Microsoft Office Tutorials and References

In Depth Information

Exhibit 3.31
Histogram of difference in each teen’s page views

is taken in our sample, the interval will be even tighter (< 0.74832

). The larger

the sample, the smaller the interval for a particular conﬁdence interval.

...

Let us now move to a more sophisticated form of analysis, which answers

questions related to our ability to generalize the sample result to the entire teen

population. In the
Data Analysis
tool, there is an analysis called a
t-Test
. A t-test

examines whether the means from two samples are equal or different; that is,

whether they come from population distributions with the same mean or not. Of

special interest for our data is the
t-Test: Paired Two Sample for Mean
.Itisused

when
before
and
after
data is collected from the same sample group, in our case the

same 100 teens being exposed to both the
new
web-site and the
old
.

By selecting
t-Test: Paired Two Sample for Means
from the
Data Analysis
menu,

the two relevant data ranges can be input, along with a hypothesized mean differ-

ence, 0 in our case, because we will assume
no
difference. Finally an
alpha
value is

requested. The value of alpha must be in the range 0 to 1. Alpha is the signiﬁcance

level related to the probability of making a
type 1 error
(rejecting a true hypothe-

sis); the more certain you want to be about
not
making a type 1 error, the smaller the

Search JabSto ::

Custom Search