rdssim: A simple respondent-driven sampling simulator

Description Usage Arguments References Examples

Simulates referral chains/trees for Respondent-driven sampling

1 2	rdssim(adjList, referral.type = c("sRW", "acRW"), wRreplacement = TRUE, nSamples, nReferrals, seedNode, rseed)

`adjList`	is the object `adj2list` returns. It has the adjacency list of the input social network. It assumes that adjList represents an undirected network. Refer to `adj2list`.
`referral.type`	referral type. It can be "sRW", simple random walk, or "acRW", anti-cluster random walk.
`wRreplacement`	if TRUE, then the referral chain/tree will be collected with replacement.
`nSamples`	number of samples to be collected.
`nReferrals`	number of referral per participant (or node)
`seedNode`	the starting seed participant (or node).
`rseed`	random seed for random number generators inside the code. It is useful for reproducing results.

Khabbazian, Mohammad, Bret Hanlon, Zoe Russek, and Karl Rohe. "Novel sampling design for respondent-driven sampling." Electronic Journal of Statistics 11, no. 2 (2017): 4769-4812.'

generate.network <- function(N,con){
   A <- matrix(0,ncol=N,nrow=N) # initialize with zeros
   A[upper.tri(A)] <- rbinom(N*(N-1)/2,1,con) # fill up the upper triangle
   A[lower.tri(A)] <- t(A)[lower.tri(A)] # fill up the lower triangle to obtain a symmetric matrix
   A
}

ER <- generate.network(300,0.6)

adjL <- adj2list(ER)
res  <- rdssim(adjL, referral.type="sRW", wRreplacement = T, 
                nSamples=100, nReferrals=3, seedNode=10, rseed=1)