Statistics: IQR

[Function]

Hello everyone

 

In order to write a paper on the differences in epidemiology of pandemic influenza A(H1N1)pdm09 during that pandemic, between Europe and Africa, I need to understand a concept in a German article I'm using:

 

"Der Altersmedian der Einzelfälle liegt bei 16 Jahren (IQR 10;28)."

 

Which means as much as:

 

"The age median of the individual cases is 16 years old (IQR 10;28)."

 

Now, I know that IQR means inter-quartil-range, but can someone explain the meaning of this range 10;28 in this specific context?

Would it just mean that Q1 = 10 and Q3 = 28?

 

Thanks.

 

F.

Edited February 28, 2015 by Function

[mathematic]

My guess is the same as yours. Range is 10 to 28.

[studiot]

Hello, function, please recheck your original.

 

I don't see how the median can be 16, I think it should be 19.

Edited February 28, 2015 by studiot

[Function]

Hello, function, please recheck your original.

 

I don't see how the median can be 16, I think it should be 19.

IQR1.jpg

 

Note that it's the median age of the cases! It may very well be possible that some ages have had no (registrated) cases

At least, that's what I think. I've checked the article once more and it does state 16 as median age.

[studiot]

In which case there is more than you have stated since the age must be plotted as a cumulative frequency against some flu event. This techniqe is useful as it allows an open ended age range to be divided into quartiles.

 

The median is the centre of the the central 50% (ie the inner two quartiles) of the frequencies of catching or recovering from or transmitting or whatever the flu.

 

Edit You have a PM

Edited March 1, 2015 by studiot

[CharonY]

The interpretation in OP seems correct. What they will have analyzed is the age of each case. There is no reason why the patients should have a perfect continuous age distribution.

[studiot]

Good morning, function.

 

Thinking about this further it may be more important to you to consider how you are going to use the data from the german paper.

After all you are not going to repeat their survey or statistical analysis, just use the results.

 

I have sent you some stuff on box plots which may be of interest.

 

It is likely that the figures quoted (percentiles and median) were derived from a frequency plot (cumulative or otherwise).

That could indeed lead to the median not coinciding with the centre of the age range.

It would, however, be illogical to miss out ages because there is no data for that particular age so the axe axis must be complete.