Source: https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1467-8543.2009.00723.x
It should be normally distributed, but it is skewed left towards negative estimates, suggesting studies that show positive employment effects are less likely to be published.
I should note an important point they mention is employment effects of minimum wage changes may not be static across time and place. So it wouldn't be impossible for there to be both true negative and positive employment effects. It just again is pointing out the systemic effects are much more complex than just basic theory.
Source : https://www.aeaweb.org/articles?id=10.1257%2Faer.20180310&
This one shows clustering around significance with insignificant results less likely to be published, and a tendency towards negative significant results being published.
Researchers/publications were reticent to publish positive effects that went against economic theory dogma. It would be fascinating to see before and after the seminal Card/Krueger paper. It probably helped publication rates of positive employment estimates all else equal.
As for the publication bias around significance, that is not unique to minimum wage research, or even econ. That's a harder issue to deal with. Researchers don't have a lot of incentive to publish null results. Pre-registration helps a lot in that regard.