RobertNiles.com
About Robert
Reporting Help
Finding Data on the Internet
Questions for Candidates
Stats Lessons
Mean
Median
Percent
Per capita
Standard Deviation
Margin of Error
Data Analysis
Sample Sizes
Stats Tests
Help Board
Bookstore


The Cartoon Guide to Statistics

This isn't some dumb-downed whitewash. It's a must-read for a beginner taking his or her first steps toward understanding stats.
More information
-->

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Calculating Percentiles
Message posted by Tiffani (via 209.77.70.17) on March 9, 2001 at 7:39 PM (ET)

I am trying to calculate the 60th percentile, but do not have the original data set. The only data that I have to work with are n=798, mean=3883, std dev=1365, 25th%tile=2841, median=3684, 75th%tile=4663 and 90th%tile=5824. Is it possible to calculate or estimate an accurate 60th%tile with this information? How?


READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Calculating Percentiles
Message posted by Phil (via 216.175.112.150) on March 10, 2001 at 11:29 AM (ET)

You will be estimating, so there are probably several ways. It is not a normal distribution since it is skewed, so the z-table won't help.

You do know four points on the Cumulative Distribution Function. You could plot it and see how it looks. You could estimate the 60th percentile from that.


Re: Calculating Percentiles
Message posted by Darius (via 200.23.217.10) on March 12, 2001 at 7:19 PM (ET)

Interpolate between Median 50 (percentil) and 75 percentil.

Y = 1726 + 39.16* Percentil

60 percentil = 4075.6


Re: Calculating Percentiles
Message posted by Beth (via 38.178.11.17) on July 25, 2001 at 4:40 PM (ET)

Tiffani's question is exactly the question we have. The only number I can't get is the 1726 in Darius' formula. How is 1726 calculated.

Thanks for any assistance.


Re: Calculating Percentiles
Message posted by Darius (via 200.23.217.10) on July 27, 2001 at 5:22 PM (ET)

25=2841,
50=3684,
75=4663
90=5824
60=??

Interpolate between Median 50 (percentil) and 75 percentil.
Y = 1726 + 39.16* Percentil

60 percentil = 4075.6
--------------------------------------------------
The only number I can't get is the 1726 in Darius' formula. How is 1726 calculated.

*************************
Sorry, let me explain, I used only X1=50, Y1=3684, X2=75, Y2=4663

Y=A*X+B
A=(Y2-Y1)/(X2-X1)
B=Y1-A*X1

I used only the closest numbers in the interpolation to 60 (50 and 75), you can use all and the equation will be (using regression)
Y= 1598.551 + 44.24*X
Y(60)= 4253
But I tink the first approach is better (because you don愒 care how the distribution was before 50 and after 75 and this is offen suggested for interpolation in statistical tables).


Re: Calculating Percentiles
Message posted by Darius (via 200.23.217.10) on July 27, 2001 at 5:57 PM (ET)

Also, if interested in a more exact value, with all the values (using polinomial regresion), but more complicated ones:

Y=912.08+115.17*X-1.846*X^2+0.013*X^3

Y(60)=3985


Re: Calculating Percentiles
Message posted by desirie gamil (via 208.142.142.21) on July 31, 2001 at 10:50 PM (ET)

I want to know the concept
of problem.


What's Percenting definition ?
Message posted by Patrick (via 193.250.38.53) on November 12, 2001 at 6:46 PM (ET)

I'm french. Please, what's the definition exact of percenting ? With an french exemple possible ? Thanks to answer by Email directly.
Patrick


Re: Calculating Percentiles
Message posted by Tomi (via 154.32.143.5) on November 13, 2001 at 9:34 PM (ET)

Bonjour, Patrick. Je parle Francais un peu seulement, mais j'essayera ecrire en Francais. Votre "email address" n'est pas visible.

Conaissez-vous la distribution de fr廦uence cumulative? Les "percentiles" sont similaire. Si on a des valeurs, un quartier des valeurs sont inferieur au "25 percentile".

Par exemple:

Data: 3, 6, 7, 9, 10, 11, 13, 16, 17, 20

Il y a 10 valeurs. Le valeur "50 percentile" est 10.5 parce que 50% des valeurs sont inferieur a 10.5. Le valeur "25 percentile" est 7 parce que 25% des valeurs sont inferieur a 7. Le valeur "75 percentile" est 16 parce que 75% des valeurs sont inferieur a 16. Naturellement avec 10 valeurs les percentiles sont approximations.

Les percentiles sont utiles pour comparaison un valeur avec les valeurs normales. Par exemple, ma fille etait 3300g apres 2 semaines. Etait-elles trop petite? Elle etait inferieur au valeur 10 percentile mais superieur au valeur 5 percentile. Non, elle est petite mais n'etait pas trop petite.



Your $5 contribution helps cover part the $500 annual cost of keeping this site online.

Niles Online Sites:RobertNiles.comTheme Park InsiderViolinist.com

RobertNiles.com™, the site, content and services 咀opyright 1996-2002, Robert Niles.
All rights reserved. Questions? Comments? Read my Privacy Policy, or E-mail me!