By Richard Morin
By Richard Morin
President Clinton should have been a pollster. Anyone who can quibble over what the meaning of "is" is has got precisely the right stuff to appreciate how even seemingly small differences in the wording of survey questions can produce significantly different results.
The latest example of the instability of words comes from the flurry of public opinion polls last week that asked Americans what Clinton should do if he is impeached.
The deluge started Dec. 15 when The Washington Post and ABC News reported the explosive results of a weekend poll that asked this question: "If the full House votes to impeach Clinton, what do you think Clinton should do: Fight charges in the Senate, or resign from office."
Nearly six in 10 58 percent said Clinton should quit. It was the first time any media pollster had asked a version of this question. Instantly, that result was everywhere (it appeared on local TV news in Washington, D.C., even before it appeared in The Post), spread by those desperately seeking fresh news to report as the impeachment vote in the House approached. (For the record, this desperate crew included this writer and his colleague in The Post's polling department, Claudia Deane, who shared the byline on The Post's story on the poll findings.)
Two theories immediately surfaced to explain the stunning result.
The first and most widely reported explanation was scandal fatigue. Americans were simply tired of the whole dirty business and were willing to sacrifice Clinton in order to stop the madness in Washington. That theory was partially supported by another survey finding: Only 29 percent of those interviewed said they would be "angry" if Congress eventually impeached Clinton and forced him from office, a number that suggested that most of the country was willing to accept, albeit grudgingly, a White House without Clinton.
The second theory call it the Know-nothing Hypothesis is that many Americans were simply confused by the impeachment process. Specifically, they falsely believed that impeachment meant removal that the game was over when the full House votes on the articles of impeachment. Thus the option that Clinton should fight the charges in the Senate would seem like so much sour grapes. The evidence for this hypothesis came from other polls: About 30 percent in a recent CBS survey believed that impeachment meant automatic removal.
The Post-ABC survey did explain that the House impeaches and then the Senate conducts a trial to see whether or not the president should be removed. But this information was in a question at the beginning of the 24-question survey. The resignation question came at the end of the poll. Perhaps some respondents didn't understand or simply forgot what they were told earlier about the process.
To test those two theories and to learn more about what America meant when a majority said it wanted Clinton to resign The Post and ABC tested different wordings of questions in two separate surveys, both conducted the night of Dec. 15.
Two versions of the "fight or resign" question were asked to different halves of a sample of 759 randomly selected adults. One version of the question carried an introduction explaining the impeachment process, the other version was simply the core question without the introduction.
If people said Clinton should resign, they were asked a new follow-up question: "Do you feel that way mainly because of what [Clinton] did in the Lewinsky matter, or mainly because it would be the quickest way of bringing this matter to an end?" This would be used to measure scandal fatigue.
A knowledge question at the very end of the survey asked respondents if impeachment meant that "Clinton is removed from office and no longer is president," or whether it meant that the president "has to stand trial in the Senate, which would decide whether to remove him from office."
The test worked. Both versions of the questions produced a result nearly identical to the weekend poll: About 57 percent said Clinton should quit, not fight. The test found little confusion over the meaning of impeachment: At the end of the survey, nearly nine out of 10 respondents knew what impeachment meant.
The fatigue theory was confirmed: A clear majority of those who favored resignation said they mainly wanted Clinton to quit to quickly end the scandal.
Another test of question language was conducted that same night on a separate sample of 600 randomly selected adults. The language of the resign-or-fight question was modified. The choice option "fight the charges in the Senate" was replaced by the phrase "continue to serve and stand trial in the Senate."
Those few words made a significant difference: In response to the new question, fewer than half 43 percent said Clinton should resign a dismaying number from the perspective of the White House, but certainly smaller than the 57 percent who wanted Clinton to quit if he was impeached.
Is one question "wrong?" No quite the opposite. They measure different and important things. America fears a continuation of the fight that has paralyzed official Washington for nearly a year. In fact, the data suggest America is far less fearful of a Senate trial (most people currently think Clinton will survive) than it is of a raucous partisan battle.
The latest wave of polling "seems to suggest there is a real risk that his support will be shaken by an impeachment vote, perhaps motivated by a fatigue factor, a desire to see this over and perhaps motivated concerns about whether he can carry on effectively," says Andrew Kohut, director of the Pew Research Center for the People & the Press, who reviewed the results.
But Kohut says it's impossible to predict how public opinion may or may not move in the next few critical weeks. He suspects that a defining moment will occur in just a few weeks, when "people wake up in January and see that the New Year begins with more of this. We could see a backlash [against Republicans] or a throwing in the towel [by Clinton supporters] that will be decisive in a public opinion sense."
Translation: Watch this space.
© Copyright 1998 The Washington Post Company