For some values of 'plummeting'

Based on tonight's No. 2 story at the Fair 'n' Balanced Network, would you say that over the past year, "trust in government" has:

Soared
Increased
Decreased
Plummeted
Or stayed about the same?

Shall we have a hint?

When it comes to Washington controversies, most American voters think Benghazi, the IRS and the government’s electronic surveillance program are serious matters. A Fox News poll also finds that less than four in 10 voters trust the federal government.

The new poll, released Thursday, finds 37 percent of voters answer “yes” when asked: “would you say you generally trust the federal government?” Six in 10 say they don’t trust the government, down a touch from a high of 62 percent (June 2013 and July 2011).

That's an interesting result on distrust. What do you suppose might be the corresponding result on "trust"? 

You could fairly conclude that "trust in government" has plummeted from 35% to 37% over the past 11 months, much as "distrust" has soared from 62% to 60%. Neither of those changes, needless to say, is statistically significant* -- but each is a hell of a lot more likely to represent a positive (trust is higher, distrust lower) than a negative change in the population value.

Fox, as usual, has paid for and gotten a competent, professionally executed survey. Give Fox credit for posting the results in a way that lets a lay reader see how fundamentally dishonest Fox is in describing the numbers it pays for. Surely, though, at some point, there has to be a corrosive effect on the soul of the Fox staffer whose assignment is to make the numbers look good for the people who write the checks. We should be prepared to offer these folks amnesty; who knows what might happen if they all turn up at the gates someday pleading to be considered as real journalists?

* Today's exercise: Calculate the standard error of proportion for a random sample of 1,025, assuming p(1-p) = .25. How does that compare with the result if p =.37?