I got those citations from NASA's website. I have a lot of trust for those nerds. You'l never get a 100% accurate statistic on the issue, but the evidence is overwhelmingly strong that it's around 97%.
Right. You can never get a 100% statistic.
However, your sample should be representative.
On a side note.
You can skew a lot of statistics in a lot of ways.
* Not taking a representative sample - Common in polls (% of Americans believe - samples taken from CNN (Left) or Fox News (Right) only) - If you can get a different number in the same poll with a different sampling body, then you are not taking a representative sample. Commonly in polls they do not relay where the information comes from and if they do then it buried so that most people will not find it.
* Not taking a random sample - Common in studies looking for a specific answer - Asking residents in Wyoming about how they feel about the wilderness rule and than stating the Americans feel this way. Looking for the answer that it is a good rule or Asking non resident hunters in Wyoming about how they feel about the wilderness rule. Looking for the answer that it is a bad rule.
* Questions - You can write questions that will lead a person to answer one way - (Do you agree the government using necessary means to identify terrorists vs Do you agree the governement should be able to tap phone lines to look for potential terrorists) You may agree with the first and disagree with the second.
* Results interpretation - If you had a poll - Donald Trump, Hillary Clinton, and other. And they split 33.34 of the vote each. Then the trump team could say that only 33% of Americans agree with Hillary Clinton implying that 66% of Americans support him. The Hilary team could say that 66% oppose Donald Trump, implying that the 66% support here. Neither are inaccurate, but both are misleading.
* Results units - If you did a study and the result came out as 2 kg of methyl ethyl bad shit in 100 acre of corn soil, but then you punish the results in ug. 2,000,000,000 ug comes out in the report. It is then something that you can say that it is 20,000,000 ug/acre of methyl ethyl bad shit and use that to justify your position.
There is a lot of things that have to be taken into account to produce quality statistics.