> qbinom(.025,100,.5)
[1] 40
> qbinom(.975,100,.5)
[1] 60
So in this experiment, if the number of heads we saw was between 40 and 60 (out of 100
tosses), we would tentatively accept the null hypothesi
the levels of the cultivar, i.e. do t-tests for 1 vs. 2, 1 vs. 3, and 2 vs. 3. This is a very bad
idea for at least two reasons:
1. One of the main goals of ANOVA is to combine together all our data,
15.1
Linear Regression
Linear regression is a very popular procedure for modeling the value of one variable on the
value(s) of one or more other variables. The variable that were trying to model or pr
mkput = function(sym)cfw_
function(.)cfw_
calcinp <- paste(calcinp,sym,sep=)
tkconfigure(display,text=calcinp)
Notice that were refering to an object called display, even though we havent dened it
y
This correction is performed by default, but can be shut o by using the var.equal=TRUE
argument. Lets see how it works:
> t.test(x,y)
Welch Two Sample t-test
data: x and y
t = -0.8103, df = 17.277, p-
318
Chapter 16
Analysis of Variance
319
16.1
Analysis of Variance
In its simplest form, analysis of variance (often abbreviated as ANOVA), can be thought of
as a generalization of the t-test, because
Note that only the specic widget for which background= was set changes if you want
to change the background for the entire GUI, youll probably have to pass the background=
argument to every widget you
sense for the wine data, suppose we wanted to add Cultivar and all the interactions between
Cultivar and the independent variables to our original regression model. The rst step is
to create a vector
Many of the variables seem to be correlated with each other, so its dicult to see which is
causing the problem. A statistic known as VIF (Variance Ination Factor) can be very useful
in situations like
The plots look much better, so well continue with the analysis of the log of retention.
Df Sum Sq Mean Sq F value
Pr(>F)
level
2 15.588
7.794 22.5241 7.91e-09 *
treatment
1 2.074
2.074 5.9931 0.01607