I concur with those who suggested that it be released after your item is published. In doing so, you're able to facilitate additional research opportunities. Later, after you've released it, then other may attempt replication after your study is published.
A very good question! Probably more researchers would be comfortable with posting/releasing their data after publication than before--the worry being that somebody might beat them to publication with their own data.
No matter what your point of data sharing is, be sure that you've screened the data for accuracy, documented any score modifications (or choices to flag cases as outliers, etc.), and removed any personal identifiers if the data are from human respondents prior to releasing data to others.
Here's a good, multi-disciplinary listing of data repositories: https://www.nature.com/sdata/policies/repositories
I do not like posting them before publication! You can often write more than one article from one study, and I like to wait until most of it is written up.
Here is a standard formulation we used in a study of a PhD fellow:
"Availability of data and materials
The data used in this study forms the basis of a still ongoing PhD study which will be finalized in 2018. According to current Norwegian regulations and practice, the data will be anonymized 31.12.18 and will then be deposited to the publicly available data repository of the Norwegian Centre for Research Data."
Article Improved drug-use patterns at 6 months post-discharge from i...
Research data should be shared in the paper to be published, that is, before publication. This makes for credibility and promotes replication of the research. Research should be credible and replicated.
I concur with those who suggested that it be released after your item is published. In doing so, you're able to facilitate additional research opportunities. Later, after you've released it, then other may attempt replication after your study is published.