There is much you could
say when reporting a regression. Most importantly:
-
You can usually get away with the
absolute minimum of reporting on your intermediate diagnostic tests.
Generally, you will not show
tables of graphs from these tests (e.g. residual graphs, collinearity
or outlier tables, and the like). Rather, succinctly report what diagnostic
tests you did and anything major you did to adjust the regression.
This can often be done in a short paragraph.
-
Overall, it usually suffices to
present a fairly concise description of various core statistical findings
and then to reflect on what you have found in some depth. This applies
to all major statistical analyses in fact. In regression specifically,
your report will usually revolve around a single table that integrates
fit and slope information, like that seen in
Figure 13.29 Example of an integrated and formatted regression table below.
-
It is very important when reporting
major analyses such as a regression that your report flows
together as a coherent whole. Many beginners fail
to appreciate this, and put together a patchy report that fails to
read as an integrated whole.
-
Finally, reflecting on the practical
and – in the case of academic research theoretical –
implications of your findings is a key capstone for such analyses.
What do your findings mean for management, strategy, your customers,
and the like? Can you extrapolate your findings to some key business
outcomes as we do in Chapter 17?
Ultimately –
especially in academic research – good presentation of statistical
findings can make a big difference to the reception the work gets
by readers.