The Willow Of Darkness
Reality is biased. Truth is biased. If X is, then not-X cannot be in the given situation.
What you are doing amounts to throwing your hand in the air and ignoring that which is(conflicting data, for example, might be showing both issues exist indifferent circumstances).
I love statistics, the best parts of my job involve doing statistical regressions, however, whenever we start doing them there are a plethora of things we need to take into account. When testing for causality, that is, creating a model that tries to show how a dependent variable is linked to a series of independent variables, it is important to come up with a theory that first explains what is going on. Why were certain explanatory variables picked, why were others excluded? What is the expected sign of each variable, and is there any inclination towards their respective magnitudes?
Having two studies that each try to test a result in a similar fashion, and come up with conflicting results might simply mean that there is no general applicable theory behind why a dependent variable happens. Instead, there are place and time sensitive variables that are unaccounted for in the model. This is called omitted variable bias, and will skew the statistical significance and magnitude of studies.
If there was large inconsistencies in the past research, individuals would be unable to make any claim, as clearly there would be variables with a strongly determinant effect on growth that were excluded and masking the real effects.
It should be noted however that statistics is not truth, it is a syllogism used in logic to try and add a level of empiricism to theory.