Page 197 - Intermediate Statistics for Dummies

15_045206 ch09.qxd 2/1/07 10:14 AM Page 176
176
Part III: Comparing Many Means with ANOVA
The results of the t-test done to compare the spitting distances of males and
females in the section “Comparing Two Means with a t-Test” (see Figure 9-1)
showed that males and females were significantly different on mean seed spit-
ting distances. So I would venture a guess that if you include gender as well as
age group thereby creating what statisticians call a two-factor ANOVA (or two-
way ANOVA), the resulting model would fit the data even better, resulting in
2
2
higher values of R and R adjusted. (See Chapter 11 for two-way ANOVA.)
Up-front rejection the best policy
for most refusal letters
Many medical and psychological studies use
designed experiments to compare the responses
sonable. The data were also shown to have a
bell shape.
of several different treatments, looking for differ-
ences. A designed experiment is a study in enough possible values that ANOVA isn’t unrea-
The null hypothesis would be Ho: Mean
which subjects are randomly assigned to treat- responses to the three types of rejection letters
ments (experimental conditions) and their are equal, versus Ha: At least two forms of the
responses are recorded. The results are used to rejection letter resulted in different mean
compare treatments to see which one(s) work responses.
best, which ones work equally well, and so on.
In the end, the researcher did find some signif-
One example of one such experiment that
icant results. In other words, the different ways
employs ANOVA is from The Ohio State
the rejection letter was written affected the par-
University research press release Web site. The
ticipants in different ways. Using multiple com-
experiment tested three traditional principles
parison procedures (see Chapter 10), you would
of writing refusal letters:
be able to go in and determine which forms of
Using a buffer — a neutral or positive sen- the rejection letters gave different responses
tence that delays the negative information and how the responses differed.
Placing the reason before the refusal So in case you have to write a rejection letter at
some point, the researcher recommends the
Ending the letter on a positive note as a way following guidelines for writing it:
of reselling the business
Don’t use buffers to begin negative
Subjects were randomly assigned to treat-
ments, and their responses to the rejection let- messages.
ters were compared (likely on some sort of Give a reason for the refusal when it makes
scale such as 1 = very negative to 7 = very pos- the sender’s boss look good.
itive with 4 being a neutral response).
Present the negative positively but clearly;
This scenario can be analyzed by using ANOVA. offer an alternative or compromise if
It compares three treatments (forms of the possible.
rejection letters) on some quantitative variable A positive ending isn’t necessary.
(response to the letter). You can argue that this
isn’t a continuous variable, because it has