The mean point
Every regression line passes through (x̄, ȳ)
- In 6 of 20 recent exams
- Paper 1 or 2
- 1 to 3 marks
How it is set in the exam
Paper 2 (calculator) mostly (8 of 9 logged parts), 1–3 marks a part; once on Paper 1 (no calculator) with both lines given. Common (6 of 20 exam versions). It is often a small part inside a longer regression question, and twice it led into a missing data value.
Exam phrases to spot (5)
“find the mean of x and the mean of y”
“show that the mean of y is …”
“write down the mean point”
“draw the regression line on the scatter diagram”
“the two regression lines intersect at …”
In disguise (3)
- "the lines L1 and L2 pass through the same point (p,q); find p and q": that point is (xˉ,yˉ)
- a line of best fit "by eye": it must go through (xˉ,yˉ)
- "a line through the origin and the mean point": its gradient is xˉyˉ
Variants (2)
- given: both lines; one line and one mean (the mean itself found from the table first); the data
- asked: both means, one mean as a "show that", the line drawn on a scatter diagram, a line through the origin and the mean point
Looks like it, but isn't (2)
- both lines are given and an estimate is asked for: Predict x from y
- the question asks for a value missing from the table: Work backwards from the mean or median. With the regression line given, this type is only its first step, the mean from the line (Example 3.2)
Two lines given: where they cross
For a group of trees, x is the height in metres and y is the trunk diameter in centimetres. The regression line of y on x is y=0.5x+20, and the regression line of x on y is x=1.2y−16.
(a) Find the mean height and the mean trunk diameter of the trees. [3]
(b) The heights of the trees add up to 1000 m. Find the number of trees in the group. [2]
Try it yourself first, then compare.
Now try a question
Questions of this type are on their way. Try another one meanwhile: write your working on paper or on screen and see where each mark went. Free, with no account.
| Given | First move |
|---|---|
| both regression lines | solve them simultaneously: they cross at (xˉ,yˉ) |
| one line and xˉ (or yˉ) | substitute the known mean into the line and solve for the other mean (exact, because (xˉ,yˉ) lies on every regression line: not the rearrangement that Predict x from y forbids) |
| the raw data | the means from the GDC (one-variable statistics on each list, or the regression output) |
- 1
Say it:
every regression line passes through the mean point (xˉ,yˉ). Writing this (or sketching the two lines crossing) is the recognition mark.
- 2
Both lines:
substitute one into the other and solve (by hand on ; simultaneous linear equations on ). Label the answers xˉ=… and yˉ=…, in context.
- 3
One line and one mean:
substitute the mean you know. If it isn't given, find it from the complete list first.
- 4
A missing value:
total =n× mean, then subtract the known values (Work backwards). Never put the missing point's x into the line: the line need not pass through any data point, only through (xˉ,yˉ).
- 5
Draw the line:
plot (xˉ,yˉ) and a second point of the line far from it (the y-intercept, or the value at the far end of the axis), and rule a straight line through both across the data.
- 6
A line through the origin and the mean point:
its gradient is xˉyˉ (rise over run from (0,0)), from the unrounded means. Write y=…x, with no constant.
Substituting the missing point's x into the line
to find a missing y: the line gives an estimate, not the recorded value. It scores nothing (Example 3.2 would give 10.4 instead of 9.7).
Taking yˉ from the incomplete table
(dividing the known values by 7, or by 6).
Rounding a mean before using it
carry xˉ and yˉ unrounded into the next part.
More mistakes (4)
Unlabelled means
in (a) of a two-lines question: say which is xˉ and which is yˉ, in context.
Drawing freehand,
through the wrong point, or with two points too close together to fix the line's direction.
A line through the origin with the gradient upside down
it is xˉyˉ, rise over run.
Reading x=cy+d as y=cx+d
when solving the two lines (or typing it into a graph screen as y=cx+d): keep each line's subject and substitute one into the other. In Example 3.1 the misread lines give 0.5x+20=1.2x−16, x=51.4 instead of 20.