simdata87 {miWQS}R Documentation

Simulated Dataset 87

Description

The 87th dataset from the simulation study with 10 percent of observations were below the detection limit (BDL) based of a real epidemiological dataset. Out of 1000 subjects, fourteen correlated chemicals are completely observed (in X.true). In this simulation design, each chemical was simulated from independent normal distributions.

BDLs were created using the bottom 10th percentile of the true data. Three covariates are considered: the child’s age, the child’s sex (Male/Female), and the child’s ethnicity/race (White, Non-Hispanic White, and Other). After creating a model matrix, male white newborns (age = 0) serves as the reference. The age is simulated from a normal with mean of 3.78 and standard deviation of 1.85 truncated between 0 and 8. The categorical variables are simulated from independent binomial distributions. The outcome will be simulated using a logistic WQS model with complete data:

logit(\mu_{i}) = -1.25 + log(1.75)*WQS_{i} + 0.032*z_{age}+ -0.0285*z_{sex} + 0.540*z_{His} + 0.120*z_{other}

where

WQS_{i}= \Sigma_{j=1}^{c}(w_{j}*q_{ij})

with four of the 14 weights w_j's being 0.25 and the rest 0. The q_ij refers to the quantile score of the jth chemical in the ith subject.

Usage

data(simdata87)

Format

A list that contains:

References

Ward, M. H., Colt, J. S., Metayer, C., Gunier, R. B., Lubin, J., Crouse, V., … Buffler, P. A. (2009). Residential Exposure to Polychlorinated Biphenyls and Organochlorine Pesticides and Risk of Childhood Leukemia. Environmental Health Perspectives, 117(6), 1007–1013. https://doi.org/10.1289/ehp.0900583

Examples

simdata87 <- data(simdata87)

[Package miWQS version 0.4.4 Index]