DHT22 gives incorrect humidity readings by 20%

Ive searched far and wide for a solution to my problem.
Ive tried multiple DHT22 devices and even a couple of the many available libraries.
But still the DHT22 reads about 20% too high.
If I place it in a calibration bag with a hygrometer, it stops increasing its readings at 72.9% ( it should read 75% ) so I'm not even sure if a simple offset can solve the problem.
Its stranger still that the DHT22 claims +/- 0.1% accuracy
Based on all the other topics here, most people have trouble getting readings - I test for nan - its all working, just wildly wrong. Even the wide voltage supply and different pull-up resistors are said to make no difference.
Any ideas?

From: DHT22.pdf (sparkfun.com)

Accuracy humidity +-2%RH(Max +-5%RH); temperature <+-0.5Celsius

Humidity hysteresis +-0.3%RH

Yes agreed - so it shouldnt be reading 20% higher than actual, eh?

Yea. That's why I tossed using the DHT's for temperature readings, too many cheap knockoffs.

Only 20% higher ? That is actually very good for a DHT22.
The sensor can be influenced by certain solvents from plastic bags.
Trying to re-calibrate it is hard to do and not worth it.
The accuracy in the datasheet is just a random number.

This is a big test: Compare different i2c temperature and humidity sensors (SHT2x, SHT3x, SHT85, ..)

Thanks @Koepel - that link is great - really good to know Im not going mad