Homework 8
Math 114Q, Sections 9, 11

Due in class Thursday, November 20.

  1. This is (essentially) a problem from an old Math 114 final.

    AT&T offers the following three calling plans for long distance land-line phones:

    Calling Plan	        Description	             Monthly Fee  Long Distance rate
    Unlimited Plus Plan	Unlimited calling to         $29.95        Included in monthly fee
                            anyone in the U.S., 24
                            hours a day, 7 days a week	
    
    One Rate 7 Plus Plan	One low long distance rate,  $3.95         7 cents per minute
                            all day every day!
    
    One rate 10 Plan	One low long distance rate,  None	  10 cents per minute 
                            all day every day
                            with no monthly fee	
    

    Use Excel to draw one chart showing how the monthly bill (y axis) depends on the number of minutes you use the phone for long distance calls (x axis) for each plan.

    Create a sequence of cells in column A for the various possible numbers of minutes. (Start with 0. What's a good step to use? What's a reasonable place to stop?) Don't type in all the entries by hand: use Excel to create the list.

    Use columns B, C and D for each of the three plans. The fixed charge and charge per call should be in cells in those columns too, so you can use the same formula everywhere in the data table. (That will call for judicious use of the '$' to keep Excel from changing row numbers when you don't want it to.)

    Since we want to see how you did the problem as well as the result, print the spreadsheet (with the graph on it) so that we can see the formulas, not the values. There's a trick for that. If you click anywhere in the spreadsheet outside the graph and then hold down the control key (it may be something else on the mac) while typing the backwards single quote (`) Excel will display the formulas, not the values (I learned this trick from Excel help). Do that, print the spreadsheet, and then control-` again to see the numbers.

    Write a paragraph explaining to your friend how she should go about choosing the plan that's best for her. (Hint: finding the places where the lines in your Excel chart cross is the key to the problem. You can do that with Excel experiments or with algebra.)

  2. Use Excel to plot toll as a function of miles travelled for a car starting in Boston and driving to the end of the Mass Pike. You can get that data here: www.masspike.com/user-cgi/tollcalc.cgi.

    You can get the tolls all on one page from http://www.masspike.com/pdf/tolls/toll_class1_08.pdf but need to look one interchange at a time to get the miles.

    Does the graph look linear? Ask Excel to find a regression line. What is its slope (with proper units, please.) How good a fit is it?

  3. In class we talked about correlation and causation. The important point is that the observation that data is correlated does not mean that there is a causal effect. We will look at an example to illustrate this. Open the Excel spreadsheet TVData.xls. The data there show the life expectancy (the average for men and women) for each country, along with the number of people per tv set.

    1. Which countries have the highest and lowest life expectancy at birth (if you are not sure what life expectancy means, look it up!) Which countries have the highest and lowest number of people per television set?
    2. Make a scatter plot of the data and find the trendline and R-value. Write a sentence that explains the slope of the trendline.
    3. Does a low number of people per television set cause high life expectancy? Would people in countries with low life expectancy live longer if we sent them shiploads of television sets?
    4. Does a high life expectancy cause there to be fewer people per television set? If we improved the life expectancy in a country by providing better medical care would that cause there to be fewer people per television set?
    5. What else could be going on here? Why might high life expectancy be strongly correlated with a low ratio of people per tv set?