St. Lawrence, Brasfield, Shirley, Jefferson, Alleyne, and O'Brannon (1995) investigated the effects of an 8 week Behavioral Skills Training (BST) program aimed at reducing the risk of HIV infection among African-American adolescents. The study followed males and females from a pretest to a 12 month follow-up, recording the frequency of condomprotected sex. (They also had a control condition, but I am going to look only at the males in the BST condition for this exercise.) The actual dependent variable is one thousand times the natural logarithm of the frequency of protected sex. (I multiplied the $\log$ by 1,000 to eliminate decimal values.) The data for males follow.
$$
\begin{array}{lccc}
\text { Pretest } & \text { Posttest } & \begin{array}{c}
\text { Follow-up } \\
\text { 6 Months }
\end{array} & \begin{array}{c}
\text { Follow-up } \\
\text { 12 Months }
\end{array} \\
\hline 07 & 22 & 13 & 14 \\
25 & 10 & 17 & 24 \\
50 & 36 & 49 & 23 \\
16 & 38 & 34 & 24 \\
33 & 25 & 24 & 25 \\
10 & 07 & 23 & 26 \\
13 & 33 & 27 & 24 \\
22 & 20 & 21 & 11 \\
04 & 00 & 12 & 00 \\
17 & 16 & 20 & 10
\end{array}
$$
(a) Calculate and plot the means.
(b) Use the analysis of variance to draw the appropriate conclusions.