In many cases, empirical articles actually represented
blends of different theory-building and theory-
testing components. For example, an article
might describe an examination of a previously unexplored
relationship as a relatively minor facet of
a study while focusing more attention on replicating
previously examined effects. Alternatively, an
article might present some theory-grounded hypotheses
hypotheses
and describe others as based on past findings.
In such cases, in coding we averaged ratings
on the multiple components while giving more
weight to the ones that were more central to a focal
article. As a result, we allowed our coding to include
half-points in addition to integers (1.5, 2.5,
3.5, 4.5).
For analyses involving the five discrete categories