zhu_dispersal.Rmd

---
title: "Zhu dispersal"
output:
  word_document: default
---

```{r load_data, warning=FALSE, message=FALSE, echo=FALSE}
source("init.r")

library(pander)
library(data.table)
library(dplyr)

# create dummy for whether/not each person still lives in their natal household (adults only)
mosuo.adults$HasDispersed = ifelse(mosuo.adults$HHID2012 == mosuo.adults$NatalHousehold, 0, 1)

mosuo.adults$DispersalStatus = case_when(
  mosuo.adults$HHID2012 == mosuo.adults$NatalHousehold ~ "Never dispersed",
  mosuo.adults$NatalHousehold %in% c("9", "", "N") ~ "Dispersed from outside Lugu Lake",
  substring( as.character(mosuo.adults$HHID2012), 1, 1) == substring( as.character(mosuo.adults$NatalHousehold),1,1) ~ "Dispersed within village",
  TRUE ~ "Dispersed between villages"
)

mosuo.adults = subset(mosuo.adults, EthnicGroup=="Mosuo")
```

```{r, echo=FALSE, message=FALSE}
# code modified from: http://stackoverflow.com/a/34587522
res = mosuo.adults %>%
  group_by(zhubo1, DispersalStatus) %>%
  dplyr::summarise(n = n()) %>%
  dplyr::mutate(freq = n / sum(n))

#make an 'export' variable
res$export = with(res, sprintf("%i (%.1f%%)", n, freq*100))

# sort out factors
res$zhubo1 = as.factor(res$zhubo1)
levels(res$zhubo1) = c("Not zhu", "Zhu")

res$DispersalStatus = factor(res$DispersalStatus, levels=c("Never dispersed", "Dispersed from outside Lugu Lake", "Dispersed between villages", "Dispersed within village"))

#reshape again
output = dcast(DispersalStatus~zhubo1, value.var="export", data=res, fill="missing") #use drop=F to prevent silent missings
```

```{r, results='asis', echo=FALSE}
knitr::kable(output)
```

```{r, echo=FALSE, results='hide'}
zd.table = table(mosuo.adults$DispersalStatus, mosuo.adults$zhubo1)
( zhubo.dispersal = chisq.test( zd.table ) )

df = zhubo.dispersal$parameter
chi = zhubo.dispersal$statistic
```

Being labelled  *zhu* had no effect on dispersal ($\chi^2_{`r df`}$ = `r round(chi, 3)`; *P* = `r round(zhubo.dispersal$p.value, 3)`).