I must apologize for not being able to be with you at this conference, especially because the subject of this particular panel interests me so much. The paper which I have read presents a significant methodological advance in the study of what I will call, for want of a better term “captive populations.” As many of you know, scholars have long been frustrated in their efforts to study populations whose governments place them beyond the reach of our efforts to study them. The Cuban case has been especially frustrating. On a number of occasions when I directed the Canter for Latin American Studies at the University of Pittsburgh, we had delegations of Cuban scholars visit us for conferences and consultations regarding our long-term exchanges, established by Carmelo Mesa-Lago. Invariably, we would hold a discussion about the prospects of conducing a survey of public opinion in Cuba. I would argue for a participatory process, so that Cuban scholars would be engaged in every phase, from the questionnaire and sample de- sign up through the analysis of the data. Yet, in each case, the idea was stillborn. Whenever I asked for assurances that the survey would be done as an academic exercise, free of government control and censorship, I was met with silence. And silence meant “no.” In the end, the studies were never carried out.
The present effort confronts all of the difficulties faced by those who have tried to use emigre data in the past. We know that emigre data about the former Soviet Union was based on a biased sample; presumably the dissidents who left or escaped, were different in a wide variety of ways from those who did not. In prior studies, little or no systematic effort was made to control for those differences. In this paper, an heroic effort is made, and very successfully, up to a point, to control for this sample bias. The authors use the known demographic and socio-economic parameters of the Cuban population to adjust the sample of emigres so that it more accurately mirrors the population of those who have not (yet?) left the country. The adjustment methodology used seems entirely sound to me and produces results which have little demographic or socioeconomic bias.
Unfortunately, the researchers are unable to control for the single most important bias in a sample of public opinion, and that is bias in the opinions them- selves. We have to assume that those who have left Cuba, whether on rafts or via visas or third countries, hold views different from those who have not left. Some of those differences of opinion are, of course, reflected in demographic and SES parameters. But most are not. How can we explain why one 25 year- old college educated female leaves the island, while another with the same age and education stays be- hind? We have to assume that the two vary in their opinions about the regime, which is, after all, what we are very interested in measuring in the first place. So, it is fair to conclude that the data presented here are about as good as one can get, but cannot be said to reflect the opinions of Cubans who reside on the island.
Let us look at some of the results. On the first question analyzed, namely, humanitarian aid, 89 percent of the survey respondents said that they had not benefited from it. Now it may be that this figure accurately reflects the parameters of Cubans still on the island, but one suspects that it does not. The second questions concerns the locating the blame for the lack of goods in Cuba. Few respondents blame the United States, but, as the authors note, this could well be a function of the respondents’ need to ingratiate themselves to their new hosts in their now home. They explain that such a bias is unlikely, whereas I suspect that it is very likely. We know a great deal about what is called “social desirability response set,” and in this case there are real pragmatic reasons for emigres to avoid insulting the country that one day they hope to be citizens of.
In another portion of the analysis, the authors ex- press surprise that “only 40 percent of the population had seen the protest film Guantanamera.” I find that a very high number. I suspect that a blockbuster like Titanic reached that large a proportion of the U.S. population, but most films, even the most popular, have been seen by a far smaller proportion of the population. Many people just don’t like going to the movies.
Perhaps most telling in the sample bias are the results of the evaluation of political leaders. Fidel Castro himself ends up with a negative rating of 93. It is simply not possible to believe that among those on the island this figure could be anywhere nearly accurate and still have the system survive. I have never seen public opinion evaluations that low for any extant political leader. After all, there is considerable evidence that Stalin himself, responsible for so many millions of state-sponsored murders of civilians during his long rule of the Soviet Union, was (and re- mains) highly popular in certain circles. Therefore, these numbers almost certainly overstate the views of Cubans on Castro and reveal that the sample as a whole does not reflect the political views of most Cu- bans. Sadly, what we don’t know is how far off those results are. If we could measure it, would Castro end up with a -80 or a -50 or a -20? We just don’t know. We could have a clearer idea, perhaps, if the respondents had also been given some questions on other figures, like Vesco or Noriega. Then we might be able to anchor their views and control for some of those biases.
In sum, this is the first work I have seen using emigree data. It tells us a great deal about the emigres, and something about Cubans still on the island. But, until such time as the regime allows free and unfettered social science research on the island itself, we will not know from public opinion surveys what Cu- bans are thinking about politics.