BTW, here's why I think the official results are closer to reality than this guy's estimates.
First of all, there are UN polls (which I trust more than any Russian or Ukrainian polls, given how politicized this matter has always been between the two countries). Before the election of Yanukovich as president, in 2009, these shown 70% in favor of joining Russia, and only 15% against it. These figures gradually decreased to 65% under Yanuk. After he was ousted and maidan took over in Kiev, I'd expect this number to shoot back to 70%, at the very least, and likely grow even further because of heavy Russian propaganda pitching maidan as "fascists" and "banderovtsy" (and most people in Crimea were getting their news mainly from Russian TV channels, not Ukrainian ones). With 16% undecided in the original polls, I can totally see support at 75% or so.
The other, more indirect indicator is language. This is more reliable than ethnicity, because Ukraine has a very blurry line between the two, with plenty of people self-identifying as Ukrainian on the basis of their family name or their parents' self-identification, but culturally and linguistically behaving as Russians in all other aspects. For example, in Crimea, less than 60% self-identified as Russians, but 77% listed Russian as their native language of communication (and 90% in Sevastopol) - whereas only 10% listed Ukrainian as such. It's even more skewed if you ask people which language they primarily use at home with family (since some would consider native language tied to ethnicity) - in this case you're looking at something like 95% in Crimea. Crimea also had extremely low levels of bilingualism, with only 30% of the population proficient in Ukrainian at all - a stark contrast with the rest of the country, where bilingualism is the norm, except for Western Ukraine where Ukrainian dominates with a similar proportion.
Now if you look at these language figures and consider them a proxy for political affiliation (an oversimplification, but not an unreasonable one), they also match pretty closely to what the official claim was: 80% turnout, 97% in favor of the union. What this looks like to me is that most everyone who was against the union did not vote at all, considering the referendum illegal (Crimean Tatars, in particular, had an open boycott); and most of those who wanted the union came and voted. The real numbers are probably closer to 70% for turnout and 90% in favor or thereabouts, but still a clear supermajority.