The government sends out a survey to get employment data, but they don’t get responses in time for their initial reports, so those are usually off and have to be revised later.
The surveys are always incomplete, and a lot of statistical guesses have to be made.
The once a year reports are better; maybe they should only have the once a year reports?
The difficulty of measuring country-wide employment in a short period of time and methodological flaws are the causes of revisions, not political bias.