mdata <- read.csv("~/Documents/Hybrid migration/Hybrid migration data/hybmigr-submit.csv", header = TRUE, sep = ",", row.names = NULL) mdata$length <- as.numeric(mdata$length) mdata$pred_by_cormorant <- factor(mdata$pred_by_cormorant) mdata$survived_to_spring <- factor(mdata$survived_to_spring) mdata$migr_trips <- as.numeric(mdata$migr_trips) mdata$obs_period_days <- as.numeric(mdata$obs_period_days) mdata$first_migr_numeric <- as.numeric(mdata$first_migr_numeric) mdata$first_migr <- as.Date(mdata$first_migr,"%m/%d/%Y") mdata$first_migr_noNA <- as.numeric(mdata$first_migr_noNA) library(car) library(multcomp) library(MuMIn) #ANALYSIS 1: Migration trips between species and hybrids (GLM) #Limit to individuals that showed activity after the winter springsurv <- subset(mdata, mdata$survived_to_spring =="1") # Poisson GLM for species differences in migration trips trips.model <- glm(migr_trips~Species, data=springsurv, family=poisson) summary(trips.model) anova(trips.model, test = "Chisq") # Tukey Post Hoc tuk.species <- glht(trips.model, linfct = mcp(Species = "Tukey")) summary(tuk.species) #ANALYSIS 2: Date of lake departure between species and hybrids kruskal.test(first_migr_numeric ~ Species, data = mdata) pairwise.wilcox.test(mdata$first_migr_numeric, mdata$Species, p.adjust.method = "BH") #ANALYSIS 3: Model predicting cormorant predation # Limit used variables to avoid NA's in the dataset mdata.log <- subset(mdata, select = c(Species,length, pred_by_cormorant, obs_period_days, migr_trips, migr_frequency, first_migr_noNA)) #Full model full.logmodel <- glm(pred_by_cormorant~Species+obs_period_days+Species*obs_period_days +length+migr_frequency+first_migr_noNA, data=mdata.log, family=binomial) summary(full.logmodel) Anova(full.logmodel) options(na.action = "na.fail") output <- dredge(full.logmodel, rank=AIC) output options(na.action = "na.omit") #Final model log.model <- glm(pred_by_cormorant~Species+migr_frequency+obs_period_days, data=mdata.log, family=binomial) summary(log.model) Anova(log.model)