dat.gen: Generate bivariate meta analysis studies

Description

Generate bivariate meta analysis studies based on random-effects model, some studies with smallest weighted sum of the two outcomes are suppressed.

Usage

dat.gen(
  m.o,
  m.m,
  s.m,
  angle.LC = pi/4,
  mybeta,
  tau.sq,
  rho.w,
  rho.b,
  s.min = 0.01,
  m.m.o = 0,
  s2.dist = 2,
  verbose = F
)

Arguments

m.o: number of observed studies
m.m: number of missing / suppressed studies
s.m: vector of the mean of the variances of the two outcomes
angle.LC: direction of suppressing line, default is pi/4, i.e. the studies on the left bottom corner are missing
mybeta: the true center of the effect sizes
tau.sq: between-study variance, the larger it is the more heterogeneity.
rho.w: within-study correlation of the two outcomes
rho.b: between-study correlation of the two outcomes
s.min: minimum of the variances of the outcomes, default is 0.01
m.m.o: number of studies on one side of the suppressing line been observed, i.e. non-deterministic suppressing, default is 0, i.e. deterministic suppressing
s2.dist: options for generating the outcomes' variances. 1=runif, 2=runif^2, 3=runif^4, 4=rnorm
verbose: logical, galaxy plot the studies? Default FALSE

Author

Chongliang Luo, Yong Chen

References

Luo C, Marks-Anglin AK, Duan R, Lin L, Hong C, Chu H, Chen Y. Accounting for small-study effects using a bivariate trim and fill meta-analysis procedure. medRxiv. 2020 Jan 1.