The finding that many social science studies fail to replicate should be an embarrassment to social scientists. I suppose any participant has an explanation. I can say after being involved in social science research for many years that there is a lot of pressure for numbers (sounds like the excuse educators used when cheating on student tests).
There were many ways to play the game. My observation recognized strategies that would not necessarily result in failure to replicate., but would boost numbers For example – questionnaire studies rather than studies based in observable behaviors, multiple authorships to boost the number of lines on the vita, short duration studies. I assume the large and expanding number of journals plays to the same motivational forces.
I always thought tenure was to allow an escape from this kind of pressure and a greater focus on quality. All systems seem to revert to variables that are easy to quantify.