Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

R: Re: st: AW: Sample size

From   Carlo Amenta <>
To   <>
Subject   R: Re: st: AW: Sample size
Date   Sat, 15 May 2010 20:59:09 +0200

54 seems to me an adequate size for inferential statistics. What you 'infere' from that reduced sample in term of results to attribute to the population maybe it's the main problem with your analysis.
Carlo amenta

----- Messaggio originale -----
Da: Fabio Zona <>
Inviato: sabato 15 maggio 2010 16.34
Oggetto: Re: st: AW: Sample size

I have a full model, control and 3 indep Var1 Var2 Var3.
I have two subgroups, identified by a dummy D

When D = 1, the subgroup has only 54 companies (for D=0 I have 139 companies, for a total of 193 companies).

I run interactions:
Var1 x D
Var2 x D
Var3 x D
I introduce these interactions one per time; afterward, I include them all-together (i.e., a chow test).

How can I be sure if my analysis is OK, given that the second subsample has only 54 corporations?

----- Messaggio originale -----
Da: "Martin Weiss" <>
Inviato: Sabato, 15 maggio 2010 16:19:51 GMT +01:00 Amsterdam/Berlino/Berna/Roma/Stoccolma/Vienna
Oggetto: st: AW: Sample size


" How can I figure out the minimum sample size per each subgroup required to run correct regression (OLS)?"

How is a "correct regression" defined theoretically? Stata does have commands like -h sampsi-, but apparently you are looking for something else?


-----Ursprüngliche Nachricht-----
Von: [] Im Auftrag von Fabio Zona
Gesendet: Samstag, 15. Mai 2010 16:12
Betreff: st: Sample size

Dear Statalist

I have a sample of about 193 corporations; this sample is composed by:

- a subgroup of 139 corporations (subgroup A)
- a subgroup of 54 corporations (subgroup B), identified by a dummy SUBG

I run interactions (i.e., the dummy SUBG x indep var); I also run a chow test.

Do you think that 54 companies in the interacting subsample are not enough? How can I figure out the minimum sample size per each subgroup required to run correct regression (OLS)?

Thank you

Il messaggio che segue e' inserito automaticamente dal server di posta
dell'Universita' Bocconi.

Il 5 per mille per gli studenti meritevoli della Bocconi.
E' un atto volontario, non costa nulla e non sostituisce l'8 per mille.
Scegli Bocconi: codice fiscale 8002461015

[Messaggio originale troncato]

*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index