fun to play with -- try it yourself! (If it's not working on your computer, make sure your browser has Java applets enabled. Also, if you have a slow connection, it might take some time to load.)
  
• The page asks good questions about the domain: for what temperatures does the equation provide a realistic model? Identifying the realistic domain is always an important part of doing linear modeling and regression.
  

If you try this activity, please let me know what you think of it, OK? That will help decide whether to have the whole class try it next year.