negative values from numpy.random.normal #24
Description
For smaller values or with large margins of error, the numpy.random.normal in approximate_mean may return a negative number which won't make sense in context. We should probably just use max(0, simulated_value) instead.