The Sierra Leone Web

Pipul dem Tok’

An Opinion Poll of Freetown Residents on Elections and the Peace Process

Monitor
A Project of the Campaign for Good Governance
Freetown, Sierra Leone

 

EXECUTIVE SUMMARY

As at many times in its recent troubled past, Sierra Leone faces complex and difficult choices over the next several months. A range of stakeholders, from the international community to local chiefs to the government and civil society, are currently grappling with charting the best path to securing lasting peace, reconstruction and rejuvenation for the country. All have come to realize that in charting this path, the voices of the people must be heard. It is in this context that the ‘Monitor’ team at the Campaign for Good Governance decided to undertake an opinion poll of Freetown residents. Our aim was to garner and make public the population’s views on key issues in the peace process and the elections in order to help frame and guide public debate and policy making - to "Leh di people dem tok". Between November 30th and December 11th, Monitor staff interviewed a representative sample of 3039 adult Freetown residents. Individuals were asked 19 questions, as well as their age, income, educational and tribal background and IDP status. The full results of the poll and explanation of our methodology are set out in the report below.

From inspiration to implementation the motive behind the poll has been to provide the public in Sierra Leone and abroad with a non-partisan, apolitical measurement of the views of Freetown residents. We would prefer to refrain as much as possible from interpreting the results and rather let them speak for themselves. However, in an effort to summarise the findings, we have drawn out a few key themes:

First, a degree of confidence has been expressed in the peace process, with 63% of respondents confident that peace has come, only 15% opposed to the amnesty that has been offered to junior level ex-combatants, and 67% reporting that they would welcome ex-combatants back into their community.

Second, mixed results were returned on the level of trust in and support for the government. Only 48% of respondents trusted the National Electoral Commission (NEC) to hold free and fair elections and 34% see the Civil Defense Forces as a threat to the peace. However, at 24% support, the SLPP is the most popular political party and 80% of respondents think President Kabbah has been a good president.

Thirdly, the poll demonstrates some opposition to the election track endorsed by the National Consultative Conference held in mid-October. A majority (57%) of respondents favour a postponement of the elections scheduled for May. Also, not only do 82% of people not understand the district block electoral system (DBS) proposed by the NEC, but 63% of those that do understand the DBS would prefer constituency elections.

A final theme observed is some lack of confidence in Sierra Leone’s current political scene. 61% of respondents say they do not support any political party, and 78% would be in favour of an international trusteeship to run their country.

We are confident that this poll is an accurate representation of the views of Freetown residents. However, there are several challenges and pitfalls associated with opinion polling anywhere, and accurate polling in Freetown is particularly difficult. A number of concerns, therefore, should be addressed. First, due to logistical and funding constraints in conducting a properly representative poll, Monitor was forced to limit the scope of the poll to Freetown residents. There is a danger that the poll results will be understood to be representative of the country as a whole, and we would wish to warn against this at the outset. While many Freetown residents (40% according to the poll) are internally displaced persons from other parts of the country, the concerns and conditions of Freetown are notoriously distinct from those in the provinces. We hope that our effort will inspire other institutions to seek the resources necessary to engage in similar polling efforts at a national level.

Secondly, there is the question of whether the poll was truly ‘randomised’, i.e. whether its 3000 interviewees were a perfectly representative sample of Freetown residents. Getting a perfect sample in a poll is always difficult, if not impossible. The situation in Freetown, with limited demographic and other information available and large transient and displaced populations, makes it even more so. This may be one reason why a poll like this one has never, to our knowledge, been conducted in Sierra Leone before. However, Monitor has employed as robust a methodology as possible under the circumstances (see pg.9) and based on this methodology and the demographic breakdown of our sample (see pg.4) we are confident that we have captured a highly, if not necessarily perfectly, representative sample of Freetown residents.

Finally, there is the issue of bias and impartiality in conducting the poll. Monitor is a project of the Campaign for Good Governance, an organisation which has become famous for its outspoken contributions to political debates in Sierra Leone, in particular its criticisms of the RUF and the Government. Might the organisation not be tempted, even if only in subtle ways, to influence the outcome of the poll? In response to this last concern, we can only offer complete transparency of our methodology, recruiting and staff members, and reiterate our motives of impartial provision of information that lie behind the poll and the Monitor project.

There are many hopeful messages in the poll. In general we were surprised by the lack of divergent views among people of different tribal backgrounds, certainly a hopeful sign for the future of Sierra Leone’s politics. In addition 68% say they intend to vote in the upcoming elections, including a large proportion of young people. Finally the results on Freetown residents’ willingness to accept ex-combatants back and accept the amnesty given to them is testament to their personal commitments to taking difficult steps toward peace.

These results provide a snapshot of public opinion in the capital. All in all, despite its necessary faults, we hope that this poll will contribute to public debate in Freetown and elsewhere about the future of Sierra Leone and the many difficult choices that lie ahead.

The Monitor Team

Campaign for Good Governance

###

 

THE SAMPLE

3039 Freetown residents were interviewed for approximately 15 minutes each between the 30th of November and the 11th of December (inclusive), 2001. The following figures illustrate the composition of this sample. The extent that the sample succeeded in being representative, it provides a snapshot of the demographics of Freetown.

Sex

Age Range

Monthly Income
(1000 Le)

Highest Level of Education Attained

Ethnic Background

 

In addition, 78% of the sample described themselves as literate, 41% as having been internally displaced by the war and 2.5% as ex-combatants.

THE RESULTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Margin of error ±1.8% on a 95% confidence interval.

 

ANALYSIS OF RESULTS

From inspiration to implementation the motive behind the poll has been to provide the public in Sierra Leone and abroad with a non-partisan, apolitical measurement of the views of Freetown residents. We would prefer to refrain as much as possible from interpreting the results and rather let them speak for themselves. However, in an effort to summarise the findings, we have drawn out a few key themes:

High Degree of Confidence in and Commitment to the Peace Process

First, a degree of confidence has been expressed in the current peace process, with 63% of respondents confident that peace has come, 67% saying that they will accept ex-combatants back into their community, and only 15% arguing against the amnesty provided by the Lome Agreement for almost all combatants. The results on accepting ex-combatants and the amnesty are particularly heartening indications that the people of Freetown are committed to making the peace work. In another vote of confidence, only 14% believe the army is likely to mutiny against the state once again. The credit for these developments goes more to the British than the UN, though not overwhelmingly so.

Mixed Results on Trust and Support of Government

Second, the level of trust in some key government institutions appears mixed. Only 48% of respondents trust the National Electoral Commission to hold free and fair elections, and 34% see the government-allied Civil Defense Forces as a possible threat to the peace. Standing against these results however are the relative popularity of the ruling SLPP, which at 24% enjoys 10% more support than all the other parties put together, and of Kabbah himself, who returns an 80% approval rating. One of the reasons for Kabbah’s popularity might be the belief, held by 34% of Freetown residents, that he is personally responsible for the presence of the international community in Sierra Leone.

Serious Concerns about the Current Elections Track

Third, the current election track seemingly adopted by the government and the international community of using a district block electoral system to hold elections in May, has been called into question by the poll. Only 38% of respondents favour holding elections in May over postponement (57%) of the date. The answers to this question had a relatively low number of ‘unsure’ responses (5%), indicating that people had considered views on it. In addition, not only do 82% of people not understand the district block electoral system (DBS) proposed by the NEC, but 63% of those that do understand the DBS would prefer constituency elections.

No-Confidence Vote in Domestic Politics

Finally, some poll results are indicative of a dissatisfaction of Freetown residents with the current political system. 61% of respondents indicate that they do not support any political party and an equivalent number are unsure that any of the current 24 political parties represent the interests of the people. This may explain why 78% of respondents would be willing to turn the country over to an international trusteeship by the British or the UN until elections can be held.

Hopeful Signs

Finally, it should be noted that there are a number of messages of hope in the poll. It’s worth reiterating the commitment of the people of Freetown to taking their own difficult steps to contribute to the peace process, with 67% ready to welcome ex-combatants back. The low level of people (15% ) insisting on justice for all perpetrators might even indicate a level of forgiveness for what ex-combatants have done. Also, despite the general lack of support for the current parties, over 70% of Freetown residents say they intend to vote in the upcoming elections, which raises the question of for whom they will vote.

The last question in the poll might have perplexed some readers. The Monitor team, which includes both national and international staff, decided to ask this question to confirm or reject a popularly held view outside Sierra Leone that as a result of the country’s poverty and sufferings, its people must be miserable. The team believed that the people of Sierra Leone have an extraordinary resilience and are wealthy in terms of the quality and integrity of their family and community structures, as well as their religious life. The result, that 58% of respondents describe themselves as generally happy people, (or in Krio, think their ‘hearts are sweet’) is equivalent to happiness ratings in OECD countries. Such a rating for the country at the very bottom of the Human Development Index is food for thought for those who would adopt what we call ‘crusading development’, which seeks to lift the people out of their supposed misery through growth strategies which can disrupt community and family structures and traditional ways of life.

Breakdown of Results

The database collected in this survey is large and bears much analysis. For the sake of brevity, we have not included any of the breakdowns in this report. If you are interested in knowing any particular breakdowns – e.g. how responses to questions varied with gender, education or ethnic background; or what % of people who answered ‘yes’ to question X also answered ‘yes’ to question Y –please contact the Monitor team at sierraleonemonitor@yahoo.com, stating clearly which breakdown(s) you wish to receive and providing an email address to which to send the result.

THE METHODOLOGY

The methodology adopted to conduct the poll was of random sampling of the population of Freetown. 11 interviewers worked for 10 days approaching individuals throughout Freetown – on the street, in homes and in offices. Interviewing took place in the early morning, midday, evening and weekends. In all cases, the interviewer randomly selected and approached the interviewee, explaining the poll and its purpose and asking them if they would be willing to participate in the poll. Interviewers reported a 96% response rate to a first approach – in other words, only 4% of people refused to participate upon being asked. The attempt to conduct completely random sampling was made within ‘strata’ or sections of the population. The sample was stratified on the basis of available or estimable knowledge about the population of Freetown, such as the percentage of the population living within a given neighbourhood, or the percentage of women in the population.

Controlling for District of Freetown

The first strata, and basis for choice of interviewees, was location. Priority was placed on ensuring that each neighbourhood of Freetown was proportionally represented in the poll, in the belief that the demographics of the city, including tribal background, income, and political views, vary significantly by neighbourhood. The Central Statistics Office conducted a comprehensive population survey of the city in the year 2000, which gave precise numbers of residents for every street, area and neighbourhood of the city[1]. We divided the city into 8 districts, and apportioned our questions according to the relative populations of the district implied by this survey. For example, West III district (encompassing an area from Hill Station to Lumley to Aberdeen) was noted to contain 20.09% of the population of Freetown, so 20.1% of our sample number of 3000 questions, or 610 questions, were allotted to this district. IDP camps were considered as separate districts and each apportioned their share of the questions accordingly.[2]

Assumption: The relative population shares of the 8 districts of Freetown have not changed significantly since the Central Statistics Office survey last year.

In addition to population figures for districts of Freetown, the CSO also provided a comprehensive list of all streets within each district, first compiled in the 1986 census and updated since. Deciding upon a number of 5 interviews per street, we then determined the number of streets per district on which interviewers should ask questions (e.g. 120 streets in West III), and randomly chose this number of streets from the available list.

Controlling for Location within District

It was also considered that within any given neighbourhood, the next most significant demographic variation would be among people in different locations at any given time, such as on the street or in an office. We therefore attempted to accurately apportion our questions to these ‘locations’. We postulated that within a given district of Freetown, persons were likely to be found in one of three places at any given time: the home, the ‘street’ (a category which included bars, community hangouts, and open-air garages as well as the pavement), and the ‘office’, which included street-side shops as well as formal closed offices. Using a focus group consisting of the 15 monitor staff members, which included residents from all 8 of our districts, we estimated for each of the districts the percentage of people who would be found in each of these three places on any given day, over the course of 3 separate time periods per day. These proportions varied both with the day and the time of day. Interviewers were then given ‘selection sheets’ indicating to them, based on day and time of day, the proportion of interviews from each location they should obtain. The sheets also provided a random sequence in which the interviewer should decide among the three locations for each interview.

Assumption: Monitor focus group estimates for percentage of the population found in the ‘home’ on the ‘street’ or in the ‘office’ at any given time were roughly accurate.

Controlling for Gender

At one point during the survey the results yielded by this process were checked against known data, such as the percentage of women in the population (estimated on the basis of the results of the1986 census of Sierra Leone, which shows women in the slight majority). The focus group estimates for location were altered accordingly. A slightly larger percentage of men than women were returned by the original estimates, and Monitor staff believed that women were more likely to be found in the home than men at most times of day. The proportion of interviews conducted in the home was correspondingly increased, in order to obtain an equivalent number of male and female interviews. Gender therefore became the third strata within which randomization took place.

Assumption: Women and Men each account for roughly 50% of the population of Freetown.

Controlling for Time of Day and Day of Week

Concerned that our attempts to control for ‘home’, ‘street’ and ‘office’ effects faced particular difficulties during weekdays and working hours, an attempt was made to concentrate polling time on the weekends and in evenings when office workers in particular were likely to be at home. Well over 60% of the questions asked in the poll were asked during these times.

Assumption: Those who work far from home or in offices are more likely to be at home in the evenings and during weekends.

Selection Protocol : Randomly Choosing Interviewees

Interviewers were trained in rigorous randomizing selection methods in order to eliminate any personal biases they might have in whom to approach. When interviewing on the street, interviewers had to turn alternatively to the left and right, choose a person, and then count 11 people away from that person in order to choose the person whom they approached. They had a similar procedure for selecting houses and offices/shops. Once in a home or office/shop, interviewers were required 3 times out of 4 to ask to speak to someone other than the person who greeted them, by asking 1 time out of 4 to speak to someone of a different gender, and 2 times out of 4 to speak to someone older or younger. This was to avoid a gender or age bias in who is likely to meet a stranger at the door of a home, office or shop.

Training, Approach and Questions

The full 15 members of the Monitor poll team participated in development of the methodology, approach and drafting of questions for the poll. The Core Team drew up a list of questions which were then critiqued and altered by the larger team, who also collectively decided upon standard principles and methods for dealing with questions from interviewees which were consistent with the spirit of randomized and impartial polling. The full team also participated in drafting the approach/introduction used by interviewers. The approach, shown in the box below, was intended to assure the interviewee of their anonymity and our non-partisanship, and prepare them to answer controversial questions.

 

 

An afternoon was then taken for each interviewer to practice the approach and polling technique in trial interviews, after which the group came back together to discuss problems and make alterations. All interviewers wore clear ID cards which gave their full names, positions and institution.

Monitoring and Evaluation

Each day, each interviewer was required to conduct 30 interviews. They were met in the evening by a member of the Monitor core team who discussed any problems, and closely scrutinized answer sheets for errors or irregularities. Data was then inputted into an excel spreadsheet and the standard deviation of responses to each interviewer was checked against the rest.

Calculating Error

We used an online computational program to calculate the margin of sampling error for our poll, which can be found at www.surveysystem.com. To calculate the margin of error we had to assume a population size of adult Freetown residents, based our data for the total population of Freetown plus the population of Freetown IDP camps – 949,895. We chose 750,000 as our estimated population size. This may seem a fairly arbitrary number but it is reasonably certain that it is an overestimate, and therefore if our error calculation is flawed as a result it is only in overestimating the margin of error. Each poll result has a different margin, but since they all range between + 1% and + 1.8%, we chose to cite the maximum range of 1.8% as the margin of error for a 95% confidence interval.

Problems Encountered

Three principal problems were encountered. First, there was difficulty in achieving gender balance, even when the proportion of homes approached was increased. After a group dialogue with all the interviewers, it was determined that women were much more likely to be uninterested in participating in the poll, especially if approached by other women. As a result, interviewers were required to conduct a certain number of interviews with women in order to achieve a better gender balance in the poll.

A second problem encountered concerned the difficulty with movements of populations between districts of the city during the working hours of weekdays. Our focus group indicated that many residents moved from outlying districts, particularly the western districts, into the centre district for work. Lacking the ability to estimate these numbers accurately, and concerned to avoid over-polling residents of the central district, we chose to maximize polling at the weekends and in the evening in these districts, rather than attempt to shift more questions to the central district. It is therefore possible that we may have somewhat under-polled professional working residents of some of the outlying districts.

A final problem concerned staff failing to follow the selection protocols. Thankfully, this problem was had only with one staff member, and it was caught early on in the poll by our daily monitoring of the results. The staff member was dismissed, and their quota of interviews filled by another staff member.

Budget

Monitor is committed to absolute transparency of its activities, including financial. The total project budget for this poll was $1100. The project is not currently funded by any donor agency, and the poll was funded entirely from the personal donations of staff members of Monitor and Campaign for Good Governance.

Acknowledgements

Also deserving of acknowledgement are: Professor Claudius Thomas, founder of the Fourah Bay College Public Policy Unit, who has been a valuable source of counsel and who we know will follow up this effort with far more rigorous and comprehensive public policy polling work; Professor Kandeh at the Central Statistics Office, who provided invaluable data extremely promptly; Tom Perriello, Yale Law School Fellow, who lent us his advice and his car at critical moments; Comfort Ero of the International Crisis Group, who also gave valuable advice and invaluable support; Attilla Alpman, for his help in producing the pictures above; and finally Lena’s husband to be, Babs, for his unpaid heroism in putting this report on the web.

About Monitor

Monitor is a new project of the Campaign for Good Governance. A hybrid of investigative journalism and think-tank policy development, the team works at a fast pace to publish analytical reports on some aspect of the public sector in Sierra Leone every 3 weeks. The aim of Monitor is to contribute to the substance of public debate through the provision of information, analysis and concrete public policy options and strategies. The project is funded by individual private donations from staff members of Monitor and the Campaign for Good Governance.

The Monitor Team

The Monitor poll team consisted of a diverse, talented and dedicated group of individuals, all of whom are university graduates and also have a proven commitment to public service and the betterment of their country. They were selected from among a much a larger number of candidates who applied in response to an ad in several newspapers entitled: ‘Interested in working long hours for low pay to make a difference in Sierra Leone?’ The team is diverse, of different gender and ethnic backgrounds, which allowed us to check whether such interviewer characteristics had any effect on responses to the poll. There were no effects of this kind. The team members are shown below.

The Core Team

Abdul Kpakra-Massally has recently worked as Executive Secretary of PRIDE, an organisation working with ex-combatants. He has a BA in Politics and History from Fourah Bay College, where he was Secretary General of the Students Union.

Ricken Patel has worked for Harvard University and the International Crisis Group. He has a BA Hons in Philosophy, Politics and Economics from Balliol College, Oxford University and a Master in Public Policy from the John F. Kennedy School of Government at Harvard University.

Tom Pravda has recently worked as a Consultant at the United Nations Development Program in New York. He has a BA Hons in Philosophy, Politics and Economics from Balliol College, Oxford University.

Lena Thompson is a Lecturer at the Political Science Department, Fourah Bay College. She has a BA Hons in Political Science from Portsmouth University and an MA in International Relations from the University of Kent, Canterbury.

The Interviewers

Abdul-Gibril Bah recently graduated from Fourah Bay College with a BA Hons in Political Science.

Nasratu Bangura recently attended Njala College, where she gained a BSocSc. (Sociology)

Zynab Binta-Kamara is an actress at Spence Productions. She has a BA in English and Political Science from Fourah Bay College.

Lovette T. Braima has worked as a teacher and as a volunteer with SEGA, an NGO working with socially exploited women in Bo. She recently graduated from Fourah Bay College with a BA in Sociology and History.

Adama Kamara-Harding has recently worked at the Ministry of Finance and the Ministry of Development and Economic Planning. She has a BSC in Accounting from Fourah Bay College.

Randolph Katta has worked as a teacher, a Supervisor for Catholic Relief Services and a reporter for Radio Democracy. He has a BA Ed. in Linguistics from Njala University College.

Allen Lahai Murana has worked as a teacher and recently graduated from Fourah Bay College with a BSocSc. (Economics).

Shirley Osho recently graduated from Fourah Bay College with a BA Hons in Political Science.

Caleb Michael Thomas is a graduate of Fourah Bay College, where he gained a BSocSc. (Sociology).

Celia Thompson is an agent at Spence Productions. She recently graduated from Fourah Bay College with a BA Hons in Linguistics.

Vincent Williams has recently worked as a columnist and a librarian. He has a BA in Politics and History from Fourah Bay College; he also has a Diploma in Library Studies from the same institution and a City and Guilds of London Institute in Librarianship.

----

[1] "Population Density of Freetown, 2000 Estimate", Central Statistics Office, 2000.

[2] Figures for numbers of IDPs in each camp were courteously provided by the National Commission for Resettlement, Rehabilitation and Reconstruction.