Microsoft Office Tutorials and References
In Depth Information
2. There is a broad range in the difference of data (Delta), with 51% occurring
from 2 to 6 pages and only 21% of the teens not responding positively to the new
website.
3. The 95% confidence interval for our sample of 100 is approximately 0.75 units
about (
) the sample mean of 11.83. In a sense, the interval gives us a measure of
how uncertain we are about the population mean: larger intervals suggest greater
uncertainty.
4. A t-Test: Paired Two Sample for Means has shown that it is highly unlikely that
the means for the old and new views are equal. This reinforces our growing
evidence that the website changes have indeed made a positive difference in page
views among teens.
5. To further examine the extent of the change in views for individual teens, we
find that our Correlation tool in Data Analysis suggests a relatively low value of
positive correlation. This suggests that although we can expect a positive change
with the new website, the magnitude of change for individuals is not a predictable
quantity.
±
3.7 Summary
Data analysis can be performed at many levels of sophistication, ranging from sim-
ple graphical examination of the data to far more complex statistical methods. This
chapter has introduced the process of thorough examination of data. The tools we
have used are those that are often employed in an initial or preliminary examina-
tion of data. They provide an essential basis for a more critical examination of data,
in that they guide our future analyses by suggesting new analytical paths that we
may want to pursue. In some cases, the analysis preformed in this chapter may be
sufficient for an understanding of the data’s behavior; in other cases, the techniques
introduced in this chapter are simply a beginning point for further analysis.
There are a number of issues that we need to keep in mind as we embark on the
path to data analysis:
1. Think carefully about the type of data you are dealing with and ask critical ques-
tions to clarify where the data comes from, the conditions under which it was
collected, and the measures represented by the data.
2. Keep in mind that not all data analysis techniques are appropriate for all types of
data: for example, sampling data versus population data, cross-sectional versus
time series, and multi-attribute data versus single attribute data.
3. Consider the possibility of data transformation that may be useful. For example,
our cross-sectional data for the new and old website was combined to create a
difference or Delta data set. In the case of the time series data, we can adjust data
to account for outliers (data that are unrepresentative) or one-time events, like
promotions.
Search JabSto ::




Custom Search