Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

Re: st: seqlogit with repeated observations

From (Brendan Halpin)
Subject   Re: st: seqlogit with repeated observations
Date   Wed, 20 Jun 2012 21:29:30 +0100

On Wed, Jun 20 2012, wrote:

> I didn't quite understand when you said 'implicit dependent variable'.

I just meant that you gave a complex description of the situation, from
which you were implicitly constructing an ordinal variable. 

> The example you gave is the exact problem I have with my dataset, albeit
> with the repeated case in the second step. It is a binomial outcome in
> all stages, except in the second stage 1s can be repeated.
> I hope I am making sense. I am a little confused as far as the
> significance of the second stage is concerned and its implementation in
> the sequential setup when the decision to pay dividend multiple times is
> expected to have a more significant effect on the response variables
> than a firm that pays dividend only once.

I read your description as saying that for a given firm that was
eligible to pay dividends, you observed more than one opportunity where
they could pay or not pay. If that is the case, a clustered model would
make sense at that point. If it is a case of observing the number of
times they paid, but not the number of times they chose not to pay, then
it wouldn't be appropriate (and it might make sense to collapse it into
paid/didn't paid).

It's actually your final state that bothers me, in that a subsequent
cancellation of the dividend policy doesn't necessarily depend on having
paid, simply on having had one. However, that's a substantive question
that you can judge better than I. 

Brendan Halpin,   Department of Sociology,   University of Limerick,   Ireland
Tel: w +353-61-213147  f +353-61-202569  h +353-61-338562;  Room F1-009 x 3147    ULSociology on Facebook:         twitter:@ULSociology
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index