Incorrect data that you think is correct is obviously worse than no data:
a) You're likely to stop looking for better data once you have it.
b) You're more likely to act on it than acting on no data, and the work you do based on your incorrect data is likely to be either be less useful than is acceptable (otherwise you would have acted on no data at all and gotten a similar result) or it could even be harmful/counterproductive, depending on how wrong the data is.
c) Worst of all, you might help disseminate the incorrect data (and perhaps even argue against the correct along the way), compounding the damage demonstrated above.
a) You're likely to stop looking for better data once you have it.
b) You're more likely to act on it than acting on no data, and the work you do based on your incorrect data is likely to be either be less useful than is acceptable (otherwise you would have acted on no data at all and gotten a similar result) or it could even be harmful/counterproductive, depending on how wrong the data is.
c) Worst of all, you might help disseminate the incorrect data (and perhaps even argue against the correct along the way), compounding the damage demonstrated above.