extremely unbalanced (rare case) sample statistics?

[Jinkuk Hong]

Dear all,
I have a sample of three groups, where the size of each group is vastly different:
Group1: 14 cases (it's a sample with a very rare situation)
group2: 200 cases
group3: 3600 case

I'd like to run simple bivariate analysis (Crosstab and one-way ANOVA) to see
if there are significant differences among 3 groups for a set of outcome
variables. What worries me is the sample size being extremely unbalanced.
I'm wondering if there is certain way to account for this unbalanced sample
size statistically in Crosstab or ANOVA (or in general).
Any suggestion would be greatly appreciated.

Best,
Jin

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Hector Maletta]

Unbalanced it is, but the main problem is the extremely small size of one of the samples. If you could miraculously multiply the three sample sizes x 1000, the unbalance would remain but the problem would disappear (there are corrections for unbalanced ANOVA, but no solution for an exceedingly small sample).

 

Hector

 

De: SPSSX(r) Discussion [mailto:[hidden email]] En nombre de Jinkuk Hong
Enviado el: Thursday, November 11, 2010 12:46
Para: [hidden email]
Asunto: extremely unbalanced (rare case) sample statistics?

 

Dear all,
I have a sample of three groups, where the size of each group is vastly different:
Group1: 14 cases (it's a sample with a very rare situation)
group2: 200 cases
group3: 3600 case

I'd like to run simple bivariate analysis (Crosstab and one-way ANOVA) to see
if there are significant differences among 3 groups for a set of outcome
variables. What worries me is the sample size being extremely unbalanced.
I'm wondering if there is certain way to account for this unbalanced sample
size statistically in Crosstab or ANOVA (or in general).
Any suggestion would be greatly appreciated.

Best,
Jin

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

---

No virus found in this message.
Checked by AVG - www.avg.com
Version: 10.0.1153 / Virus Database: 424/3249 - Release Date: 11/10/10