Among other things, a demonstration, again, that "science" cannot be trusted, from the declaration that vaccines are necessarily not doctored with poisons to the insistence that chemtrails are not changing the world's weather.
Note, incidentally, how overt failures of many "disciplines" are. If "economics" was necessarily valid, it could have predicted many major problems in the markets. "Psychology" is questionable in that anything beyond basic response tendencies is derived from case studies of seriously ill people, leading to "psychologists" viewing necessarily everyone as mentally ill. And no one in "science" seems to have questioned the fact that "psychology" recognizes dozens of mutually distinct "theories" about how all people's personalities work.

Among other things, too, an emphasis on statistics can be questionable. If a phenomenon occurred in 100 percent of all cases, or even only 90 percent, statistics would not be necessarily to identify it as present. Statistics serves those who want to posit manifestations that wither almost never occur, or are so vaguely "defined", they can be said almost not to exist,

People fear and resent things they dont, and cant, understand. This includes 'people' like poor julian above, who 'knows' that respect for evidence will be the end of his fantasy 'realm'.

Cold turkey julian. Its the only way to get clean.

Among other things, too, an emphasis on statistics can be questionable

Yes. especially when they show that the results of prayer are statistically indistinguishable from wishful thinking.

Ditto with prayer and sugar pills.

Statistics serves those who want to posit manifestations that wither almost never occur

Statistics serve those who wish to isolate an effect out of a group of other factors, for example a biased coin in a gambling game which prefers heads instead of tails; such effects are never 100% or 0% but they still make the house win in the long run.

Likewise, one can use statistics to recieve a milliwatt-level radio signal from the other side of the globe, which compares to the sound of a fart in a hurricane, yet it is definitely there and definitely real.

The problem with statistics only comes from weak studies with small and/or biased samples, where there study data doesn't cover enough material to detect a signal even if it was there, and can easily mistake the noise for the signal. It's this jumping to conclusions which is at the root of the issue, partly because of the publish or perish culture that rewards results with "more research needed".