Re: st: How to model data with cross-sectional dependent variable but panel explanatory variables

[Austin Nichols <austinnichols@gmail.com>]

Chao Wang <chao.wang@obs-gyn.ox.ac.uk>:
Number of decelerations over some time interval (e.g. last 10 days) is
a cross-sectional summary of a longitudinal series, so you have a
cross-sectional design where you can use many lags and differences of
X as predictors of current Y.  A "wide" dataset in the sense of
-reshape- might aid you in identifying clusters of predictors but a
"long" dataset will be easier for most computations, and for graphing
trajectories of X.

On Thu, Oct 11, 2012 at 11:11 AM, Chao Wang <chao.wang@obs-gyn.ox.ac.uk> wrote:
> Hi all
>
> I am wondering if anyone can advise which model to use for modelling data with cross-sectional dependent variable but longitudinal/panel explanatory variables? I am doing study on fetal heart rate analysis, and I would like to model the effect of patients' heart rate patterns during labour such as the number of decelerations (measured at 15min interval, so longitudinal/panel) on fetal health (measured by the arterial pH value at birth, so cross-sectional). Traditional panel data models seem require the dependent and explanatory variables to be in the same form (i.e. both panel). However in my study the dependent variable (i.e. the labour outcome for patient i) is only measured once.
>
> The ultimate goal is to predict (and prevent) the adverse labour outcome by monitoring the fetal heart rate patterns (in this sense it also seems to be a time-series analysis for a specific patient/operation, but again, the outcome is static so not consistent with the format of the predictors which are time-series).

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/