Before actually
analyzing data to answer questions such as the CEO’s queries
above, you will need to assess the data for integrity, clean any obvious
errors and mistakes, and prepare the data for final analysis. These
checks may include:
-
-
Size of the fifth respondent
is captured as “Bigg,” obviously a typographical error.
-
The “Satisfaction03”
score for Respondent 2 is captured as a “55”, but this
is supposed to be a 1-7 scale.
-
These are data entry
mistakes. While easy to spot in such a small with the eye, you’ll
not see this in a bigger table easily. Mis-entered data can seriously
impact any analysis.
-
Missing
data: There is missing data; we need to assess
and possibly deal with this as discussed in Chapter 4.
-
Multi-item
scales assessing
trust and satisfaction: We need to assess and
aggregate these into single measures of the variables if possible.
We need to pre-assess
and clean our data. We usually do these sorts of assessments through
basic descriptive statistics and variable associations. Therefore,
the next four chapters will sequentially discuss the following:
-
Chapter 6 discusses how to create,
change and manipulate data, as well as give an overview of some other
topics. To do things like create aggregated variables from multi-item
scales, we’ll need these skills.
-
Chapter 7 discusses the essential
descriptive statistics we use for single variables.
-
Chapter 8 discusses basic measures
of variable association.
-
Chapter 9 discusses using these
analyses in an initial set of steps for the purposes of data checking,
cleaning, and preparation.