Because the poll was done via the Internet, it didn't include "the people who most depend on the Postal Service and Saturday delivery — those who aren’t online all the time, like seniors and people in rural areas without good broadband," Save the Post Office notes.
Unlike scientific, random-sample polls, this one used “a blended sample of panel and non-panel,” USPS said in its press release. It didn't explain those terms, but Save the Post Office does: "That means some of the people surveyed were recruited to participate. The Postal Service doesn’t provide any information about how the recruitment was done or who was selected to participate or what portion of the 1,000 people surveyed was pre-selected. One can only imagine how the selection process might have skewed the results."
Save the Post Office also questions the questionnaire, saying it "includes questions that are framed in ways that inevitably affect how people responded. The first question on the survey, for example, is: 'Before today, do you recall hearing anything about the financial losses that the Postal Service experienced last year, of approximately $15.9 billion?' The survey thus begins by presenting the participants with a huge deficit number, which undoubtedly inclines the average person to favor cost-cutting measures like ending Saturday delivery." (Read more)