Introduction to R

Independent samples t-tests are easy in R. Try:

t.test(bmi ~ gender, var.equal = TRUE, conf.level = 0.95, data=data)

This uses the confidence level of 0.95 (conf.level = 0.95) which is the default, but you could use 0.99 if you like.  Note that this test assumes equal variances (var.equal = TRUE), which is again the default, but you could change this to FALSE if need be.  To test whether or not the variances are equal, use:

var.test(bmi ~ gender, data = data)

There is also the Fligner-Killeen test of homogeneity of variances, which a non-parametric test and apparently robust against departures from normality:

fligner.test(bmi ~ gender, data=data)

Oneway analysis of variance is relatively straight forwards. However, this is best done by creating an object for the results of the ANOVA, and requires that the grouping variable be a made into factor.  For example for a comparison of BMI be boot comfort rating, we would use comfortf as it is a factor, and an object “anova_bmi_comf” to store the results:

anova_bmi_comf <- aov(bmi ~ comfortf, data = data)

anova_bmi_comf

summary(anova_bmi_comf)

Non-parametric versions of the t-test and ANOVA are also available.  For example:

kruskal_bmi_comf <- kruskal.test(bmi ~ comfortf, data = data)

kruskal_bmi_comf