Re: internet speed data

2001-01-10 Thread David Kirschtel 



dennis roberts wrote:
> 
> this morning ... silly as it sounds ... i went to the following site 
> http://msn.zdnet.com/partners/msn/bandwidth/speedtest500.htm
> which tests your internet throughput speed (nice site!)
> and with both ie and netscape open ...
> 

While not questioning your analysis, I think that the validity of the
underlying data needs to be questioned. Did you happen to notice the
letters "MSN" occuring in the site address.  


  David Kirschtel, Ph.D. * [EMAIL PROTECTED] * 517.432.0898
112 N Kedzie Lab * Mich State Univ * E Lansing, MI * 48824


=
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
  http://jse.stat.ncsu.edu/
=

internet speed data

2001-01-10 Thread dennis roberts 

this morning ... silly as it sounds ... i went to the following site

http://msn.zdnet.com/partners/msn/bandwidth/speedtest500.htm

which tests your internet throughput speed (nice site!)

and with both ie and netscape open ...

toggled back and forth between the two ... took 10 readings each (by 
hitting refresh or reload) ... and here are the data

data are in k bytes/sec ... a reading of 150 is 1.5 Mbps ... or a T1 line speed

  Row ie  netscape

1268   246
2268   233
3260   246
4252   228
5267   233
6268   245
7252   246
8260   232
9253   239
   10260   246

MTB > desc c1 c2

Descriptive Statistics: ie, netscape


Variable N   Mean Median TrMean  StDevSE Mean
ie  10 260.80 260.00 261.00   6.76   2.14
netscape10 239.40 242.00 240.00   7.24   2.29

Variable   MinimumMaximum Q1 Q3
ie  252.00 268.00 252.75 268.00
netscape228.00 246.00 232.75 246.00

MTB > twos c1 c2

Two-Sample T-Test and CI: ie, netscape


Two-sample T for ie vs netscape

N  Mean StDev   SE Mean
ie10260.80  6.76   2.1
netscape  10239.40  7.24   2.3

Difference = mu ie - mu netscape
Estimate for difference:  21.40
95% CI for difference: (14.79, 28.01)
T-Test of difference = 0 (vs not =): T-Value = 6.83  P-Value = 0.000  DF = 17

MTB >

  Row C5C6

1268   ie
2268   ie
3268   ie
4267   ie
5260   ie
6260   ie
7260   ie
8253   ie
9252   ie
   10252   ie
   11246   net
   12246   net
   13246   net
   14246   net
   15245   net
   16239   net
   17233   net
   18233   net
   19232   net
   20228   net

MTB >

notice there is no overlap between the two sets of speeds ... for my sample 
... ie was always faster ... than the fastest netscape

i am sure there are many reasons for this ... but, i just pass this along 
as a simple "experiment" where two treatments are imposed ...

i would be interested in whether some of you would also do this ... and 
report your results



=
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
  http://jse.stat.ncsu.edu/
=