Sec 2.

8 – Measures of Position
If it is an easy test then lots of students will
be getting high marks. Maybe 65% isn’t
high when compared to other marks.
If it was an
easy test …?
Suppose Knowing
you get 65% BAD the mark
in a test. isn’t always
Is this a useful.
good mark? If it was an hard
test …?
GOOD If it is a hard test then lots of students
will be getting low marks. Maybe 65% is
high when compared to other marks.
 BUT
Knowing how
many marks were “ 𝑃𝑖 ”
smaller might be
useful. The 𝑖 𝑡ℎ percentile

“Rank”

What percent
𝑖%
of the marks of the other values
were smaller are smaller than, or
than yours? equal to, that value
 Examples

𝑃50 𝑃80
50% of the values 80% of the values
smaller than, or equal 𝑃75 smaller than, or equal
to, it. 75% of the values to, it
smaller than or
equal to it.
Quartiles split
the data into 4
median parts
8th decile
3rd quartile Deciles split the data
into ten parts.
 Quartiles for Raw Data

Splits the data

into quarters.

Example:
40 data
points.
How many in
each quarter? 25% 25% 25% 25%
10
Example 3, 4, 6, 7, 8, 10, 11, 13, 14 (n=9 odd)

Find median = 𝑄2

3, 4, 6, 7, 8, 10, 11, 13, 14

Find median of “set 1” Find median of “set 2”

4+6 11 + 13
𝑄1 = =5 𝑄3 = = 12
2 2

So: 𝑄1 = 5 , 𝑄2 = 𝑚𝑒 = 8 , 𝑄3 = 12
Example 4, 7, 13, 13, 18, 20, 20, 32 (n=8 even)

13 + 18
Find median = 𝑄2 = 15.5
2
4, 7, 13, 13, 18, 20, 20, 32

Find median of “set 1” Find median of “set 2”

7 + 13 20 + 20
𝑄1 = = 10 𝑄3 = = 20
2 2
So: 𝑄1 = 10 , 𝑄2 = 𝑚𝑒 = 15.5 , 𝑄3 = 20
 Quartiles and Percentiles: Raw Data

= 𝑄𝑈𝐴𝑅𝑇𝐼𝐿𝐸. 𝐸𝑋𝐶 𝑎𝑟𝑟𝑎𝑦 ; 𝑞𝑢𝑎𝑟𝑡

𝑞𝑢𝑎𝑟𝑡 = 1, 2, 3

= 𝑃𝐸𝑅𝐶𝐸𝑁𝑇𝐼𝐿𝐸. 𝐸𝑋𝐶 𝑎𝑟𝑟𝑎𝑦 ; 𝑘

0<𝑘<1
STEP 1: Enter the data.

STEP 2: Enter the Excel

command.

STEP 3: Press Enter

 Quartile Deviation • Data is not
symmetric
• Data is • Has outliers
symmetric
• No outliers
𝑄3 − 𝑄1 Median is better
𝑄=
2
Mean is
good Measure of
variation: Use
Measure of variation: use Quartile
standard dev. (s) deviation
Example 3, 4, 6, 7, 8, 10, 11, 13, 14
𝑄1 = 5 , 𝑄2 = 𝑚𝑒 = 8 , 𝑄3 = 12

12−5 You can expect a typical data point to

𝑄= = 3.5
2 differ from the median by 3.5 units

Example 4, 7, 13, 13, 18, 20, 20, 32

𝑄1 = 10 , 𝑄2 = 𝑚𝑒 = 15.5 , 𝑄3 = 20
20−10 You can expect a typical data point to
𝑄= =5
2 differ from the median by 5 units
 Grouped
Data
Median
Percentiles
𝑀𝑒 Quartiles
𝑃𝑖
𝑄1 𝑄2 𝑄3

Definition: Percentile class

The class that contains the
percentile of interest
 Position of
Class width percentile
Cumulative
frequency of
previous class

𝑛×𝑖
𝑐 − 𝐹𝑙𝑒𝑠𝑠
100
𝑃𝑖 = 𝐿𝑖 +
𝑓𝑖

Lower boundary Frequency of

of percentile class percentile class
Example:
Find:

1st quartile,

3rd quartile,

n = 50
 𝑛×𝑖 50×25
1st quartile Position:
100
=
100
= 12.5 Look for 𝑥12.5
𝑥1 → 𝑥4

𝑥5 → 𝑥14

𝑥15 → 𝑥22
𝑥12.5
in 𝑥23 → 𝑥37
here
𝑥38 → 𝑥46
𝑥47 → 𝑥49
𝑥50
 𝑓𝑖 = 10

𝐹𝑙𝑒𝑠𝑠 = 4
𝐿𝑖 = 41.5

𝑛×𝑖
= 12.5
𝑐 = 57.5 − 53.5 = 4 100
 𝑛×𝑖 𝐿𝑖 = 41.5
𝑐 − 𝐹𝑙𝑒𝑠𝑠
100 𝑐=4
𝑃𝑖 = 𝐿𝑖 +
𝑓𝑖
𝑛×𝑖
= 12.5
100
4 12.5 − 4
= 41.5 + 𝐹𝑙𝑒𝑠𝑠 = 4
10
𝑓𝑖 = 10
= 44.9

𝑸𝟏 = 𝟒𝟒. 𝟗
 𝑛×𝑖 50×75
3rd quartile Position:
100
=
100
= 37.5 Look for 𝑥37.5

𝑥23 → 𝑥37

Is between
these two

Use the upper class 𝑥38 → 𝑥46

 𝑛×𝑖 𝐿𝑖 = 53.5
𝑐 − 𝐹𝑙𝑒𝑠𝑠
100 𝑐=4
𝑃𝑖 = 𝐿𝑖 +
𝑓𝑖
𝑛×𝑖
= 37.5
100
4 37.5 − 37
= 53.5 + 𝐹𝑙𝑒𝑠𝑠 = 37
9
𝑓𝑖 = 9
= 53.7

𝑸𝟑 = 𝟓𝟑. 𝟕
Percentiles from Ogives Example: Ogive showing results for a
test out of 80 marks

How many data

points there are?

Cumulative frequency
40 data points 80 was the
= scores for 40 max mark
students

10 scores in
each quartile
 Half the Only 25% of
class got a the students
score of at had a score
most of more than
44
______ 52 marks
_____

A pass is
50% Students in the
i.e. 40 marks 1st decile had a
maximum mark
14
Only ____ 29
of ______
students
failed