Conflict and Education - Data analysis on the public s views on education in conflict-affected countries

Transcription

1 2011/ED/EFA/MRT/PI/31 Background paper prepared for the Education for All Global Monitoring Report 2011 The hidden crisis: Armed conflict and education Conflict and Education - Data analysis on the public s views on education in conflict-affected countries Pia Horvat 2010 This paper was commissioned by the Education for All Global Monitoring Report as background information to assist in drafting the 2011 report. It has not been edited by the team. The views and opinions expressed in this paper are those of the author(s) and should not be attributed to the EFA Global Monitoring Report or to UNESCO. The papers can be cited with the following reference: Paper commissioned for the EFA Global Monitoring Report 2011, The hidden crisis: Armed conflict and education. For further information, please contact efareport@unesco.org.

2 Commissioned by UNESCO Conflict and Education Data analysis on the public s views on education in conflict affected countries Pia Horvat 6/21/2010 This report considers public opinion on the importance of education in conflict-affected and fragile countries. 1

3 Contents Executive summary... 4 Introduction... 5 The connections between conflict and education... 5 Methodology... 6 Data sets... 6 Countries... 6 Key variables... 7 The most important problem... 7 Government performance... 7 Views of aid donors... 7 Socio-demographic characteristics... 7 Analysis... 7 Main findings... 8 Africa Liberia Results from opinion surveys Secondary Sources Nigeria Results from opinion surveys Niger Delta Ethiopia Results from opinion surveys Uganda Results from opinion surveys North Uganda Asia Afghanistan Results from opinion surveys Secondary sources Pakistan Results from opinion surveys Secondary sources Pakistanis views of foreign aid Kashmir (Pakistan administered provinces) India Secondary sources Kashmir (India-administered provinces) Sri Lanka Secondary Sources Polls of Northern Tamil and Sinhalese communities Indonesia Results from opinion surveys Thailand Results from opinion surveys The Philippines Results from opinion surveys The Middle East

4 Palestinian Territories Results from opinion surveys Iraq Secondary sources Latin America Colombia Results from opinion surveys Haiti Secondary sources Europe Georgia Results from opinion surveys Serbia Results from opinion surveys Turkey Results from opinion surveys References Data References Appendix Conflict-affected countries List of Surveys Regional Global World Values Survey Comparative Results

5 Executive summary Unemployment, poverty and management of the economy are consistently identified as the biggest problems at the national level by citizens of conflict-affected and fragile countries. When citizens are asked to identify the most important problem or concern at the national level in single response questions education is rarely seen among the most pressing problems, which the government should prioritise in its policy responses. While in conflict-affected and fragile countries citizens may not identify education as the most immediate concern, this does not mean that education is seen as unimportant. In multipleresponse questions education is consistently mentioned by a significant proportion of respondents in their subsequent responses. As a result the total number of responses identifying education as one of the most important problems at the national level is usually great. In closedended questions education is usually identified as the most important problem by a significant proportion of citizens in conflict-affected and fragile counties, although the proportions mentioning poverty and unemployment or economy are still greater. Compared to other sectors and areas of concern citizens in conflict-affected countries feel that the government s performance in addressing education needs is good. In some countries this also corresponds to high absolute levels of satisfaction with the government s handling of the education sector, while in other countries absolute levels of satisfaction are low. Citizens in conflict-affected countries show the greatest dissatisfaction with the government s handling of the economy. Compared to other sectors and public services managed or provided by governments, the public is dissatisfied with the national governments overall management of the economy and/or handling of particular economic policies, such as policies pertaining to inflation, unemployment and income inequality. There is some indication that education is prioritised more by citizens living in the poorest countries and countries with severe and widespread conflict (often the two coincide). The results from analysis of conflict-affected regions within countries are inconclusive. Only in a subsample of countries was the number of respondents from conflict-affected regions large enough to allow for a disaggregated analysis. Citizens views of education in conflict-affected regions within countries are not significantly different from the national average in a systematic way. Citizens in conflict-affected countries are relatively confident in non-governmental, charitable and humanitarian organisations and mostly satisfied with the influence these organisations have in their respective countries. 4

6 Introduction The study of education and conflict represents a relatively neglected and under-researched field and yet inadequate educational provision in conflict-affected and fragile countries remains one of the most important obstacles to achieving the Dakar goal of universal primary education by 2015 (Tomlinson and Benefield, 2005). However, recently the field of conflict and education has been the subject of increasing attention by policy-makers, in the media and the academic community (Novelli and Lopes-Cardozo, 2008). Scholarly work in the field has often focused on defining the research agenda (e.g. Novelli and Lopes-Cardozo, 2008), while empirical efforts have been more limited. In particular, the public s views on education in conflict-affected and fragile countries have only rarely been the subject of explicit attention and systematic examination. This report tries to fill some of the empirical gap by analysing the public s views on education in a large number of countries, which have recently experienced violent conflict. This work addresses six main research questions: What do citizens in conflict-affected countries see as the biggest challenge or problem for the government, and how does education rank within this? What priority are citizens giving to education in conflict affected countries compared to other sectors or areas of public concern? How satisfied is the population with the government s delivery of education services in conflict affected countries compared to other sectors or areas of public concern? What does the population in these countries see as being the government s the greatest challenge within the education sector? How satisfied is the population with the role of aid donors in supporting development activities in conflict-affected countries? Can differences be identified in relation to their satisfaction of their role in supporting education compared with other areas of activity? In the above areas, can any significant differences be noted between the public s view in conflict and non-conflict affected areas of these countries? In order to answer these questions large-scale public opinion surveys conducted in conflictaffected and fragile countries since 2005 are analysed. The connections between conflict and education The connections between conflict and education are complex, multifaceted and not yet fully understood. Tomlinson and Benefield (2005) identify the following recurrent themes: First, providing education in the time of conflict is a huge challenge. Schools and educational personnel are targets of violence, and infrastructure is often destroyed. Schooling forgone by youth during conflict years results in a vulnerable society, impeding successful recovery and reconstruction. Investment in education declines in times of conflict as other areas and sectors are given priority. Fear and disruption lead to poor learning outcomes. Inadequately paid teachers may be more prone to corruption. Emergency donors may be reluctant to fund education, if it is considered a development issue. There is a risk that provision of education to refugees will be of a higher standard than that for host populations, leading to tensions between groups. The legitimacy of governments and of their curricula is questioned when in civil war. International agencies may unintentionally undermine the capacity of governments to provide education, for example, by offering high salaries which may attract former education professionals away from education sector. People become focussed on survival and education is not prioritised. Disarming children where they had been given arms is a difficult and uncertain process. Schools in conflict areas may become disconnected from the school system. Monetary and time costs of providing education to hundreds of thousands of children during a 5

7 humanitarian emergency are very high. Primary education becomes the focus, while secondary and adult education are drained of resources. Agencies may not be willing to pay teachers salaries, as this may be considered the responsibility of governments. Second, ensuring adequate educational provision in the time of conflict has important benefits for individuals and societies, especially in the long run. Attending school brings a sense of normality in peoples lives, in a context where normality has been badly disrupted. It provides protection to pupils and allows the spread of survival messages such as information about HIV/AIDS. It enables the spread of peace messages, for example, by including peace or citizenship education in schools. An educated society may facilitate post-conflict reconstruction and redevelopment. Third, education is not inevitably a positive thing. Education may also facilitate conflict. Educational institutions and curricula may be harnessed by hostile groups and used to spread messages of hatred or to help justify and institutionalise conflict. Methodology Data sets The main data sources used in this work are publically-accessible, multi-country public opinion surveys, and include Arab Barometer, Afrobarometer, Asia Barometer, (East) Asian Barometer, LatinoBarometro, Pew Global Attitudes Project and the World Values Survey. All surveys are national probability samples and were chosen for their representativeness in each country. These are surveys of adult population. Sampling procedures employed multi-stage, stratified random sampling designs. In most instances the sampling procedure followed four basic steps: First, larger segments were selected based on population censuses using probabilistic sampling techniques. Second, smaller conglomerates or dwellings were selected from within the preselected segments, again using probabilistic techniques. Third, single households were selected from within these smaller conglomerates. Finally, respondents from within the households were chosen using a method for random selection. Often, but not always, the sampling procedure applied gender, age and education quotas. Completed sample sizes range from 710 to 2,500. The margin of error generally falls within 3 percent for a 95 percent confidence interval. Response rates ranged from satisfactory to very high (99 %). The surveys included a number of social, economic and political topics. The types of questions asked were broadly similar across surveys. Though, there are some important differences between surveys and these are given appropriate consideration. Countries To uphold UNESCO s commitment to the principles of international peace, global cooperation and common welfare of all mankind, this work has aimed to include as large a number as possible of conflict-affected and fragile countries, for which primary (or secondary) opinion survey data is available and in the public domain. This inclusiveness had to be balanced against data quality. If some countries are considered in more detail, this is purely for reasons of methodology. In countries where conflict was limited to specific regions but where regional analysis was precluded by data limitations or the contribution of these regions to the overall sample was small, the results may not fully reflect the link between conflict and education. Survey years were chosen in accordance with the following criteria: 1.) because the country or region was the scene of violent conflict at the time of the survey, 2.) the conflict is ongoing so 6

8 the most recent available time point was used, and 3.) the country is fragile and the length of time elapsed from the conflict is minimized. For several countries multiple data sources were available. A more detailed description of countries and survey years is included in the Appendix. Key variables The most important problem Question wording and response collection varied across surveys. In most surveys the question was open-ended. Respondents were asked to indicate what they consider to be the most important problem at the national level, and were mostly allowed up to three responses. The responses from multiple response questions were subsequently combined and the total proportion of responses pertaining to any given problem was calculated. In one exercise responses were weighted in order of importance so that the first problem mentioned by the respondent contributed twice and three times as much as the second and third responses to the overall result. Where respondents were asked to also identify the most important problem at the local and individual levels these results were compared to the results obtained at the national level. In a subsample of surveys the questions was closed-ended and respondents were asked to choose the most important problem from a short list (e.g. five options) of predefined problems. Government performance Satisfaction with government s performance in different sectors and areas was consistently measured on a modified Likert scale. Views of aid donors The surveys paid relatively little attention to the role of aid donors. The most commonly included question was a measure of the public s confidence or trust in a range of institutions and intentional organisations, including non-governmental, charitable and humanitarian organisations, which are involved in the provision of aid. A slightly more direct approach is found in the Pew Global Attitudes Project, where citizens opinion on the influence these nongovernmental, charitable and humanitarian organisations have in their respective countries was sought. Socio demographic characteristics The measures include urban-rural status, respondent s education, gender, age and income. The measures of socio-demographic characteristics were almost without exception categorical. Age is the only characteristic consistently measured on an interval scale following an open-ended response. Income was usually measured in categories. Depending on the purpose of the exercise some of the measures, namely income and education, could also be treated as ordinal. Analysis The analysis was performed in STATA, version 10. Data sets available only in alternative formats were converted into STATA compatible formats using Stat/Transfer. Descriptive statistics, such as frequencies and proportions, are reported and where desirable significance tests are performed. The proportions of don t know and refused responses are reported to allow for a more transparent interpretation of the results. 7

9 Problems and challenges identified by respondents are ranked from the highest to lowest in order of frequency. Tied ranking is avoided. Main findings The most important problems identified by citizens in conflict-affected and fragile countries at the national level are unemployment, poverty and management of the economy. This pattern is consistently observed in both the open-ended and closed-ended variant of the most important problem question. In open-ended questions unemployment, poverty and management of the economy emerge as the problems identified by the largest proportions of respondents in their first response and generally remain prominent as subsequent responses. The precise ranking and the proportion of respondents mentioning each problem vary across countries, and in some instances country-specific concerns emerge as equally important. For example, infrastructure is among the top three problems in Liberia, while it is not perceived as a salient issue in most other countries. However, the overall pattern is very similar across countries, and at least one of the problems from this group (most often unemployment) is always among the top three most important problems identified at the national level. In multi-response, open-ended most important problem questions education does not rank prominently, when only the first response is considered. However, the proportion of responses pertaining to education increases notably with subsequent responses. To illustrate, in Liberia only 2 percent identified education as the most important problem in their first response but 18 percent of the population did so in their second response followed by another 17 percent in their third response. This pattern is consistently observed across countries. At first education is mentioned by few respondents but the proportion of respondents ever mentioning education increases notably, when subsequent responses are considered. One way of dealing with multiple-response, open-ended most important problem questions is to combine all responses from each respondent in a single sample and analyse the proportion of total responses pertaining to each problem. In other words, responses are calculated as the percentage of responses rather than respondents. This also allows for a comparison to be made between single-response and multiple-response most important problem questions. When responses are combined the situation with regard to poverty and unemployment is little changed. These remain the most important problems identified by the public in conflict-affected countries. Figure 1 shows the results from the combined responses for a selected number of countries. The first and second most important problems defined as a proportion of total responses are pitted against education. Unemployment consistently ranks first or second. The highest percentage of responses pertains to unemployment in Nigeria and Colombia, while in Uganda, Liberia, and Indonesia unemployment is second. 8

10 Figure 1: Education s rank among the most important problems at the national level in conflict-affected countries (percentage of total responses) % Liberia Nigeria Uganda Indonesia Colombia Terrorism/Violence Infrastructure Poverty Unemployment Education Source: Afrobarometer 2005 (Nigeria); 2008 (Liberia, Uganda). East Asia Barometer 2006 (Indonesia). LatinoBarometro 2005 (Colombia single-response). Notes: Entries are percentages of the total number of all responses to the most important problem question (unweighted). Liberia: N=3,600; Nigeria: N=7,089; Uganda: N=7,293; Indonesia: N=4,248; Colombia: N=1,200. As the percentage of total responses education is more often perceived as a significant problem. The proportion of responses pertaining to education in combined samples is consistently greater than in single-response most important problem questions. While in some instances combining multiple-responses does not result in a ranking order significantly different from the singleresponse questions, it does in most. In Liberia for instance, the single most important problem is management of the economy and inadequate education is mentioned by a negligible minority. As the proportion of total responses in Liberia management of the economy does not even make the top three problems, while education ranks first at 13 percent. There is an ongoing debate about the meaning of multiple response most important problem questions (see Soroka, 2002). It has been suggested that the first and subsequent responses may not be substantively the same. Firstly, respondents characteristics vary more systematically with the first than subsequent responses. Secondly, more educated respondents are more likely to provide subsequent answers. This would imply that combining responses may not always be suitable. There is some indication that combining responses is problematic for time-series analysis (Soroka, 2002) but cross-sectional analysis is probably less likely to suffer from the same limitations. Moreover, there are no grounds to suggest that subsequent answers are somehow irrelevant. But if subsequent responses are, in fact, less calculated or less important it might be worth taking this possibility into account. If the first response is weighted to contribute twice and three times as much to the combined sample as the second and third response, for reasons of being mostly named in subsequent responses education usually drops two or more places down the ranks. Interestingly, the most important problems identified in single-response most important problem questions (i.e. unemployment, poverty, economy) are more robust against the different methodological specifications. 9

11 These patterns are largely reproduced in closed-ended most important problem questions (see example in Table 1). Again, when included among the available options poverty, unemployment and economy are identified as the most pressing problems by the highest proportion of respondents. If subsequent choices are allowed education is again increasingly mentioned. Table 1: Most important problem Gender Discrimination Poor sanitation and disease Inadequate education Environmental pollution Country Poverty Total Turkey ,346 Georgia ,500 Thailand ,534 Indonesia ,015 Serbia ,220 Ethiopia ,500 Source: World Values Survey The identification of the most salient issues at the national level is reflected in the publics evaluations of government performance in different areas and sectors. Table 2 illustrates the public s (dis)satisfaction with government performance in a selection of countries. Across countries the national governments performance is seen as being most dissatisfactory with regard to the management of the economy, job creation and improving the situation of the poor. Compared to other sectors and services the public appears to be generally satisfied with the governments handling of education. Even so, over a third of citizens in countries presented in Table 2 felt that the government was not doing a good job in addressing education needs. While the absolute level of dissatisfaction with regard to education is notable, in terms of the overall management of the economy, unemployment and other specific economic policies the absolute levels of dissatisfaction with the governments performance are much greater still. Like with education, the public is relatively satisfied with the government s performance in providing basic health services. Table 2: Evaluation of government performance in different areas and sectors (Percentage saying Very Bad plus Bad) Liberia Nigeria Uganda Palestine Addressing educational needs Managing the economy Improving living standards of the - poor Creating jobs Keeping prices down Narrowing income gaps Reducing crime Improving basic health services Source: Afrobarometer 2005; Arab Barometer

12 The public s opinion on international, non-governmental, charitable and humanitarian organisations in conflict-affected and fragile countries varies quite a bit across countries (see Table 3 for an example). On the whole, the public s views of these organisations are relatively positive; the proportion with a positive view is usually higher than the proportion with a negative views. On occasion NGOs, charitable and humanitarian organisation command more of the public s confidence than the national government (e.g. Palestinian Territories). Table 3: Percentage with a positive opinion on NGOs, charitable and humanitarian organisations Percentage confident, satisfied with influence or saying help more than a little Liberia Afghanistan Pakistan Palestine National government NGOs/charitable/humanitari an organisations United Nations Source: Afrobarometer 2005; Arab Barometer Pew Global Attitudes Survey, To reiterate, the main conclusions of this analysis are as follows: First, unemployment, poverty and management of the economy are the most important problems identified by citizens in conflict-affected countries and also the most immediate concerns for the government to address. Second, education does not emerge among the most pressing problems but on a closer examination the publics in conflict-affected countries clearly recognize the importance and value of education. Third, one possible reason why education does not appear to be the most immediate concern is the relatively high public satisfaction with the governments performance in the area of education. In contrast, national governments are seen as performing very poorly in terms of economy, unemployment and improving the situation of the poor. 11

13 Africa Liberia Results from opinion surveys Background In the 1970s and 1980s the deteriorating economic situation and individual living conditions lead to general public discontent across the country. (Afrobarometer Briefing Paper, 2009) These developments later contributed to the escalation of conflict into a full-blown civil war, which continued to cripple the country for the next 14 years. Only with the signing of a comprehensive peace agreement in Ghana in 2003 did the prospects for reconciliation start to look encouraging. The United Nations Mission in Liberia (UNMIL), which maintains a strong presence throughout the country, completed a disarmament program for former combatants in late 2004 but the security situation remained volatile and the process of rebuilding the social and economic structure of this war-torn country remains sluggish (Global Security.org, 2005; CIA Factbook). Today Liberia remains one of the poorest countries in the world; the conflict has left its economy in ruin and the country overflowing with weapons. Post-civil war economic and social revitalisation will continue to dominate the development agenda of this country for many years to come. The international community rates the education situation in Liberia as poor. In 1999/2000 the primary-level education net enrolment ratio was 79% for boys and 61% for girls (United Nations). Methodology The results presented in this section are based on the Afrobarometer 2008 Liberia survey. The sampling frame was designed on the basis of Liberian Population Census and the sample employed a random, clustered, stratified, multistage area probability sampling procedure with gender quotas to achieve the objective of a nationally representative sample (Afrobarometer Liberia Codebook, 2008). The response rate was very high at 99.1% and the proportion of missing responses on most questions was negligible, increasing our confidence in the results. This section presents the frequencies and proportions of responses given to each option, weighted so as to account for sampling probabilities. Most important problem faced by Liberia Liberians were asked to indicate their answers to the question: In your opinion, what are the most important problems facing this country that the government should address? The question was open-ended and Liberians were instructed to give up to three answers. Where more than three answers were offered the interviewers prompted respondents to indicate the three problems considered the most important. The results are presented in Tables 1 to

14 Table 1: The most important problem faced by Liberia R Problem No. of observations Percent 1 Management of the economy Unemployment Infrastructure/ roads Food shortage/ famine Wages, incomes and salaries Poverty/ destitution Farming/ agriculture Agricultural marketing Education Transportation Rates and taxes Loans/ credit Electricity Communications Health Drought Crime and security Land Total 1, Source: Afrobarometer, 2008 Note: R: Rank. At the national level the most important problem is managing the economy, mentioned by almost 27% of respondents. Concerns about the economy are followed closely by concerns about unemployment (~24%). Poor infrastructure ranks third and is mentioned by one tenth of respondents. In the first instance education was mentioned by only 2% of Liberians, and ranks ninth. While rank order is certainly interesting and illustrative, it should not be considered out of context, as a relatively small change in the total number of responses could significantly affect the rank assigned to education. If education was not among the most oft cited problems in the first instance, the picture changes dramatically when subsequent answers are considered (see Table1.1). Now, 18% of respondents have mentioned education. Education ranks first. It might be suggested that Liberians consider education to be the second most important problem at the national level. Education is followed by food shortages and famine (14%) and concerns about infrastructure and roads (13%). Unemployment remains a significant consideration, and is mentioned by another 10% of respondents. Economy no longer receives a mention. In third instance health surfaces as the most important problem and is mentioned by some 26% of respondents (see Table 1.2). Concerns about education now come second and are mentioned by another 17% of respondents. 13

15 Table 1.1.: The second most important problem faced by Liberia R Problem No. of observations Percent 1 Education Food shortage and famine Infrastructure, roads Transportation Unemployment Farming and agriculture Poverty and destitution Agricultural marketing Health Water supply Supply of electricity Rates and taxes Housing Wages, incomes and salaries Land Crime and security Loans and credit Drought Communications Orphans/ street children/ homeless children Services (other) AIDS Corruption Political instability/ political divisions Discrimination/ inequality Gender issues/ women's rights Other (i.e., some other problem) Total 1, Source: Afrobarometer 2008 Note: R: Rank. Infrastructure and roads remain important and security concerns rise in importance (see Table 1.2). The results so far suggest that the most immediate concerns identified at the national level, which the government should address, pertain to the management of the economy and creation of employment opportunities. Other more strictly development problems, such as education, infrastructure and healthcare, seem to be important though not immediate concerns. 14

16 Table 1.2.: The third most important problem faced by Liberia R Problem No. of observations Percent 1 Health Education Infrastructure/ roads Crime and security Transportation Water supply Electricity Housing Corruption Food shortage/ famine Loans and credit Poverty and destitution Communications Farming and agriculture Land Rates and taxes Services (other) Gender issues and women's rights Democracy/ political rights Political violence Agricultural marketing Discrimination/ inequality Drought Sickness and disease Political instability/ political divisions Civil war Management of the economy Orphans/ street children/ homeless children Wages, incomes and salaries Unemployment AIDS War (international) Other (i.e. some other problem) No further reply Total 1, Source: Afrobarometer 2008 Note: R: Rank. In addition it might be interesting to consider what proportion of all responses was education. These considerations are examined in Tables 1.3. and 1.4 below. 15

17 Table 1.3.: Summated proportions of Liberians mentioning the following issues as one of the three most important problems faced by their country R Problem Percent (summated) 1 Education Unemployment Infrastructure Health Famine Economy Poverty/destitution Income and wages 10.3 Source: Afrobarometer 2008 Note: R: Rank. N=1,200 Table 1.3. shows the proportion of respondents mentioning each of the problems as opposed to the proportion not mentioning the given problem. In effect, we are comparing dummy variables. We can see that nearly 40% of all respondents mentioned education with regard to the most important problem (MIP) question, while some 60% did not. Interestingly, out of all prominent problems the greatest proportion of respondents mentioned education, although as we have seen in most cases this was not done in first instance. The proportion of respondents ever mentioning unemployment is 33% and infrastructure was mentioned by 31%. Economy, despite being the most oft given answer in first instance, was not actually mentioned by an overwhelming number of respondents in total. If respondents answers to the MIP question are pooled to construct a single sample, then education is mentioned well over one tenth of the time (Table 1.4. left). In other words almost 13% of all 3,600 responses pertain to education. 11% pertain to unemployment and infrastructure, and 10% to health. As expected the results reflect the total proportion ever mentioning each problem (Table 1.3). If the responses in pooled sample are assigned weights so that the first problem mentioned is worth twice as much as the second one and three times as much as the third one, then education drops to fifth place and is mentioned in about 10% of responses (see Table 1.4. right). Unfortunately, this weighing is arbitrary because we do not know the actual degree of importance respondents ascribe to first and subsequent problems. However, this constitutes a useful statistical comparison allowing for the possibility that the first and subsequent responses are not on an equal footing. 16

18 Table 1.4.: Most important problem faced by Liberia (unweighted and weighted pooled samples) Problem N. obs. % Education Unemployment Infrastructure/roads Health Food shortage/famine Economy Transportation Poverty/destitution Farming/agriculture Wages, income, salaries Agricultural marketing Crime and security Water supply Electricity Housing Loans Taxes/rates Total 3, Problem Weight % Unemployment Economy Infrastructure Famine Education Health Wage, income Poverty Transportation Farming/agriculture Agricultural market Electricity Water supply Crime/security Taxes/rates Loans Housing Total 3, Source: Afrobarometer 2008 Note: R: Rank. Evaluations of the government s performance and responsibilities Table 2 shows Liberians evaluations of the national government s performance with regard to various areas. Some 65% of Liberians feel that the government is addressing educational needs well, of which 25% believe that the government has been performing very well. In fact, the government s performance with regard to education received one of the best ratings overall. This is in stark contrast to way in which Liberians evaluated government s performance in creating jobs and managing the economy both on the whole and with regard to specific policy areas (e.g. inflation, income inequality). 76% of Liberians feel that the government s record in creating jobs has been poor, 79% believe that the government has been failing in curbing inflation and 71% of respondents feel that the government has not succeeded in improving living standards of the poor. Evaluations of management of the economy in general are a bit better, but still more than half are dissatisfied. These results are consistent with Liberians identification of the most important problems faced at the national level, which the government should address. Also supported is the expectation that the first problem mentioned by respondents is usually also the problem considered most pressing or acute and in need of intervention. 17

19 Table 2: Ratings of government s performance in different areas Very badly Fairly badly Fairly well Very well DK/ haven t heard enough Addressing educational needs ,200 Managing the economy ,200 Improving living standards of the poor ,200 Creating jobs ,199 Keeping prices down ,200 Narrowing income gaps ,200 Reducing crime ,200 Improving basic health services ,200 Providing water and sanitation services ,200 Ensuring enough to eat ,200 Fighting corruption ,199 Maintaining roads and bridges ,200 Providing reliable electric supply ,200 Protecting rivers and forests ,199 Empowering women ,200 Combating HIV/AIDS ,200 Source: Afrobarometer 2008 N Table 3: Actors primarily responsible for providing the different services in the opinion of Liberians Central government Local Gov. Traditional leaders Community members None DK Managing schools Managing health clinics Collecting taxes Solving local disputes Source: Afrobarometer 2008 N=1,200 Interestingly, a significant proportion of Liberians believe that the primary responsibility for managing schools lies with the members of the community (Table 3). While the majority believe it lies with the central government, the proportion is lower than in the case of health care or tax collection. Given that the locus of responsibility for managing education does not unanimously lie with the national government, this suggests that some caution be taken when interpreting the results with regard to government performance. It would be interesting to see, if the pattern of public opinion with regard to education is the same at the local and national level but unfortunately this distinction was not made in the survey. 18

20 International actors and aid donors Table 4: Liberians rating of help received by international actors Do nothing Help a little Help somewhat Help a lot DK African Union ECOWAS United Nations International donors/ NGOs (other than UN) International business investors South Africa Nigeria United States Source: Afrobarometer 2008 N=1,200 The question presented in Table 5 comes closest to addressing Liberian s views of international aid. Liberians see the United Nations as a very positive actor; 70% believe that the UN helps Liberia a lot. Liberians views of international donors and NGOs other than UN are more mixed. Overall, the majority still feel that they do help but 22% feel they help only a little and 27% feel that they help only somewhat - contrasted by the 70% who feel that the UN helps a lot. The results are also mixed with regard to international business and investors, 30% feel that they help only a little, while another 12% are unsure. Liberians feel that their county receives a lot of help from the United States. They view Nigeria in a similar capacity, while South Africa is seen as providing less help. Liberians also feel that the African Union has not been among the most helpful actors. Effect on education of respondents socio-demographic and other background characteristics This section considers the effect of socio-demographic characteristics on respondents evaluation of the government performance in addressing educational needs. Table 6: Ratings of government performance in addressing educational needs by respondent s education (in percentages) Respondents education Badly Well n No formal education Informal education Primary education Secondary education Post-secondary education University education Source: Afrobarometer 2008 Note: Kruskal-Wallis test is significant at p 0.01 level. Don t knows were few and are excluded from the analysis. Spearman s rho is 0.06 and is significant only at p 0.1. Total N: 1,187 19

21 Respondents own educational attainment does not have a dramatic effect on the evaluation of the government s performance in addressing educational needs in substantive terms. Respondents without any formal education are less likely to be satisfied with the government s performance in this area, which is something that one would expect from self-interested actors. Broadly speaking having at least some or potentially having more education is related to better evaluations of government performance in this area. Table 7: Ratings of government performance in addressing educational needs by urban/rural status (in percent) Area Badly Well n Urban Rural Source: Afrobarometer 2008 Note: Wilcoxon-Mann-Whitney test is significant at p Don t knows were few and are excluded from the analysis. Total N: 1,193 As one might expect, rural dwellers are significantly more likely to feel that the government is performing badly in addressing educational needs as compared to their urban-dwelling counterparts. Table 8: Ratings of government performance in addressing educational needs by respondent s language (in percentages) Language Badly Well n English Bassa Belle Gbandi Gio Gola Grebo Kissi Kpelle Krahn Kru Lorma Mandingo Mano Mende Vai Simple Liberian English Sarpo Total ,200 Source: Afrobarometer 2008 Note: Kruskal-Wallis test is significant at p 0.05 level. The proportion of don t know responses was very small = 0.6% and for the most part equally distributed across the various language groups. Language, often a convenient approximation of ethnic and other differences between populations, does seem to have some bearing on respondents evaluation of government 20

22 performance with regard to education. However, caution is advisable as the number of respondents per language is often small. The same case can be made for region but the number of respondents per region is also mostly small. Gender has no effect on evaluation of government performance in the area of education. Secondary Sources Our results are broadly similar to those of other opinion surveys conducted in Liberia at approximately the same time. One such example is the survey, commissioned by the International Republican Institute (2008), which was conducted in three Liberian counties. The samples are made to be representative of the local population in those counties, and the estimated margin of error is ±3%. An obvious limitation of this survey is the limited number of counties included in the sample. Regional composition of the sample: Margibi County n=395 Grand Bassa County n=406 Nimba County n=505 Total N= 1,306 The most important problem Respondents from all three counties agree that the most important problem faced by Liberia today is food insecurity (see Table 1). Food insecurity is an especially salient issue in Grand Bassa. The second most important problem is lack of employment opportunities and it is particularly pronounced in Margibi. In our survey the most important problems were unemployment and the economy but on the whole the pattern is quite similar. Although, all three counties were included in Afrobarometer the number of respondents at the level of counties is only small, so direct comparison would not be sensible. Numerous reasons can be responsible for the observed differences between the results of studies. First, only one of the studies is a nationally representative sample. Second, alterations in question wording and response options can produce significant differences in the results obtained. The methodology for the International Republican Institute survey is not precisely documented, so it cannot be determined whether the most important problem question was of an open or closed type. In this survey education does not appear to be perceived as one of the most acute problems but it is nevertheless mentioned by 4% of respondents in each county (quite similar to the 2% mentioning education as the first most important problem in Libaria Afrobarometer 2008). Table 1: The biggest problem facing Liberia today (percent) Problem Nimba Margibi Grand Bassa All Food/hunger Unemployment/jobs Roads/transportation Crime Education Corruption Other

23 Source: IRI Author s own calculations. N=1,306 How much time should politicians spend discussing important issues Liberians were asked to indicate their answer to the question: In general, do you think political parties spend too much, too little or the right amount of time talking about [issue]? In Grand Bassa 65% of respondents thought that politicians spend too much time discussing education. At the same time a notable proportion of the people of Grand Bassa would like politicians to spend more time discussing provision of electricity, unemployment and provision of clean and safe water. Table 2.1.: Issue salience: Grand Bassa Time spent on discussion Issue Too much Too little Right amount DK Education Unemployment/jobs Food/hunger Clean and safe water Providing electricity Providing healthcare Eliminating corruption Reconciliation and healing of the country Source: IRI N=1,306 A similar situation with regard to education is observed in Margibi and Nimba (see Tables 2.2 and 2.3.). The people of Margibi would especially like too see more political discussion of unemployment and food insecurity. Table 2.2.: Issue salience: Margibi Time spent on discussion Issue Too much Too little Right amount DK Education Unemployment/jobs Food/hunger Clean and safe water Providing electricity Providing healthcare Eliminating corruption Reconciliation and healing of the country Source: IRI N=1,306 22

24 Table 2.3.: Issue salience: Nimba Time spent on discussion Issue Too much Too little Right amount DK Education Unemployment/jobs Food/hunger Clean and safe water Providing electricity Providing healthcare Eliminating corruption Reconciliation and healing of the country Source: IRI N=1,306 Ratings of government s performance In all three provinces, and Grand Bassa in particular, the majority feel that the government is performing well with regard to educational provision. This is consistent with the identification of the most important problem at the national level and results which suggest that Liberians would like to see political discussion focus elsewhere. In contrast to education, Liberians from all three provinces also broadly agree that the government s performance in dealing with food insecurity, action taken to eliminate corruption and in improving electricity supply has been poor. Table 3.1.: Government performance: Grand Bassa Performance rating Issue Good job Fair job Poor job DK Education Unemployment/jobs Food/hunger Clean and safe water Providing electricity Providing healthcare Eliminating corruption Reconciliation and healing of the country Source: IRI N=1,306 From Table 3.1. we can see that in Grand Bassa satisfaction with the government s performance in the area of education is high, indeed. 23

25 Table 3.2.: Government performance: Margibi Performance rating Issue Good job Fair job Poor job DK Education Unemployment/jobs Food/hunger Clean and safe water Providing electricity Providing healthcare Eliminating corruption Reconciliation and healing of the country Source: IRI N=1,306 Compared to Grand Bassa satisfaction with the government s performance in the area of education is somewhat lower in Margibi and Nimba, but still some two thirds believe that the government has been doing a good job. Table 3.3.: Government performance: Nimba Performance rating Issue Good job Fair job Poor job DK Education Unemployment/jobs Food/hunger Clean and safe water Providing electricity Providing healthcare Eliminating corruption Reconciliation and healing of the country Source: IRI N=1,306 These results are very similar to the results of the Afrobarometer survey for the country as a whole, where about 66% of respondents said that the government has been performing well with regard to education. 24