Snopes, FactCheck and other fact checking type sites love to use weasel works. For example, in the “global warming” or “climate change” debate both attempt to deal with the question as to whether NOAA has manipulated the raw data. Clearly the answer is yes. Yet both sites come out with the answer as now even though they admit that it was “standardized” or “manipulated” in the text.

For example:

“no undue manipulation of temperature data” – FactCheck.org
“NOAA Scientists Falsely Accused of Manipulating Climate Change Data” while stating “procedures for the standardization of which datasets to use” – Snopes

None of this is the language of science. Once you have massaged the data, it is no longer data, but a hypothetical. Sure, models may be good, but they may not be – garbage in, garbage out. And a model is no longer data, it is a model projection of data. Depending on your manipulation, you can achieve many different results.

Snopes continues:

“Karl et al might reasonably be criticized for having been less than rigorous in their data documentation, their findings have been independently verified, “

First of all, if it is impossible to independently verify a revision in data processing. Sure, if you put the same input into the same process, you should get the same output. But the process here is key. If the assumptions used to “standardize” the data are biased or faulty, it doesn’t matter if it is repeatable.

“they also released a revised land record based on data” – “the old NOAA record spliced together warmer ship data with colder buoy data without accounting for the offset between the two; and the new NOAA record puts more weight on higher-quality buoy records and less weight on ship records (versus the old NOAA record which treated ships and buoys equally). ” Carbonbrief.org.

A “revised” data set based on changes in weighting of the data from equal to weighting one that is unequal is purely subjective. It is no longer data, it is, again, hypothesis.

So, when reading the fact checking from many of these sites, watch out for the weasel words to twist a statement into something it isn’t.

Look at the “Did Hillary Clinton Steal $200,000 in White House Furnishings?” article on snopes which they label as “mostly false.” Yet Snopes states at the end of the article:

All told, the Clintons paid back or returned approximately $136,000 worth of furniture, artwork, china and other household items they had kept upon leaving office, with $86,000 of that total consisting of personal gifts they would presumably have been allowed to retain but decided to pay for to avoid the appearance of impropriety.
About $50,000 of the total comprised items they had removed but were later determined to belong to the government. To say the latter were “stolen” is to say more than we know — the removal of the questioned items could have been based a clerical mistake — but in any case an accurate accounting of those items’ worth puts it at only a quarter of what has been alleged: $50,000, not $200,000.

Just be sure to read the words because they could’ve ask the question “Did Hillary Clinton Steal $200,000,000 in White House Furnishings?” And of course as a percentage $50,000/$200,000,000 is de minimis.

How the question is phrased, and how the answer is presented is key.

Did NOAA manipulate temperature data? Unequivocally yes as everyone admits when you read what is written.

Snopes and others tried the same thing with the “Al Gore invented the Internet” claim. By changing the claim from what Gore said to something else, they claimed it was false.