Glassdoor is fascinating for all the wrong reasons. I've now worked in charity/community, private sector, and government, for big, small and medium.
During my last switch (from private back to government), i noticed what I'll call the "reverse glass door" effect while browsing their profiles. It was weak but it was there. That is to say, if i had to take all the employers, consultancies and corporations I've had to deal with, and rank them from places I've enjoyed myself at the most to places where I would have to genuinely ask whether there was a wage they could pay me to make me work with them again, it seemed that the workplaces I considered the worst held the highest glassdoor star ratings, and my favourites have consistently been around the middle of the pack.
As a statistician I've even considered naming and investigating this effect: explicit rating systems that end up being the inverse of what they purport to be measuring, and I've got some theories as to what could be behind it (me being abnormal, such firms having the resources/need/desire to manage their image the most, high turnover and getting new people to leave such rankings during their honeymoon period), but I've not got round to doing it with any rigor...