Quantitative forms of analysis have long been lacking in studies of African affairs. Bratton et al.’s, assertion that Africa as a whole is a vastly understudied continent, and that data is “scarce, spotty or entirely non-existent”[1] was, until comparatively recently, fairly difficult to contend. Early attempts at quantitative research on the continent bear out the notion that broadly-based surveys are difficult, if not near impossible, to execute: a study of public opinion in Nigeria between 1973 and 1995, for example, could hardly claim to be representative of the entire population if it sampled university students alone.[2] The arrival of the Afrobarometer, however, has bucked that trend. According to its website, the Afrobarometer is an “independent, non-partisan research project that measures the social, political and economic atmosphere in Africa.”[3] It is renowned for having a highly meticulous, almost obsessive, set of rules governing the collection of data – the selection of households is entirely randomised and the barometer aims to cover a full cross-section of the peoples of each of the states under study. The Afrobarometer has so far conducted three rounds of exhaustive research (the first covered only 12 states; the second, 15; and the third, 18), the findings of which have been widely used by scholars to quantify – among other issues – the extent of the democratic consolidation in African states.

Yet the attempts of the Afrobarometer and other independent organisations to analyse African politics using quantitative methods have not been met with universal approval. A number of scholars have raised significant questions regarding the legitimacy of such studies. If the value of quantitative methodology in assessing democratic consolidation in Africa is to be properly appraised, it is important that we hold up the achievements of quantitative analysis to the critiques presented by qualitative scholars. This should allow us to determine how existing quantitative methods might be adjusted to create a truer picture of political realities. Ultimately, therefore, we should be able to assess not only what quantitative analysis has given, but also what it can give to the study of Africa.

Branch and Cheeseman’s study of opinion polling in Kenyan politics suggests that rather than commonly-assumed factors like ethnicity, public support for the government is, in fact, profoundly determined by its practical performance: indeed, when the NARC government was seen to have presided over a period of economic success at the end of 2004, public approval increased from 36% in October to 65% in January 2005.[4] It has been argued that the strength of such quantitative analyses is that they assess the views of normal African citizens, and thus force governments to direct policy according to those issues that are genuinely important to the electorate.[5]

Basing their study on the first round of Afrobarometer surveys in 12 states, Norris and Mattes find that “ethnicity is a significant predictor of party support in most, though not all, African societies under comparison”[6]:in states like South Africa and Nigeria with many languages, ethnic ties are found to be particularly strong; while in states like Lesotho and Botswana where ethno-linguistic groups are more homogenous, they are somewhat weaker.[7] The value of this study is that it includes states with similar economic and cultural histories, but with varying degrees of democracy and ethnic factionalism – comparative analysis is therefore properly facilitated. Qualitative theorists, however, would argue that no study can claim to be truly representative of Africa – even the most recent of the highly-regarded Afrobarometer surveys only covers 18 African states. Therefore, Afrobarometer-sponsored analyses such are this fundamentally misconceived. Such studies cannot come close to depicting the role of ethnicity across a continent of 46 countries. It is worth qualifying, however, that Norris and Mattes’ do not claim to account for the entire continent; in this case it is not the survey, but how its findings might be interpreted that is unrepresentative.

It has been argued that African respondents may be suspicious of the purpose of any quantitative survey; that they might see it as some government-sponsored scheme, and might censor their answers accordingly.[8] The Afrobarometer makes a very specific effort to avoid this, with its interviewers emphasising from the very outset that they represent an independent organisation and not the government, and ensuring that participation is entirely voluntary.[9] Another key weaknesses of the quantitative approach is that that geographically isolated or less well-educated respondents may have little interest in some of the questions raised, and may simply answer reflexively. Bratton et al. have suggested that this sort of discrepancy might be corrected by framing the question differently: rather than asking (for example) “do you support market reforms?”, interviewers should ask: “do you prefer to have low prices in the market even if there are shortages of goods, or do to have plentiful goods in the market even if the prices are high?”.[10] The Afrobarometer avoids asking such direct economic questions: instead of asking about market reforms, it lists various ways of managing the economy and asks respondents to rate each method.[11] In this case, it can be said that the Afrobarometer successfully engages respondents in the issues under study.

In terms of understanding the issue of democratic consolidation, the Afrobarometer takes a broad approach, asking its respondents a whole series of political questions before approaching the question of “satisfaction with democracy.”[12] In this sense, however, the Afrobarometer leaves itself open to criticism. The “Satisfaction With Democracy” question has been shown to be subject to too many variables to remain an appropriate sphere of quantitative analysis.[13] People’s interpretation of the question depend on the society under study: an Eastern European respondent would generally define democracy in different terms to those used by a Western European. Moreover, people’s opinions are never a fixed entity: they inevitably change over time and according to the performance of the government.[14] Thus, it can be said that in this case, a Tanzianian respondent is answering an entirely different question to a South African respondent. Conache et al. put it in more explicit terms:

It makes as much sense to compare satisfaction with democracy in Panama, Paraguay and Peru as it does to compare the weather in France with the cost of living in Austria with employment opportunities in Belgium.[15]

Some studies have gone as far as to describe public opinion research, conceived as it is in industrialised states, as a globalised brand.[16] Like any global brand, it is argued, quantitative methodology is dangerously unrepresentative of indigenous realities. The Freedom House Freedom in the World annual surveys, for example, categorise 193 of the world’s countries and 15 disputed territories as either “free”, “partially free” or “not free”.[17] According to the Freedom House website, the survey is based upon the assumptionthat “freedom for all is best achieved in liberal, democratic societies.”[18] Clearly this is far too narrow a categorisation; Freedom House’s singular definition of freedom is, without doubt, not applicable in each of the states it assesses. By universalising its notion of freedom, Freedom House undermines it own conclusions.

Quantitative methodology offers a number of unique insights to the study of African politics: crucially, it allows us to evaluate the opinions of the normal people who are so often isolated from world politics. Moreover, a quantitative approach also facilitates comparative analysis. Howard White argues that “Even without aspiring to a Nobel Prize, economists can in their day-to-day work learn much more from drawing on other disciplines.”[19] Here lies the crucial lesson regarding the study of democratic consolidation in African states: quantitative forms of analysis undoubtedly have plenty to offer, but only when their methods and their conclusions are properly tempered and qualified by qualitative knowledge.

Surveys like the Afrobarometer are particularly valuable in that they make it clear that they are not affiliated with the government, and thus are able to discourage participants from censoring their comments. The Afrobarometer is also careful to avoid academic style jargon: where respondents are likely to have a lack of interest or educational grounding in the issues raised, the questions are couched in user-friendly terms – participants are thus engaged in the survey. The shortcomings of quantitative methodology in Africa – such as its failure to represent the entire continent; its inappropriate use of the Satisfaction With Democracy question, with its meaning highly dependent on personal interpretation; and its universalised approach, which undermines indigenous peculiarities – also involve its failure to address qualitative concerns. It thus appears that David Laitin’s tripartite methodology, whereby scholars from opposing disciplines gather cross-sectional data, apply it to a theoretical framework, and test it against practical examples,[20] is a particularly useful means of adjusting existing methods for the study of democratic consolidation in African states. This will ensure that every study of democratic consolidation is conditional, and subject to the evaluations and the insights of other disciplines. Put simply, quantitative study can only be improved by an understanding of qualitative theory.

Bibliography

Afrobarometer website: http://www.afrobarometer.org/.

Branch, Daniel and Nic Cheeseman, “Using Opinion Polls to Evaluate Kenyan Politics, March 2004-January 2005”, African Affairs 104, (415) 2005.

Bratton, M., R.B Mattes and E. Gyimah-Boadi Public Opinion, Democracy and Market Reform in Africa, New York: Cambridge University Press, 2004.

Canache, D, J.J Mondak and M.A. Seligson, “Meaning and Measurement in Cross-National Research on Satisfaction With Democracy”, Public Opinion Quarterly 65 (4), 2001.

Freedom House Website: http://www.freedomhouse.org/.

Heath, A., S. Fisher and S. Smith, “The Globalization of Public Opinion Research”, Annual Review of Political Science, 8, 2005.

See Laitin, D., “Comparative Politics: The State of the Sub-Discipline”, in Katznelson, I. and H.V. Milner (eds.) Political Science: The State of the Discipline (W.W. Norton). Available at http://www.standford.edu/~dlaitin/papers/Cpapsa.doc.

Norris, P. and R. Mattes, Does Ethnicity Determine Support for the Ruling Party? Afrobarometer Working Paper 26. Available at: http://www.afrobarometer.org/abseries.html.

White, H., “Combining Quantitative and Qualitative Approaches in Poverty Analysis”, World Development 30 (3), 2002.

—

Written by: Cosanna Preston
Written at: Oxford University
Written for: Dr D. Pratten
Date written: 2007

Further Reading on E-International Relations

• To What Extent Can History Be Used to Predict the Future in Colombia?
• To What Extent Can Natural Disasters Be Considered State Crimes?
• Neopatrimonialism and Democratic Consolidation in Nigeria
• To What Extent Has China’s Security Policy Evolved in Sub-saharan Africa?
• To What Extent Is ‘Great Power Competition’ A Threat to Global Security?
• To What Extent Was the NATO Intervention in Libya a Humanitarian Intervention?