##### Code for Analyzing Categorical Variables in R


### Import Data


datum=read.csv(file.choose()) ### creates a data frame named datum from imported csv file
head(datum)


### In the newest version of R, we must tell R that 'Sex' is a categorical (or 'factor') variable


datum$Sex=as.factor(datum$Sex) # Tell R that Sex is a 'factor' or categorical variable
levels(datum$Sex) #Check that it worked by asking R what the groups of 'Sex' are


### Plot Data


plot(Mass~Sex,data=datum) #Plot relationship between sex and mass
#Note that R plots a box and whiskers plot because Sex is categorical


### Run a t-test using the t.test command


help(t.test) ### pull up help file for 't.test'
results=t.test(Mass~Sex,data=datum,var.equal=TRUE) ### run a t-test, call results 'results'
#var.equal=TRUE argument ensures homoscedasticity assumption for comparison to linear model
#default is var.equal=FALSE which is standard for t-tests; allows for unequal variance of groups
results ### prints results - Note absence of 'summary' statement - one of few tests that doesn't use 'summary'
#Note results don't provide estimate of effect!


### Run a t-test using the lm command


resultslm=lm(Mass~Sex,data=datum) ### runs a regression with mass as dependent and sex as independent variables
summary(resultslm) ### give results from regression analysis
confint(resultslm) ### give confidence intervals of coefficient estimates


### Run regression on dummy coded variable


resultsDum=lm(Mass~Males,data=datum) ### runs a regression with mass as dependent and sex as independent variables
# Note that Males = 1 is males
summary(resultsDum)
confint(resultsDum) ### provide confidence intervals on estimates in 'resultsDum'


###Import data 2


datum2=read.csv(file.choose()) ### creates a data frame named datum2 from imported csv file
head(datum2) ### check data was imported properly


### In the newest version of R, we must tell R that 'Sex' is a categorical (or 'factor') variable


datum$Sex=as.factor(datum2$Sex) # Tell R that Sex is a 'factor' or categorical variable
levels(datum2$Sex) #Check that it worked by asking R what the groups of 'Sex' are


### Plot Data


plot(Mass~Sex,data=datum2) #Plot relationship between sex and mass
#Note that R plots a box and whiskers plot because Sex is categorical


### run an ANOVA using aov command


results3=aov(Mass~Sex,data=datum2) ### generate an ANOVA for effect of sex on mass, call it results3
summary(results3) #### give summary of ANOVA
#returns anova table with SSE, MSE, F statistic and p-value
#Significant p-value indicates at least 2 groups are significantly different


### run an ANOVA using the lm command


results3lm=lm(Mass~Sex,data=datum2) ### run a regression with mass as dependent and sex as independent varialbes
summary(results3lm) ### Give results
#note F statistic from lm and aov give same results, F statistics and p-values are at bottom of summary output


### run an ANOVA using dummy coded variables


results3Dum=lm(Mass~Males+Herm,data=datum2) ### runs a regression with mass as dependent and sex as independent
#Note that + in the lm equation above doesn't actually mean 'add' - it means include both variables in the model
summary(results3Dum)
#Note that results3Dum and results3lm are exactly the same


### How to change the reference category


help(relevel) #use the relevel command to change reference categories
results4=lm(Mass~relevel(Sex,ref="Herm"),data=datum2) #Changes the reference group to "Herm"
summary(results4)