I once made an internal G+ post that "voluntary response data measures liquidity, not satisfaction". When you have an organization that people can freely choose to join or leave, then you should expect that any internal surveys will have very high satisfaction numbers - because anyone who is dissatisfied will simply leave and drop out of the sample. Similarly, if you get consistently low numbers for satisfaction, that could be because the people in the sample have low liquidity - they can't easily go elsewhere.
This is why organizations like Congress, Comcast, and Facebook (users, not employees) have such perpetually low approval ratings - for most people stuck with them, there are simply no alternatives, and so they remain beholden to their decisions despite hating them. And similarly, Google employees have very high satisfaction ratings because it's relatively easy to get a job elsewhere or just not need a job at all as a Xoogler, and so anyone who dislikes it just leaves.