Simple statistical analysis

User avatar
StuartR
Administrator
Posts: 12612
Joined: 16 Jan 2010, 15:49
Location: London, Europe

Simple statistical analysis

Post by StuartR »

It is a long time since I had to do any statistical analysis.

I have a set of "things", which are named "Thing 1", "Thing 2" etc. up to "Thing 20"
I have a number of events associated with each of these things.
I need to demonstrate whether the distribution of events shows a significant difference between the things, or whether it is just as likely to be random.
Can someone please tell me what to do to generate a significance rating

The numbers I have are
ThingEvent Count
15
28
31
41
50
60
76
80
90
103
110
121
130
140
151
160
170
181
190
202
StuartR


User avatar
Jan Karel Pieterse
Microsoft MVP
Posts: 656
Joined: 24 Jan 2010, 17:51
Status: Microsoft MVP
Location: Weert, The Netherlands

Re: Simple statistical analysis

Post by Jan Karel Pieterse »

Are there no replications in there, so we can assess the variability?
Regards,

Jan Karel Pieterse
Excel MVP jkp-ads.com

User avatar
StuartR
Administrator
Posts: 12612
Joined: 16 Jan 2010, 15:49
Location: London, Europe

Re: Simple statistical analysis

Post by StuartR »

I guess what I would like to be able to state is whether the high values for things 7 and 2 are significant, or are they likely to be chance.

In other words is the variable number of events per thing significant, or would you expect this number of events to have such an uneven distribution.
StuartR


User avatar
Jan Karel Pieterse
Microsoft MVP
Posts: 656
Joined: 24 Jan 2010, 17:51
Status: Microsoft MVP
Location: Weert, The Netherlands

Re: Simple statistical analysis

Post by Jan Karel Pieterse »

I don't have a boilerplate answer for you, this is beyond my expertise.
Maybe this page can give you some clues:
http://www.statsoft.com/textbook/basic- ... %20samples
Regards,

Jan Karel Pieterse
Excel MVP jkp-ads.com

User avatar
Jan Karel Pieterse
Microsoft MVP
Posts: 656
Joined: 24 Jan 2010, 17:51
Status: Microsoft MVP
Location: Weert, The Netherlands

Re: Simple statistical analysis

Post by Jan Karel Pieterse »

Regards,

Jan Karel Pieterse
Excel MVP jkp-ads.com

User avatar
HansV
Administrator
Posts: 78532
Joined: 16 Jan 2010, 00:14
Status: Microsoft MVP
Location: Wageningen, The Netherlands

Re: Simple statistical analysis

Post by HansV »

To test whether data are uniformly distributed, you can use the chi-square test. As the attached workbook shows, the chi-square value for this sample is ~69.6, with 19 degrees of freedom (20-1).
The p-value is =CHIDIST(69.6, 19) ~ 10^-7, i.e. it is extremely improbable that these data are uniformly distributed.
ChiSquareTest.xls
You do not have the required permissions to view the files attached to this post.
Best wishes,
Hans

User avatar
StuartR
Administrator
Posts: 12612
Joined: 16 Jan 2010, 15:49
Location: London, Europe

Re: Simple statistical analysis

Post by StuartR »

Hans,

That is exactly what I was looking for, thank you.

I remember doing chi-square calculations when I was studying zoology at school, many decades ago, but I couldn't remember any detail about it, even the name.

Jan Karel
Thanks for the pointers. There's lots of good stuff to read there.
StuartR


User avatar
Jan Karel Pieterse
Microsoft MVP
Posts: 656
Joined: 24 Jan 2010, 17:51
Status: Microsoft MVP
Location: Weert, The Netherlands

Re: Simple statistical analysis

Post by Jan Karel Pieterse »

Yes, an excellent resource if you ask me. They used to have a downloadable version of the textbook too, but I can't find it anymore. I do have it on my system though.
Regards,

Jan Karel Pieterse
Excel MVP jkp-ads.com