The response variable (Y_{i j}) in a (2^{2}) design can also be expressed as a regression model
Question:
The response variable \(Y_{i j}\) in a \(2^{2}\) design can also be expressed as a regression model
\[Y_{i j}=\mu+\beta_{1} x_{1}+\beta_{2} x_{2}+\beta_{12} x_{1} x_{2}+\varepsilon_{i j}\]
where the \(\varepsilon_{i j}\) are independent normal random variables and each has mean 0 and variance \(\sigma^{2}\).
Because \(\beta_{1}\) is a regression coefficient, it quantifies the change in the expected response when \(x_{1}\) is changed by one unit. The effects are calculated on a change from -1 to 1 or 2 units.
(a) Obtain the expected values of \(\bar{Y}_{1}, \bar{Y}_{2}, \bar{Y}_{3}\), and \(\bar{Y}_{4}\).
(b) Show that the expected value of the main effect of Factor \(A\) is \(2 \beta_{1}\).
(c) Show that the expected value of the \(A B\) interaction effect is \(2 \beta_{12}\).
Step by Step Answer:
Probability And Statistics For Engineers
ISBN: 9780134435688
9th Global Edition
Authors: Richard Johnson, Irwin Miller, John Freund