[Statistica / Statgraphics] - Can grouped data be analyzed?

[wtw1936]

I have many datasets consisting of grouped data -- no individual cases, only frequencies (# of cases) in each class (some open-ended upper classes).

Can such data even be entered in Statistica or Statgraphics? Can it be analyzed? I can find no reference to grouped data in the help files of either program. Perhaps there's some indirect, less obvious way to handle this.

I could of course use the conventional textbook methods to do basic calculations by hand, but I've lots (!) of data and I'd like to be able to take advantage of the more sophisticated distribution fitting procedures.

While I'd like some basic descriptive stats, I'm particularly interested in fitting probability distributions.

Thanks for any suggestions.

 

[gianmarco]

Re: Can grouped data be analyzed in Statistica or Statgraphics?

Hi,
I happen to use statistica. Do you mind to post your data or part of them so that I can give a closer look at them and get an idea of what you can do with them?

Regards
gm

 

[wtw1936]

Re: Can grouped data be analyzed in Statistica or Statgraphics?

Here's a particularly messy example of the type of grouped data I'm dealing with. There are unequal class widths, gaps between some classes, and an open ended upper class. I need mean, standard deviation, percentiles, etc. The context suggests an exponential distribution, but I'd like to know that no others fit significantly better.

For a sample of 900 cases,
Class freq
0--100 704
100--200 38
200--300 20
300--400 9
500--1000 27
1000--2500 20
2500--5000 12
5000--10000 6
10000--inf 1

Not all of my datasets are this messy: many have the more usual equal continuous intervals and no open-ended classes.

 

[gianmarco]

Re: Can grouped data be analyzed in Statistica or Statgraphics?

Hi!

Sorry for my late reply, but things are busy here

I gave a look to your example, and to be honest with you I do not know what you can do with that kind of dataset.


May be that someone else here in the Forum could give you some hint.


Best Regards
Gm

 

[JenB]

Re: Can grouped data be analyzed in Statistica or Statgraphics?

These are good questions. The summary of your data really limits what you can do with it.

Now, if you can generate some data, then you might consider STATISTICA Distributions and Simulations, which is useful for comparing fits of distributions on data. Here's a blog about this module.

But there is a big problem of getting that frequency data into a more usable form.

Best of luck to you.