mscm: Mothers' Stress and Children's Morbidity (MSCM) study:...

In gmonette/spida15: Collection of miscellaneous functions for mixed models etc. prepared for SPIDA 2009+ (development version)

Description

The Mothers' Stress and Children's Morbidity (MSCM) study is a longitudinal observational study of the causal effect of maternal stress on childhood illness (Zeger and Liang 1986, pp. 125-128).

Usage

data(mscm)

Format

A data frame with 5010 observations on 167 mother/child dyads each observed on 30 successive days with the following 14 variables.

id

id for dyad

day

day of observation: 1 to 30 for each dyad

stress

indicator for maternal stress

illness

indicator for child illness

married

a factor with levels married other

education

a factor with levels less than high school some high school high school graduate some college college graduate

employed

indicator for employment

chlth

child health status at baseline: 1=very poor 2=poor 3=fair 4=good 5=very good

mhlth

mother health status at baseline: 1=very poor 2=poor 3=fair 4=good 5=very good

race

a factor with levels white non-white

csex

a factor with levels male female

housize

a factor with levels 2-3 people more than 3 people

bIllness

proportion of child days ill in first 7 days

bStress

proportion of maternal days with Stress in first 7 days

Details

The Mothers' Stress and Children's Morbidity (MSCM) study is a longitudinal observational study of the causal effect of maternal stress on childhood illness (Zeger and Liang 1986, pp. 125-128). In the MSCM data, the daily prevalence of childhood illness was 14 questions. How would the prevalence change if an ongoing, fully effective stress-reduction intervention program were instituted? How would prevalence change if conditions worsened and all mothers were subjected to substantial stress on a daily basis? To attempt to answer these questions, we use a formal model for causal effects in longitudinal studies introduced by Robins (1986, 1987a,b). This model extends Neyman's (1923) counterfactual causal model for "point" treatment studies to longitudinal studies with time-varying treatments and confounders. We show that the methods for causal inference developed by Robins provide a better justified basis for answering the foregoing causal questions in longitudinal data in general and in the MSCM study in particular than do methods based on generalized estimating equations (GEE's).

Source

Estimation of the Causal Effect of a Time-Varying Exposure on the Marginal Mean of a Repeated Binary Outcome Journal article by James M. Robins, Sander Greenland, Fu-Chang Hu; Journal of the American Statistical Association, Vol. 94, 1999

Examples

data(mscm)
## maybe str(mscm) ; plot(mscm) ...

gmonette/spida15 documentation built on May 17, 2019, 7:26 a.m.