Ask Your Question

How to calculate the 60% of the median income in Calc?

asked 2018-12-30 12:29:46 +0200

nikos_geo gravatar image

Hello, I have 40 countries and their Gross National Income. I would like to calculate what is the 60% of their median income (poverty line) using LibreOffice Calc.

Thank you in advance.

edit retag flag offensive close merge delete



You need distributions to calculate medians. The Calc function MEDIAN works on a sample which must be representative if you want reliable results.

Lupp gravatar imageLupp ( 2018-12-30 14:07:27 +0200 )edit

As an aside, is it even possible to calculate median income, which refers to income of the citizens, from the Gross National Income which by definition includes many "incomes" which are never seen by the citizens? I doubt that dividing the GNI by the population of each country, for example, will produce a valid "income" for the population. As indicated by @Lupp, you need figures for the distribution of incomes across the population of each country in order to calculate the median of the distribution for each country.

ve3oat gravatar imageve3oat ( 2018-12-31 03:31:39 +0200 )edit

2 Answers

Sort by » oldest newest most voted

answered 2018-12-31 13:17:45 +0200

Lupp gravatar image

updated 2018-12-31 18:36:49 +0200

Just to exemplify my comments I attach this roughly sketched example.

===EDit1 2018-12-31 18:25===
Only attaching the reworked example as annonced in my comment below.
Please note: Nobody can calculate a real distribution from total values. It must be "measured".
Well, there are cases in science where the type of the distribution is knows as a "law", and where then the complete distribution can be generated from a few parameters. That should not be expected in cases as discussed here.

edit flag offensive delete link more


Thank you for your help. But I have some questions about your table. The Column B = Income, the Columns F,G = Median. About the Columns C,D I didn't understand how you computed them, because you made different computations in each cell. I have found how to construct a frequency table and how to choose the intervals (equal intervals) and the number of classes that I want.

nikos_geo gravatar imagenikos_geo ( 2018-12-31 17:55:17 +0200 )edit

Concerning column C: I simply generated arbitrary frequencies by ad-hoc formulae. Then I missed to replac the formulae by the results. This was misleading, of course. Sorry!
Column D simply accumualtes the distribution.
Since I don't know a standard function calculating a median from the distribution, I had to find the critical classes next to the 50%-level using MATCH with the type parameter set to 1: "ascending order asserted". The critical classes are emphasised with the help of CF.
I now will attach a rectified example to my answer.

Lupp gravatar imageLupp ( 2018-12-31 18:24:20 +0200 )edit

Because of your help I have managed to solve my problem. I just needed to compute the cumulative relative frequency of my data. Thanks again!

nikos_geo gravatar imagenikos_geo ( 2019-01-03 11:20:20 +0200 )edit

answered 2018-12-31 12:13:35 +0200

nikos_geo gravatar image

I have the income for a given year for each region (NUTS II) of a specific country. If I got it right, I need to create a frequency distribution of the income? If so, that means that I have to categorize the income into, lets say, 5 categories (0-2000, 2001-4000,... etc). Did I understand it correctly?

edit flag offensive delete link more


To get a useful median you surely should have smaller steps / much more classes.
In addition I would like to know how exactly the frequencies are counted.
May there be cases where people having an income at all are responsible for a group (family) of n (n unknown) persons of which (n-1) persons have 0 income, but are not counted for a respective class?
Sorry. I don't know much about the topic. (Never trusted in related statistics.)
My knowledge about statistical measures is "abstract".

Lupp gravatar imageLupp ( 2018-12-31 13:06:42 +0200 )edit

By the way: Additional information or consecutive questions should be appended to the original question by editing it. The "Answer" feature should only be used for answers.

Lupp gravatar imageLupp ( 2018-12-31 13:10:28 +0200 )edit

Unfortunately I do not know how many members are in a family of every region, I just know the pop of every region [I have 74 regions(cities lets say)]. About the frequencies now, I am searching how to construct them.

nikos_geo gravatar imagenikos_geo ( 2018-12-31 17:02:46 +0200 )edit
Login/Signup to Answer

Question Tools

1 follower


Asked: 2018-12-30 12:29:46 +0200

Seen: 59 times

Last updated: Dec 31 '18