Page 357 - Statistics for Dummies

Chapter 21: Ten Surefire Exam Score Boosters
been sold. You find the mean price is $219,100 with standard deviation of
$60,100, and you know the mean size is 1,993 square feet, with standard devi-
ation of 349 square feet. You find the correlation between size and price for
these homes is +0.90. Find the best-fitting regression line that you can use to
predict house price using size.”
Your first step is labeling everything. Knowing you use size to predict price,
you figure size must be the x variable and price must be the y variable. You
then label the means
(in thousands)
(square feet) and
(square feet) and
respectively; the standard deviations are labeled
(in thousands), respectively, and the correlation is labeled r = 0.90.
The sample size is n = 100. Now you can plug your numbers into the right for-
mulas. (See Chapter 18 regarding correlation and regression.)
When you know you have to work with a regression line and that formulas
are involved, having all the given information organized and labeled, ready to
go, is very comforting. It’s one less thing to think about. (The problem in this
particular example is solved in the section “Make the Connection and Solve 341
the Problem.”) If that example doesn’t convince you, here are six more rea-
sons to label what you are given in a problem:
✓ Labeling allows you to check your work more easily. When you go
back to check your work (as I advise in the section “Do the Math —
Twice”), you’ll quickly see what you were thinking when you did the
problem the first time.
✓ Your professor will be impressed. He will see your labels and realize you
at least know what the given information stands for. That way if your cal-
culations go haywire, you still have a chance for partial credit.
✓ Labeling saves time. I know that writing down more information seems
like a strange way to save time, but by labeling all the items, you can pull
out the info you need in a flash.
For example, suppose you need to do a 95% confidence interval for the
population mean (using what you know from Chapter 13) and you’re
told that the sample mean is 60, the population standard deviation is
10, and the sample size is 200. You know the formula has to involve , ,
and n, and you see one that does:
Because you’ve already labeled everything, you just grab what you need,
put it into the formula, throw in a z*-value of 1.96 (the critical value
corresponding to a 95% confidence level), and crunch it out to get the
answer:

3/25/11 8:12 PM
30_9780470911082-ch21.indd 341 3/25/11 8:12 PM
30_9780470911082-ch21.indd 341