|

Spinning “Lies, Damned Lies, and Statistics” in Venezuela

Some pollsters – after having long engaged in highly biased polling designed to demoralize the government’s supporters and to embolden the opposition – will issue less biased polls in a last-ditch effort to salvage their own credibility in the face of impending defeat.

“There are three kinds of lies: lies, damned lies, and statistics.” – Benjamin Disraeli, Prime Minister of England (1868, 1874-1880)

As the August 15 referendum on whether Hugo Chávez should continue as president approaches in Venezuela, anti-Chávez pollsters have begun reluctantly issuing polls showing Chávez in the lead. In June, the Washington-D.C. based polling firm Greenberg Quinlan Rosner Research Inc. – working on behalf of the opposition – conducted a poll showing that 49 percent of Venezuela’s registered voters would support President Chávez versus 44 percent that would vote to recall him. Another June poll by the Venezuelan firm DATOS – also commissioned by the opposition – gave Chávez 51 percent of support, against 39 percent who would vote against him.

Recently Chávez challenged other Venezuelan polling firms aligned to the opposition to release the results of their latest polls. Venezuelan Information Minister Jesse Chacón has claimed to have copies of these polls – which favor Chávez – and has threatened to publish them if the polling firms do not come forward.

One should not mistakenly conclude that these polls vindicate the anti-Chávez pollsters as “unbiased.” Rather, in the hour of truth, some pollsters – after having long engaged in highly biased polling designed to demoralize the government’s supporters and to embolden the opposition – will issue less biased polls in a last-ditch effort to salvage their own credibility in the face of impending defeat.

In early February 2003, the anti-Chávez Venezuelan polling firms Datanálisis and Consultores 21 held a joint press conference in Caracas claiming to be “neutral parties” in the country’s deeply polarized political conflict. Just over two weeks before the press conference, Narco News had highlighted that Datanálisis’ President Jose Antonio Gil Yepes had told the Los Angeles Times in July 2002 that Chávez “has to be killed.” Narco News pointed out that a simple glance at Datanálisis’ website revealed “the kind of blatant political partisanship that one normally does not associate with respectable polling operations” (as this report goes to print, Datanálisis’ website has been running John Kerry’s Chávez-bashing misstatement at the top of their “news” column for over a month).

Since Narco News first reported on Datanálisis’ blatant partisanship and biased polling, Gil Yepes has mysteriously disappeared as a public spokesperson for his company (although he occasionally pops up brandishing a letter from L.A. Times correspondent T. Christian Miller, who now supposedly claims that the pollster did not have criminal intent when he told Miller that Chávez “has to be killed”).

With Gil Yepes’ reputation in question, the job of restoring Datanálisis’ mythic neutrality was left to company director Luis Vicente León. Never mind that León had also been making blatantly anti-Chávez statements to the press long before Gil Yepes blurted out his homicidal fantasies to the L.A. Times. In Venezuela, where Chávez-bashing journalists abound, “neutrality” means telling the business-controlled propaganda apparatus what it wants to hear.

Thus, in the spirit of “neutrality,” León made a startling announcement at the conference of February 2003. Although it had long been established that Chávez enjoyed his highest levels of support among the poor, León declared that Datanálisis’ latest “poll” disproved the “myth” that public opinion was divided along class lines. According to León, “people of lower incomes” had become even more inclined to reject Chávez than the rest of Venezuelan society.

For anyone even slightly in tune with reality, León’s claim should have sent off alarm bells. Hardly more than two months before, Gil Yepes himself told the Associated Press that—while only 30 percent of the overall Venezuelan population supported the government—45 percent of the poor still approved of Chávez. Setting aside the question of whether or not Gil Yepes’ figures were based on methodologically sound polling (that issue will be taken up in the second part of this series), the figures suggested that the poor were more than twice as inclined to support Chávez as the rest of society, a finding that was consistent with past polls and election returns. Given that Venezuela’s poorest stratum (stratum E) accounts for just over 40 percent of the adult population, the only way Gil Yepes could arrive at an overall 30 percent approval rating amid 45 percent support for Chávez among the poor is if the President’s approval rating among the non-poor was close to 20 percent.

Did León actually expect people to believe that—in the course of two months—the poor had gone from being more than twice as likely to support Chávez to rejecting him at a higher rate than the middle and wealthier strata?

Puzzled by León’s claim, I decided to ask Jose Miguel Sandoval—an expert on Latin American opinion polls at the University of North Carolina at Chapel Hill—how the political views among Venezuela’s poor could undergo such a dramatic shift. Sandoval replied that reports of “drastic changes of opinion in a short period of time are not to be taken seriously, particularly in Venezuela, where opinions are well entrenched.”

Curiously, León’s dubious “finding” of the “myth” of a class divide appeared just in time for the Venezuelan opposition’s new U.S. campaign strategists to spin the same story.

As the Washington Post’s Scott Wilson reported, prominent members of the Venezuelan opposition traveled to Washington in January 2003 and began consulting informally with Democratic Party strategist James Carville. Soon thereafter, the Democratic Party polling firm Greenberg Quinlan Rosner Research (GQR)—the company of Carville’s associate and Clintonite pollster Stanley Greenberg—popped up in Venezuela working on behalf of the opposition.

In a clear demonstration of U.S. bipartisanship at the service of Uncle Sam’s reactionary foreign policy, GQR joined forces with the Republican polling firm Public Opinion Strategies to carry out “polls” on behalf of the opposition. In March 2003, GQR released a misleading statement that its findings contrasted with “the assumption of many analysts that Venezuela is divided between the upper- and middle-class opponents of Chavez and his lower-class supporters.”

The strategy was clear; in order to beat Chávez, GQR —like Datanálisis—sought to deny the government of its base.

The only problem was that GQR’s denial of a class divide and Datanálisis’ claim that the poor were now even more disapproving of Chávez than middle and wealthier strata were strongly contradicted by the actual results of the GQR-POS “poll”. The “poll” showed that the poor (stratum E) and the relatively poor (stratum D)—which together represent about 80 percent of Venezuela’s adult population—were more than twice as likely to continue supporting Chávez than the middle to wealthier strata.

As it turned out, Datanálisis’ claim that the poor had turned against Chávez with greater vehemence than middle to wealthier strata was plainly dishonest.

Between November 2002 and February 2003—the period of business-led economic sabotage against the Venezuelan government and people—Datanálisis temporarily stopped sending field workers into Chavista-controlled slums. Due to the heightening of resentment towards biased pollsters as well as increasing levels of crime resulting from the misery induced by the economic sabotage, field workers could not safely perform surveys.

In other words, León relied on telephone polls for his claim that lower-income respondents had turned strongly against Chávez (Datanálisis’ website acknowledged that its December 2002 poll regarding the opposition’s so-called “general strike” was conducted by telephone). The sociologist Greg Wilpert, who resides in Caracas, estimates that only 50 percent of Venezuelan households have mainline telephones, meaning that Datanálisis could scarcely have polled stratum E (the poor) during the period on which León based his deceitful claim.

Now, as the opposition’s campaign is clearly faltering and Venezuela’s poor appear poised to turn out en masse against the recall of President Chávez, the failure of the anti-Chávez pollsters’ underhanded attempts to deny the government of its long-standing base becomes increasingly clear.