[BiO BB] Mean and Standard deviation
wusu at ele.uri.edu
wusu at ele.uri.edu
Tue Feb 21 18:49:17 EST 2006
The mathematical interpretations for the reduction of SD due to a trimmed sample
size at both ends are all very good. However, I don't know why you would get
rid of the end points of your data set, because the frequencies for both end
classes or points are rather high (that is 85 + 33 = 118) compared with the
other classes. You are ignoring 118 / 394 = 29.95% of your data set. If it is
the boundary effect that you are concerned with, you may want to redesign the
experiment or simulation in my opinion.
Regards,
Su
Quoting Martin Gollery <marty.gollery at gmail.com>:
> Chris is correct. Another way to look at it is to consider that you are
> removing the values that have the greatest deviation from the mean, and so
> it makes sense that you then have a lower standard deviation.
>
> Marty
>
> On 2/21/06, Frenz, Christopher <CHRISTOPHER_FRENZ at nymc.edu> wrote:
> >
> > SD calculations involve the sample size as one of the variables. This is
> > likely what is causing the change in SD, since you are reducing the the
> > sample size when you remove the two sets of endpoints.
> >
> > Chris
> >
> >
> > -----Original Message-----
> > From: bio_bulletin_board-bounces+christopher_frenz=
> > nymc.edu at bioinformatics.org on behalf of Subhash Agarwal
> > Sent: Tue 2/21/2006 12:03 AM
> > To: bioinformatics.org
> > Subject: [BiO BB] Mean and Standard deviation
> >
> > Dear Forum Memebers
> >
> > I have a dataset with 394 data points. The frequecy distribution of
> > which is as follows:
> >
> > 0-0.02 85 0.02-0.04 63 0.04-0.06 51 0.06-0.08
> > 39 0.08-0.1 36 0.10-0.12 24 -0.12-0.14 21 0.14-0.16
> > 19 0.16-0.18 13 0.18-0.2 10 >0.20 33
> >
> > The mean and SD of the data is 0.0842 and 0.0853. But when the data
> > points from 0-0.02 and > 0.2 are removed and the mean and SD is calculated
> > it is found that mean doesnt change much (0.0819) but SD does (0.0472). My
> > query is what can I interpret from this regarding the data.
> >
> > If any of the forum members feels that this question should be addressed
> > to some other place do let me know.
> >
> > Thanks
> > Subhash
> >
> >
> > ---------------------------------
> > Jiyo cricket on Yahoo! India cricket
> > Yahoo! Messenger Mobile Stay in touch with your buddies all the time.
> >
> >
> > _______________________________________________
> > Bioinformatics.Org general forum - BiO_Bulletin_Board at bioinformatics.org
> > https://bioinformatics.org/mailman/listinfo/bio_bulletin_board
> >
> >
> >
>
>
> --
> --
> Martin Gollery
> Associate Director
> Center For Bioinformatics
> University of Nevada at Reno
> Dept. of Biochemistry / MS330
> 775-784-7042
> -----------
>
More information about the BBB
mailing list