Using data
obviously implies collecting/measuring it first, and then capturing
these measurements into a database or spreadsheet. I cannot cover
a massive amount of research methodology in this book, so the enthusiastic
researcher should also study a complete book on methods. This chapter
makes a few critical points about data collection and capture for
statistics.
Have another look at
Chapter 2, notably the process of statistics diagram. The major data
challenges discussed there are choosing the correct samples and populations,
choosing the correct constructs and variables, and asking the right
questions. I cover each of these topics here. If there is one major
point that can be made about getting research methodology right it
is this:
There
are two cardinal errors that can be made in research; all other methodology
issues are secondary to these. The two cardinal errors are measuring
the incorrect variables or having an incorrect (wrong or non-representative)
sample or set of observations. If your sample is poor
then your analysis will not be meaningful to the population you wish
it to represent, and you will be unlikely to replicate your findings.
If you measure the wrong set of variables then you have missed the
focus of the study. Even if you have measured some correct variables
but you have an inadequate variable set, your study is being conducted
without information in its correct context. We call this specification
error. Having good methodology and doing good analysis
on the wrong variables or wrong sample is often problematic at best
and frequently meaningless.
Having said this, it
is also important to do a good job on the other methodology details,
such as how you have measured and analyzed data. I make some further
comments in the next few sections.