> people who blindly click OK on every pop up are not representative of your typical user

You could unbias the data by including the metric determining how long did it took them to click "Ok" and whether they actually reviewed the data before agreeing.