Sampling Designs - Sampling Spatial Units for Agricultural Surveys - page 134

> set.seed(200694)
> fstphsrs < - srswor(2*n,N)
> framefst < - framepop[fstphsrs ¼¼ 1,]
> sndphstr < - strata(framefst,"strataid",
+ size ¼ rep(25,4),method ¼ "srswor")
> sndphstr < - getdata(framefst,sndphstr)
> par(mar ¼ c(1,1,1,1), xaxs ¼ "i",yaxs ¼ "i")
> plot(framefst$xc,framefst$yc, axes ¼ F,cex ¼ 2,
+ pch ¼ 1,xlim ¼ c(0,1),ylim ¼ c(0,1))
> abline(h ¼ 0.5,lty ¼ 2,lwd ¼ 2)
> abline(v ¼ 0.5,lty ¼ 2,lwd ¼ 2)
> box()
> points(sndphstr$xc,sndphstr$yc, pch ¼ 19, cex ¼ 2, col ¼ gray(0.8))
> points(framepop$xc,framepop$yc, pch ¼ 19, cex ¼ 0.5)
> points(framefst$xc,framefst$yc, pch ¼ 1 , cex ¼ 2)
> twoph < - framefst
> twoph$sel < - rep("FALSE",nrow(twoph))
> twoph[sndphstr$ID_unit,"sel"] < - "TRUE"
> twoph$sel < - as.logical(twoph$sel)
> twoph$Prob1 < - rep(2*n/N,nrow(twoph))
> twoph$Prob2 < - rep(NA,nrow(twoph))
> twoph[sndphstr$ID_unit,"Prob2"] < - sndphstr$Prob
> d2ph < - twophase(id ¼ list(~id,~id),
+ strata ¼ list(NULL,~strataid),fpc ¼ list(~Prob1,~Prob2),
+ data ¼ twoph,subset ¼ ~sel)
> estr2ph < - svytotal(~yobs+as.factor(q1obs),d2ph)
> estr2ph

total SE

yobs 94111.5 1934.412

as.factor(q1obs)1 289.8 44.526

as.factor(q1obs)2 343.8 46.003

as.factor(q1obs)3 366.4 46.815

6.8 Sample Coordination and Longitudinal Surveys

Sample coordination has been a main topic of interest for many years, in both its
versions: negative and positive coordination. In negative coordination, the aim is to
minimize the number of common units between different samples drawn succes-
sively from a frame, mainly to reduce the statistical burden . In positive coordina-
tion, we want to maximize the overlap (Nedyalkova et al. 2008 ).