The survey was conducted by Flexjobs, a job listing site focusing on remote positions. I'm not going to give them my contact information to check, but I would guess that it probably overrepresents temporary and entry-level positions at small to medium sized companies, not big-company corporate jobs that don't need to advertise on external job boards. These jobs would be for people with fewer options, less benefits, and more likely to get overworked.
I wouldn't conclude too much from a survey of only 3,000 employees.
Sure, if you have specific information about the people polled and the general makeup of the population, you can extrapolate these results. I didn't see that it's happening here, it's just the results of a specific group who bothered to answer the poll.
Having 3000 responses doesn't automatically mean the results are meaningful. I can poll 5000 Facebook employees and get results that say 100% of them have a vacation plan and 90% of them have taken a day off in the past year.
I wouldn't conclude too much from a survey of only 3,000 employees.