As far as I am aware, there are three options here:
Option 1: Skip the question and score no marks.
Option 2: Learn how to get the calculator to give you the equation.
Option 3: Learn how to do the problem manually.
Here is option 3:
The slope of the required line is
and the
-intercept is
and thus, the least-squares regression line of best fit is
and you can see that this matches the data. It has a negative slope. No exercise (
) corresponds to our
-intercept at about 22.2 hours of TV, and no TV (
) goes with our
-intercept at about 10.4 hours of exercise. By eye, this matches our plot.
Now, we can also be asked to find the correlation coefficient,
, which is a measure of how well the line fits the data. It must satisfy
A good fit should have an
near 1 for a positive correlation and near -1 for a negative correlation. A poor fit will have an
near 0. It is usually reported as
. In this case, we have a strong negative correlation as
and
The formula is
If you add a
column to the table, you should find that
, and hence can calculate that:
Of course, the calculator could have spat out this result too, by the option 2 method.
Before anyone comments, it is true that there are different calculating formulae for
,
, and
. They are all equivalent and will give identical results to mine above, assuming that I have made no mistakes. (ADDED BY EDIT: And
@jimmysmith560 has posted one of them while I was typing all this up!
... Question to self... why did I type this up? lol END OF EDIT)
TAKE HOME MESSGAE: (And I am paraphrasing from the Marker's notes on quite a few 2020 trials)... It is important that ALL STUDENTS learn to use the Stats mode on their calculators for regression and correlation questions.
MY ADDENDUM: unless you'd rather learn the method to solve the problem manually.