Microsoft Office Tutorials and References
In Depth Information
if we find that associates are concentrating on the sale of higher priced station wag-
ons to a small number of demographics, a decision maker may want to take steps to
change this focused selling. It is possible that other demographics will be interested
in similar vehicles if we apply appropriate sales incentives.
In Chap. 6 we will focus on statistical analysis that can be performed with tech-
niques appropriate for qualitative and quantitative data. Among the techniques that
will be examined are Analysis of Variance (ANOVA), tests of hypothesis with t-
tests and z-tests, and chi-square tests. These statistical tools will allow us to study
the effect of independent variables on dependent variables contained in a data set,
and allow us to study the similarity or dissimilarity of data samples. Although these
technical terms may sound a bit daunting, I will establish clear rules for their appli-
cation, certainly clear enough to permit a non-statistician to apply the techniques
correctly. Now back to the techniques we will study in this chapter.
5.2 Essentials of Qualitative Data Analysis
In Chap. 4 we discussed the essential steps to prepare, organize, and present qual-
itative data. The preparation of qualitative data for presentation should also lead
to preparation for data analysis; thus, most of the work done in presentation will
complement the work necessary for the data analysis stage. Yet, there are a number
of problems relating to data errors that can occur due to problems in data collec-
tion or transcription that require special consideration. These errors must be dealt
with early in the data analysis process or the analysis will lead to inaccurate and
unexplainable results.
5.2.1 Dealing with Data Errors
Data sets, especially large ones, can and usually do, contain errors. Some errors
can be uncovered, but others are simply absorbed, never to be detected. Errors can
occur due to a variety of reasons: problems with manual keying or electronic trans-
mission of data onto spreadsheets or databases, mistakes in the initial recording
of data by a respondent or data collector, and many other sources too numerous
to list. Thus, steps insuring the quality of the data entry process need to be taken.
As we saw in the previous chapter, where we assumed direct data entry in work-
sheets, we can devise data entry mechanisms to facilitate entry and to protect against
entry errors.
Now let us consider data that has been transcribed onto an Excel worksheet from
an outside source. We will focus on the rigorous inspection of the data for unex-
pected entries. This can include a broad range of data inspection activities, ranging
from sophisticated sampling of a subset of data elements to exhaustive (100%)
inspection of all data. If a low level of errors can be tolerated, then only a sam-
ple of the data need be reviewed for accuracy. This is usually the case when a data
set is very large, and the cost of errors is low relative to the cost of verification. If
Search JabSto ::

Custom Search