Cuban Schools: Too Good to Be True (Unabridged)

It is widely accepted that Cuban schools have made great strides forward. “Cuba’s education system might as well be considered the ultimate wrap-around institution for children,” claims the executive director of the American School Superintendents Association. [1] A Stanford scholar writes in the HuffPost that he has “a hunch” that Cuban schools are better than those in the United States. President Barack Obama and Vermont Senator Bernie Sanders are equally celebratory. [2]

What is the evidence to warrant this enthusiasm? Cuba does not participate in major international tests of educational achievement. The country did participate in the 1997 and 2006 waves of Laboratorio, a UNESCO-sponsored survey of Latin American elementary school students, but it dropped out of the third wave administered in 2013. Further, the Cuban results from the 1997 and 2006 waves of this survey lack credibility, as we shall see.

The Literacy Campaign

One thing is certain. Education—at least of a certain kind—is central to Cuban Communism. “Revolution and education are the same thing,” said Fidel Castro, the island’s revolutionary hero. [3] “To build communism, a new man must be created. . . . Society as a whole must become a huge school,” wrote Castro’s philosopher-in-residence, Ernesto “Che” Guevara. [4] Immediately following its 1960 revolution, Cuba embarked on a campaign to eradicate illiteracy. “Over a quarter of a million” alphabetizadors or literacy teachers were sent from schools into rural areas for “extended periods away from home” to live with the “campesinos and others whom they taught,” Samuel Bowles, a Marxist economist, writes approvingly in a 1971 article in the Harvard Educational Review. “Over 100,000 students joined the campaign when schools were closed for the year on April 15, and almost all the professional teachers in the country participated.” In an effort similar to China’s Great Leap Forward, the staff of “entire schools [were moved] to the countryside for extended periods to harvest crops and do other agriculture work,” Bowles writes. Teachers and students were “housed in simple camps and doing hard agricultural work side by side with the campesinos.”[5]

A less sanguine account, by H. S. Bola, conveys the militaristic energy of the operation:

student workers were organized into “brigades,” wore uniforms and took oaths, and “liberated” villages from illiteracy. The title of the primer, Venceremos, which means “We will overcome” or “We will conquer,” reflects a military tone, although it is well understood that the enemy in this case is illiteracy. A section of the hymn sung by brigadistas in the countryside, however, includes reference to yet another enemy: “Down with imperialism, up with liberty! We carry with the words the light of truth.”[6]

One million four hundred thousand Cubans fled their homeland for the United States in the wake of draconian measures taken to restructure Cuban society. [7] The cost to the Cuban economy is well known. But what has been the long-term educational impact of Cuba’s broad jump forward? Could it be a model for school reform in the United States?

That topic edged into the 2020 presidential race when Bernie Sanders, in a 60 Minutes interview, gave Cuba’s educational innovations high marks. Castro “educated their kids,” by means of “a massive literacy program,” he said admiringly. [8] Asked to defend the assertion, Sanders cited President Barack Obama’s own assessment of Cuban education reforms: “You’ve made great progress in educating young people. Every child in Cuba gets a basic education,” Obama says he told Castro in a conversation in which the president asked him to embrace a market economy. [9]

The Impact: How is it measured?

Sanders and Obama are correct that Cuba launched a campaign to reduce illiteracy in rural communities, and it would be inaccurate to deny that progress has been made. Even there, the gains may well be overstated, as the literacy rate in Cuba had reached 78 percent prior to Castro’s revolution. [10] But celebratory claims by left-wing academics and liberal media outlets have left the impression that Cuba, alone among Latin American countries, has created a high-quality educational system—and that its “great progress in educating young people” stands in contrast to the dismal performance of American schools. Martin Carnoy, a professor at Stanford’s School of Education, wrote in 2011 that “Cuban education may be better, on average, than [the education available in] American” schools. [11]

Cuba has resisted invitations to subject its claims to external verification. Notably, it has declined to participate in the Program for International Student Assessment, sponsored by the Paris-based Organization for Economic Co-operation and Development, which every three years since 2000 has administered achievement tests in math, science, and reading to 15-year-old students in over 70 countries. PISA tests are administered to students in Russia, China, Vietnam, Argentina, Brazil, Chile, Columbia, Mexico, the United States, and Uruguay. The PISA test has revealed the woeful performance of the United States as compared to Finland, Germany, Canada, and elsewhere. Cuba could easily establish the fact that its students outrank the United States and its sister countries in Latin America simply by participating. But Cuba has never administered PISA to a representative sample of its students.

Cuba did participate in the 1997 and 2006 waves (but not the 2013 wave) of a survey of elementary-student achievement known as the UNESCO Regional Comparative and Explanatory Study, which has been administered by Laboratorio, the moniker used here, to multiple countries in Latin America. [12] Results from these tests seem to show that Cuba outranks the rest of Latin America by wide margins. Carnoy, author of a Stanford University Press book titled Cuba’s Academic Advantage, relies on these tests when claiming that Cuba outperforms the United States.

UNESCO has given Laboratorio responsibility for design of the survey. The agency constructs appropriate questions for students at particular grade levels, and, beginning with the second survey, uses standard techniques to ensure that test results are comparable from one survey to the next. The testing organization also asks each country to include in the sample a representative number of schools by urbanicity (urban vs. rural), grade composition (primary, middle school, combined) and sector (public vs. private). However, the actual selection of schools and administration of the tests is left to the coordinating agency within each country. As a consequence, the Cuban administration of Laboratorio in the 1997 and 2006 waves of the survey was the responsibility of the country’s central government.

The 1997 wave was administered to 13 countries, a number that grew to 14 in 2006. When the third Laboratorio survey was administered in 2013, Honduras also agreed to participate, but the number remained at 14 when Cuba withdrew. Carnoy says that 7 of these 14 countries also participated in PISA’s 2006 wave. [13] He says average national performances in these countries on the 2006 Laboratorio (taken by 6th graders) are correlated with average performances on the 2006 PISA (taken by 15-year-olds). [14] That, he says, permits a statistical operation that yields an estimation of Cuba’s performance on PISA. When Carnoy finishes his calculations, he discovers, lo and behold, that Cuba would have outperformed the United States had it participated in the PISA survey. To reach such a conclusion, Carnoy makes three heroic assumptions: PISA and Laboratorio tests are comparable, student performances at age 15 can be predicted by performances in 6th grade, and tests can be linked by constructing correlations based upon a few nationwide observations.

As problematic as the assumptions are, they are not the main reason for lifting eyebrows when told the size of the “Cuba’s academic advantage.” The main concern is the credibility of the Cuban testing results themselves. There is no direct evidence of cheating, it must be said. But the following peculiarities raise concerns that would likely prompt official investigation had they been observed in the United States: 1) Performance levels are incredibly high, 2) gains from one grade to the next are miniscule, 3) socioeconomic gaps in student achievement are unbelievably tiny; 4) teachers report extraordinarily high homework assignment rates and low incidences of disciplinary problems, and 5) Cuba withdrew from participation in 2013, despite its number-one ranking in earlier surveys. Countries that win gold medals don’t typically withdraw from subsequent competitions without good reason.

Any one of these outside-the-box outcomes may have an alternative explanation, but together they point toward one conclusion: the Cuban central government had a strong incentive to demonstrate their students out-performed the rest of Latin America—and it very probably took steps to make sure that happened.

Sampling

When outcomes seem unlikely, the first suspect is the design of the sampling frame. To obtain a nationally representative sample, a survey must give all students in the country an equal probability of being chosen to participate. If U.S. performance on the PISA were ascertained by gathering information only from schools in rich suburbs, estimates would exaggerate nationwide achievement levels. Conversely, if tests were administered only in schools located in central cities, estimates of average national performance would be skewed downward.

Laboratorio left sampling decisions, apart from the guidelines mentioned above, to the discretion of the countries administering the test. Carnoy and his colleague, Jeffrey Marshall, say “it is reasonable to ask whether the very high test scores in Cuba are the result of picking a select group of schools.” [15] But they don’t think that happened because “our own classroom observations in 10 schools . . . suggested to us major differences in the level of performance of Cuban third graders compared with those in Brazilian and Chilean schools.” [16] But the authors provide no evidence that they were allowed to visit representative schools rather than Potemkin villages. So it is entirely possible that the results for Cuba are simply due to biases in the sampling frame. Other oddities, though, suggest that more than sample design impacted the Cuban results.

Beyond Superior Performance

Consider, for example, Cuba’s achievement results in language arts in Laboratorio’s 1997 wave. The median score in language arts for Cuban 3rd graders was 343 points, as compared to 264 points in Argentina, 256 points in Brazil, 259 points in Chile and 229 in Mexico, differences that range from 1.6 to 2.4 standard deviations. (A standard deviation on these tests appears to be roughly two or more years’ worth of learning.) [17] If these scores are to be believed, the median child in Cuba learns by 3rd grade what the median student in other countries learns only by 6th grade or later. This difference is so large that the distributions of achievement in these three countries barely overlap that of Cuba. The score of a student at the 25th percentile in Cuba is 305 points, while the scores of students at the 75th percentile in Argentina, Chile, and Brazil are only 305 points, 304 points and 283 points, respectively. [18]

Cuba’s eye-popping performance was not limited to language arts. In 2006 the median 3rd-grade student in Cuba performed on the math achievement test at about 1.5 standard deviations higher than the median students in Argentina, Brazil, and Chile. The same is true for the 6th-grade test. These results have been interpreted as showing an astonishing Cuban educational advantage, but they might also be interpreted as “too good to be true.” After all, Chile performed only 0.9 standard deviations lower than high-flying Finland on the 2018 math test administered by PISA. [19]

Little Value-Added

Despite the fantastic results at each grade level, the Cuban students do not seem to learn much from one grade to the next. In 1997 Laboratorio tested students in both 3rd and 4th grade, which allows one to track how much students, on average, gain over the course of a single year. In Argentina, Brazil, Chile, and Mexico, 4th graders score 22 to 25 points higher than third graders, indicating learning gains of about a half a standard deviation in the course of one year. But the students attending Cuba’s marvelous schools gained only 5 points, not enough to achieve statistical significance. The oddity is of such magnitude that Carnoy and Marshall feel required to comment, if only in a footnote, as follows:

One of the mysteries of the Cuban results is the small difference between third-and fourth-grade test scores (on the same test but different students taking the test). One possible explanation is that the test was sufficiently easy for Cuban students that a high fraction of both third and fourth graders achieved perfect scores, so that it was difficult to achieve much higher average scores in the fourth grade. [20]

An alternative explanation for this unusual phenomenon is that teachers were correcting the answers so that many students, in both 3rd and 4th grades, were obtaining perfect or near-perfect scores. As a consequence, 3rd graders appear to be doing as well as those with another year of schooling.

Eliminating the Achievement Gap

Cuba has virtually eradicated the socioeconomic-status-achievement gap, if Laboratorio results are to be believed. In Latin American as a whole, that gap is very large. According to a report of the Inter-American Development Bank, Latin American students participating in the 2006 wave of Laboratorio who were from households in the bottom 20 percent of the socioeconomic distribution had only a 10 percent probability of scoring at a satisfactory level on the 3rd grade math test, while students from households in the highest quintile had a 48 percent probability. In Brazil, these probabilities were 12 percent and 59 percent, respectively. But in Cuba, the probabilities were essentially the same—72 percent and 74 percent—for students from households in the lowest and highest quintiles of the distribution. [21] For 6th grade students taking the 2006 math test, these probabilities were 76 percent and 81 percent, respectively. [22]

The socialist paradise has also virtually eliminated the urban-rural gap, which is otherwise quite massive throughout Latin America. In Mexico and Brazil that gap in language arts is 0.62 and 0.66 standard deviations, respectively, and in Argentina and Chile, it is 0.35 standard deviations. But in Cuba that number falls to just 0.16 standard deviations. In math it is just 0.05 standard deviations, a difference that is not statistically significant. Cuba has indeed lived up to its egalitarian commitments—if it has not falsified its scores to give that impression. [23]

Stakhanovite teachers and obedient pupils

In 1935, Stalin honored Aleksei Grigorievich Stakhanov for mining 102 tons of coal in less than six hours, 14 times his quota. His followers, the Stakhanovites, tried to do likewise, and it is this kind of heroism that Cuban teachers apparently feel they need to report. When responding to a survey, the vast majority say they always assign their students homework. Elsewhere in Latin America, only a minority of teachers say they always assign homework. Fewer than 30 percent of the 3rd-grade and 6th-grade math teachers in Mexico, Argentina and Brazil, and no more than 10 percent of the Chilean ones, said they always assign math homework. By comparison, 90 percent of 3rd-grade and 6th-grade math teachers in Cuba insisted they always assign homework. In the language arts, these percentages were 87 percent and 84 percent for the two grades, respectively. [24] Homework is not popular in progressive circles in the United States, but it appears to be nearly pervasive in Cuba—or at least teachers feel compelled to claim that that is the case.

Then, too, elementary students “hardly ever” fight in Cuban classrooms, teachers say. The average “classroom fight” score on the Cuban teacher survey runs a full standard deviation below that for other Latin American countries. [25] Just as test scores are incredibly high, reports of classroom fights are dubiously low. Perhaps elementary students in Cuba are model socialist citizens, but if they are not, teacher reports under-state the factual situation on the ground, perhaps because accurate statistics are not desired by the authorities. One can only wonder about the potential consequences for teachers had they reported that their students misbehaved or did not do well on the Laboratorio tests. We do know that at least one teacher, Roberto de Miranda, was fired from his teaching position for “refusing to pass students [whom he thought] did not earn passing grades” and was later imprisoned for his political activities. [26]

Conclusions

Given Fidel Castro’s commitment to state socialism, one can hardly fault his cheating. To deceive credulous sympathizers is in the national interest, as understood from his point of view. After all, education was central to the original mission of the Cuban revolution, thousands of Cubans were uprooted ostensibly to eliminate illiteracy and equalize opportunity in rural Cuba, and a sizeable share of the country’s scarce resources are committed to primary education.

Nor should Laboratorio be faulted for launching an imperfect survey of Latin American countries which hitherto had not participated in international testing. Results from international tests can have serious political consequences. Germany was forced to re-examine its school system in 2000 when its students ranked well below those in Finland and the Netherlands. India withdrew from PISA after 2009 when results placed the country near the bottom of international rankings. The United States is embarrassed by its low math rankings each time the PISA tests are announced. When asking a country to participate in an international test for the first time, an international agency needs to be sensitive to local political circumstances, and Laboratorio was in no position to drive a hard bargain with individual countries when first attempting to construct the survey. Even PISA officials may be more lenient with countries participating in its survey for the first time. Allowing nations to draw samples and administer the tests themselves was the only option for Laboratorio.

One need not be as generous with scholars who have the responsibility to expose sham and pretense whenever and wherever it is observed. The Laboratorio data are open and available for any scholar to analyze. Yet the alleged Cuban educational advantage flogged by left-wing professors has never been subjected to the kind of rigorous scrutiny applied to impressive test-scores reported by schools in the United States. When the academic community fails to exercise its responsibilities, political leaders are not constrained from making unwarranted claims based on flimsy evidence. Under the circumstances, Bernie Sanders must be complimented for exercising restraint when he said Cuba had made progress toward ending literacy. That statement does not say much, but at least it is true.

This is an unabridged version of an article that also appears in a more streamlined version here.

Paul E. Peterson is the Henry Lee Shattuck Professor of Government and Director of the Program on Education Policy and Governance at Harvard University, a Senior Fellow at the Hoover Institution at Stanford University, and Senior Editor of Education Next.

Notes:

[1] Dan Domenech, “Cuba’s Education System: High literacy Rates-Free College Come at a Price,” Learning First Alliance Blog. May 3, 2017.

[2] Martin Carnoy, “Are Cuba’s Schools Better than Ours?” Huffington Post, June 26, 2008; updated May 25, 2011.

https://www.huffpost.com/entry/are-cubas-schools-better_b_109280. Accessed May 4, 2020.

[3] “Universidad Popular,” Educacion y Revolucion, 6^th series (Havana: Imprenta Nacional de Cuba, 1961, p. 271, as quoted in Samuel Bowles, “Cuban Education and the Revolutionary Ideology,” Harvard Educational Review, 1971, Vol 41(4), p. 472.

[4] “Man and Socialism in Cuba,” in Venceremos! The Speeches and Writings of Che Guevara, John Gerassi, ed. (New York: Simon and Schuster, 1968, p. 391, as quoted in Bowles, p. 472.

[5] Quotations in this paragraph are from Bowles, p. 488.

[6] H. S. Bhola, Campaigning for Literacy. Paris: UNESCO, 1984, p. 96, as quoted in Ruth A. Supko, “Perspectives on the Cuban National Literacy Campaign,” Paper presented before the 1998 meetings of the Latin American Studies Association, September 24-26, 1988, p. 12.

[7] Jorge Duany, “Cuban Migration: A Postrevolution Exodus Ebbs and Flows,” Migration Information Source, July 2017. https://www.migrationpolicy.org/article/cuban-migration-postrevolution-exodus-ebbs-and-flows

[8] Ian Schwartz, “Bernie Sanders ’60 Minutes’ Interview: “Unfair To Say Everything Is Bad About Cuban Revolution, Castro” February 24, 2020. https://www.realclearpolitics.com/video/2020/02/24/bernie_sanders_60_minutes_interview

[9] Usina Del Arte, “Remarks by President Obama in Young Leaders of the Americas Initiative Town Hall,” The White House, Office of the Vice President, March 23, 2016. https://obamawhitehouse.archives.gov/the-press-office/2016/03/23/remarks-president-obama-you

[10] Norman Luxenburg, “A Look at Castro’s Statistics,” Encounter, March 1984, pp. 58-61. Also, see https://www.nationalreview.com/2020/03/bernie-sanders-wrong-cuba-literacy-program/

[11] Carnoy, 2008, updated 2011. Also, see Martin Carnoy and Jeffrey Marshall, “Cuba’s Academic Performance in Comparative Perspective,” Comparative Education Review, pp. 472-500. 2005. 49: 2, pp. 472-500. This article is reprinted in Martin Carnoy, Cuba’s Academic Advantage (Stanford, 2007.)

[12] “Student Achievement in Latin American and the Caribbean: Results of the Second Regional Comparative and Explanatory Study (SERCE), 2008. Laboratorio Latinoamericano de Evaluacion de la Calidad de la Education (LLECE); Alejandro J. Ganimian, Are Latin American Children’s Reading Skills Improving? Highlights from the Second and Third Regional Comparative and Explanatory Studies (SERCE& TERCE). Laboratorio Latinoamericano de Evaluacion de la Calidad de la Education (LLECE), Working paper, 2016. No. 6. Also see Carnoy and Marshall, 2005.

[13] According to the World Bank, the number is six—Argentina, Brazil, Chile, Columbia, Mexico and Uruguay. Peru participated in 2000 but not in 2006. Emanuela Di Gropello, Maria Jose Vargas, and Monica Yanez-Pagans, “What are the main lessons from the latest results from PISA 2018 for Latin America?” World Bank Blogs, December 6, 2019. https://blogs.worldbank.org/latinamerica/what-are-the-main-results-pisa-2018-latin-america. Accessed March 9, 2020.

[14] Carnoy, 2008, update 2011. He does not report in this blog post either the size of the correlation coefficient(s) or the statistical technique he used. We have been unable to find documentation in any other online or print publication.

[15] Carnoy and Marshall, 2005, p. 238.

[16] Carnoy and Marshall, p. 239, note 32.

[17] The tests administered in the 1997 wave have a median of 250 points and a standard deviation of 50 points. The tests administered in the 2006 wave was have a median of 500 with a standard deviation of 100. Students in Mexico, Chile, Brazil and Argentina all performed about 0.5 standard deviations higher on the 1997 wave of Laboratorio in fourth grade than in third grade when co-variates are controlled. Carnoy and Marshall, 2005, Table 2, p. 248.

[18] Lavinia Gasperini, The Cuban Education System: Lessons and Dilemmas. Country Studies: Education Reform and Management Publication Series, Vol. 1, No. 5. World Bank, Washington, DC. Human Development Network. Annex I, p. 23.

[19] “PISA 2018 Results,” Program for International Student Assessment, Organization for Economic Co-opertion and Development, 2019. https://www.oecd.org/pisa/publications/pisa-2018-results.htm

[20] Carnoy and Marshall, 2005, p. 245, note 16. The fourth-grade value-added results, controlling for co-variates, are to be found in Table 2, p. 248.

[21] Jesus Duarte, Maria Soledad Bos, and Martin Merano, Inequity in Student Achievement in Latin America: Multi-level Analysis of SERCE Results According to the Socioeconomic Status of Students, Inter-American Development Bank, 2010, Table 2, p. 13.

[22] Duarte, Bos and Merano, 2010, Appendix 2, p. 48.

[23] Duarte, Bos and Merano, 2010, Appendix 2, p. 48.

[24] F. Javier Murillo and Cynthia Martinez-Garrido, “Homework and Primary-school Students’ Academic Achievement in Latin America,” International Review of Education (2014) 60, pp. 670-71, Tables 2 and 3.

[25] Carnoy and Marshall, 2005, Online Appendix Table A1-A3. Although this Online Appendix is cited in this source, it is not now available online. Martin Carnoy generously shared these tables with me by private email and these tables are in the Appendix to this essay.

[26] George W. Bush Institute, “Interviews : Roberto de Miranda,” The Freedom Collection, 2011. http://www.freedomcollection.org/interviews/roberto_de_miranda/

TABLE A1

Definitions for all Variables Included in Analyses

TABLE A2

Means and Standard Deviations (in parentheses) for Variables Included in Analysis

TABLE A3

Means and Standard Deviations (in parentheses) for Variables Included in Fourth Grade Only Analysis