[Math] Interquartile range in a symmetric distribution

statistics

My text book says:

"If a distribution is symmetric, the median plus or minus the semi-interquartile range contains half the scores in the distribution."

I have a simple example set: {1,4,8,12,15}

It is symmetric:

1 _ _ 4 _ _ _ 8 _ _ _ 12 _ _ 15

Median M = Mean = (1 + 4 + 8 + 12 + 15)/5 = 40/5 = 8

Percentile ranks:
p = (r-1)/(n-1), where r for rank, n for size of set

r (25%) = 1 + .25(n-1) = 2
r (50%) = 1 + .5(n-1) = 3
r (75%) = 1 + .75(n-1) = 4

Percentile values:
v (25%) => 4
v (50%) => 8
v (75%) => 12

Semi-interquartile range:
R/2 = 0.5(v (75%) – v (25%)) = 0.5(12 – 4) = 4

Following the text book, then
M + R/2 = 8 + 4 = 12
M – R/2 = 8 – 4 = 4

Half the values would mean to me:
(M + R/2) – (M – R/2) = 12 – 4 = 8

Total range:
TR = 15 – 1 = 14
then,
1/2TR = 7 which is not equal to 8

I believe I do not understand what the text book says.
May I know the mistake in my understanding?
Perhaps an example could help me understand.

Thank you very much.

Best Answer

You have written

$M + R/2 = 8 + 4 = 12$

$M - R/2 = 8 - 4 = 4$

where $M$ is the median and $R$ the semi-interquartile range

Clearly three of the five terms of $\{1,4,8,12,15\}$ are in the interval $[4,12]$

Three is about as close as you can get to half of five when considering integers, and this is what your textbook is saying

Related Question