Simple question: is there any reason to be concerned about the
statistical significance of cut points in ordered logit? I've included
output below of an ordered logit where the initial cut point is not
significant. (FWIW, under Stata 10 it looks like -svy: ologit- produces
p-values but regular -ologit- produces confidence intervals without
p-values. Not sure why there's this inconsistency.) The null hypothesis
must be that the cut point = 0, and there's no special reason why one of
the cut points can't be 0, so this is not a problem, right? Thanks.
Michael
. xi: svy: ologit q3future clin_35-clin_xx sp_imed-sp_mult
(running ologit on estimation sample)
Survey: Ordered logistic regression
Number of strata = 4 Number of obs
= 599
Number of PSUs = 599 Population size =
116723
N. of poststrata = 55 Design df
= 595
F( 7, 589)
= 3.11
Prob > F =
0.0031
------------------------------------------------------------------------------
| Linearized
q3future | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+----------------------------------------------------------------
clin_35 | 1.249916 .5458784 2.29 0.022 .1778334
2.321999
clin_6p | 1.736979 1.078776 1.61 0.108 -.3816925
3.85565
clin_xx | -1.362227 .8720265 -1.56 0.119 -3.074851
.3503973
sp_imed | -1.118896 .7232126 -1.55 0.122 -2.539256
.3014635
sp_obgy | -.8694114 .6167969 -1.41 0.159 -2.080775
.3419524
sp_peds | .4568723 .9388721 0.49 0.627 -1.387034
2.300779
sp_mult | .9617215 .7174359 1.34 0.181 -.4472931
2.370736
-------------+----------------------------------------------------------------
/cut1 | -.35083 .5493174 -0.64 0.523 -1.429667
.728007
/cut2 | 1.62023 .6364649 2.55 0.011 .3702395
2.870221
------------------------------------------------------------------------------
--
Michael I. Lichter, Ph.D. <[email protected]>
Research Assistant Professor & NRSA Fellow
UB Department of Family Medicine / Primary Care Research Institute
UB Clinical Center, 462 Grider Street, Buffalo, NY 14215
Office: CC 126 / Phone: 716-898-4751 / FAX: 716-898-3536
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/