Is it wise to publish updated results while the data is still being gathered? Random selection should be blind and independent and its quite possible people may change their vote to favor an 'unpopular' number or vice versa.