fieldwork in cape verde (west africa)
I have designed, organized (including the recruitment and training of a local team of 9 collaborators), and conducted a Household Survey on the Quality of Public Services (Corruption) and Migration Experiences, including 1072 interviews in 30 of the 561 census areas of Cape Verde (CV). A parallel field experiment, including 997 subjects, was also carried out on the demand for political accountability in CV.
This survey was aimed at measuring three different sets of variables:
Changes in perceived corruption (mainly focused on influence) in the public services of the country (in health care, schooling, scholarships, justice, police, customs, licenses; and in the allocation of state jobs, subsidies, supplier positions). This part of the intervention was directed at gathering data for an EXTERNAL CONTROL GROUP, for comparison with the data from the survey conducted in 2004 in São Tomé and Príncipe (STP), in what the STP OIL DISCOVERY (1997-99) is concerned. We have also asked about the before-after democratization (1988-1990) time periods, just like in STP.
This lead me to aim at interviewing 30+ year old people (one representative per household) and to consider three periods of interest in the questions of the questionnaire, as for the STP survey.
The main technique used was to begin by eliciting memories from respondents, i.e. by asking questions on demographics, and marking personal happenings in the respondent's period of life (of interest) - e.g. year children born, illnesses in the household. Then, in the questions relative to corruption, the interviewer, when referring to the three periods of interest, used the information collected earlier in the interview.
DEMOGRAPHIC characteristics relating to MIGRATION (where the household was defined as anyone living in the house sampled, plus spouse and/or children of the subject of the interview, who was at least 30, in case of migration):
Questions on current migrants in the household.
Questions on return migrants in the household.
Questions on senders of remittances.
Questions on households' businesses and sources of investment.
The DEMAND FOR POLITICAL ACCOUNTABILITY (the 'POSTCARD EXPERIMENT').
This experiment was aimed at understanding who are the demanders of the "good" (as widely viewed by the international community, especially when comparing with African average) governance of the country. For that we made use of the quality of the public services/corruption part of the survey to propose the following deal to our respondents in the end of each interview:
We (Oxford) provide the results of the survey on quality of the public services (corruption) for media broadcast in Cape Verde if we have more than 50% of the postcards we are handing out to each of the around 1000 respondents of this survey back in our mail box at the end of the data collection.
This was aimed at identifying those respondents that demand most political accountability (namely, those who want the level of corruption in public services to be widely broadcasted), and therefore willing to pay the cost of getting the postcard to the mail (the stamp cost was supported by us).
Critically, for this to be credible we had to "show" media exposure before the survey conduction (details below).
This page is intended to provide an introduction to my work in CV and to the country itself (to know more about the country, click here).
Fieldwork (December, 2005 - March, 2006):
Stage 1: Arrival and Getting to Work
When I arrived in Praia, I had a worrying set of "to do's":
getting in contact with INE - Instituto Nacional de Estatística (National Statistics Office of Cape Verde), to have census data to take care of the sampling base (for representativeness), and to have some advice on logistics (transport prices, wages) and on the intervention itself.
getting in contact with UJP - Universidade Jean Piaget, for now the only university in Cape Verde, in order to have a base for the recruitment of enumerators.
getting in contact with the media, for an announcement of this intervention (basically for added credibility in the 'Postcard Experiment' and added general participation in the survey).
getting in contact with the post office headquarters for the 'Postcard Experiment'.
finish/piloting the questionnaire, to be submitted (NO PAPER THIS TIME!) on Oxford-borrowed HP iPAQ handhelds (together with GPS bluetooth transmitters), through Pocket Survey software. I was carrying 9 handheld machines, therefore aiming at 9 enumerators working at a time (in the IT context I would like to thank Richard Payne and Andrew Zeitlin at CSAE).
Alexandre Albuquerque central square in Praia.
I met with the President of INE, Francisco Tavares and with one of my contacts in the country, Ms Deolinda Reis (INE and Afrobarometer), who was most kind and helpful throughout this intervention. I was given access to the census area maps (through Direcção de Serviço de Cartografia e Cadastro, Ministério das Infraestruturas e Transportes) and population data from INE, which enabled accurate sampling of the target population in the survey. I thank INE for that.
The sample of census areas (from the total 561) was chosen randomly weighting by the number of households in each area. This was done in two stages: four islands first (which led to choose Santiago, São Vicente, Santo Antão, and Fogo), then the 30 census areas.
I met with UPJ representatives. In this context I would like to thank Dirce Varela, who let me announce the enumerator job opportunity to her students, who became well represented in this endeavor.
I met with information director Isabel Mendes (RTC Television), director Filomena Silva (Newspaper A Semana), director Alexandre Semedo (Newspaper Horizonte), director Vlademiro Marçal (Newspaper Expresso das Ilhas), director João Nascimento (Radio Nacional), director Carlos Gonçalves (Radio Comercial). All were very welcoming and willing to provide coverage of the survey. A special thanks to Alexandre Semedo for providing me with further candidates for enumerators.
I met with commercial manager for Correios de Cabo Verde (the post office company) Mário Moreira, who was helpful in providing coordination from the post office side. Later on, I made sure local post office branches were aware of the 'Postcard Experiment' by contacting directly the Mindelo (São Vicente), Ribeira Grande (Santo Antão), and São Filipe (Fogo) offices.
Stage 2.1: The Postcards and Media Attention
The Postcards: they were typed in Tipografia Santos (Praia), after getting a postal box in the capital city on purpose for the experiment. Each postcard had the mail paid in advance in the form of a stamp recognizable by the post office company. These had the format below, when handed out to the subjects of the study (click for details).
The front of the postcard is above in the image. Note numbered card – this is an unused number (977). The message in the card is (translation):
“I wish that the conclusions of the survey on the quality of national public services (health, education, justice,…), conducted by the University of Oxford (UK) in the first months of 2006 to 1000 households in the islands of Santiago, São Vicente, Santo Antão, and Fogo, are made public in the Cape Verdian media.”
The back of the postacard is below in the image. Note “PAID STAMP” note and stamp. It is addressed to a University of Oxford postal box in Achada Santo António, Praia).
National Television Station - Rádio Televisão Cabo-Verdiana RTC - 3min 40sec (including interviews with myself and two enumerators) broadcasted in the first half of the main prime-time news at 8pm (24/01/06)
Radio Station Nova - interviewed at the same time as for the TV (24/01/06)
National Radio Station - interviewed on the phone, broadcasted later that day (24/01/06)
Radio Station Comercial - news based on distributed press note (24/01/06)
Newspaper Expresso das Ilhas - news based on distributed press note and Radio Nova interview (25/01/06)
Newspaper A Semana, based on a two-hour interview (26/01/06)
Newspaper Horizonte - news based on distributed press note (26/01/06)
Stage 2.2: Training of the Interviewers and Trialing of the Questionnaire
training and trialing took close to three weeks; each enumerator had a total (on average) of 18 hours of training in groups of 2-3 people; training included lectures on the content and objectives of the questionnaire, answering the questionnaire, and trialing (at least once by interviewer) with a random person. These sessions took place in Universidade Jean Piaget, Palácio da Cultura Ildo Lobo, Centro Cultural Françês, Sofia (all in the capital city, Praia).
Training Session at Palácio da Cultura Ildo Lobo - enumerators Lívia and Ângela.
From an initial group of 32 (!) people (who have done at least one training session), I had the honor to work with the following 9 ladies and gentlemen:
While the training was underway one of my main worries was to improve the questionnaire. Apart from the many contributions for the earlier STP instrument which was repeated in CV, I would like to thank the inputs of Abigail Barr and Pieter Serneels on the full CV piece. Cátia Batista and Aitor Lacuesta are collaborators in the migration part of this project: therefore they are also responsible for the migration-related questions and for several full-piece suggestions. I also would like to thank the initial remark of Paul Collier who has drawn our attention to migration issues on Cape Verde. The final version of the questionnaire can be seen here (pdf files are for 'paper' versions of the questionnaire; the 'real' pocket survey file is also available):
Stage 3: Survey in the Santiago island
Map with the Sampled Areas (places visited in the survey) in Santiago - click for details.
Survey in Santiago had several obstacles, from trained enumerators quitting to robbery attempts. However, we managed to have 9 enumerators, with the 9 machines, doing an excellent job for many of the collection days; in some days, namely before going to the interior of Santiago on Feb. 7, we had to go on with smaller teams (therefore at a lower pace).
This is the schedule we followed in Santiago island (including all towns/neighborhoods visited):
January, 28: Plateau (district of Praia)
January, 29: Chã de Areia, Achada Santo António (district of Praia)
January, 30: Achada Santo António (district of Praia)
February, 1: Tira Chapéu (district of Praia)
February, 2: Achada Eugénio Lima (district of Praia)
February, 4: Achadinha / Vila Nova (disrict of Praia) - robbery attempt I to the team
February, 5: Vila Nova / Safende (district of Praia)
Alto de Safende view: enumerators Rónyo and Hamilton before starting the day.
February, 6: Sáo Martinho Grande (district of Praia) - a very welcoming village just outside the capital
February, 7: Pedra Badejo - Achada Fátima / Achada Fazenda (district of Santa Cruz)
February, 9: Boa Entrada, Poilão da Boa Entrada / Assomada - Trás de Empa (district of Santa Catarina)
Poilão da Boa Entrada, the name of a gigantic tree (behind): enumerators Rónyo, Vladmir, Edimilson, and Hamilton (left);
Vladmir interviewing (right).
February, 10: Cutelo Gomes, Mato Afonso, Mato Martins (district of São Domingos) / São Jorge, Ribeirão Galinha (district of Santa Cruz)
São Jorge: (the team) starting from me, Ângela, Lívia, Célia, Nilson, Vladmir, Edimilson, Admir, Rónyo, and driver Mário (an emigrant in Boston) (left);
interviewing, always with a census area map, in Cutelo Gomes (middle);
Admir interviewing in Mato Afonso (right).
February, 11: Figueira Muita (district of Tarrafal) / Tarrafal - Codje Bicho (district of Tarrafal) - robbery attempt II to the team at arrival to Plateau
February, 13: Chão Tavares, Cruz Grande (district of Santa Catarina) / Achada Monte (district of São Miguel)
Stage 4: Survey in the islands of São Vicente, Santo Antão, and Fogo
Map with the Sampled Areas (places visited in the survey) in São Vicente - click for details.
Map with the Sampled Areas (places visited in the survey) in Santo Antão - click for details.
Map with the Sampled Areas (places visited in the survey) in Fogo - click for details.
Survey in the islands of São Vicente, Santo Antão, and Fogo was more straightforward, as the "machine" was already oiled. Due to budget constraints we opted to take 5 enumerators to these other islands (which needed per diem pay and internal flights). The best people were chosen (Admir, Edimilson, Hamilton/Lívia, Rónyo, Vladmir) - which worked as a kind of bonus for their great Santiago work. As in STP, it was very nice to see some first-time flights in the team!
The shedule was:
Sáo Vicente Island:
February, 15: Mindelo - Morada (district of São Vicente) - this is the historical center of Mindelo, a very problematic area for a survey (on participation), which led to seeking the local Radio Nova announcement referred above.
February, 16: Mindelo - Morada / Mindelo - Madeiralzinho, Chã de Alecrim (district of São Vicente)
February, 17: Mindelo - Monte Sossego (district of São Vicente)
February, 18: Mindelo - Bela Vista (district of São Vicente) - a bad day
February, 19: Mindelo - Fonte Inês, Espia (district of São Vicente)
Mindelo - Fonte Inês, Amilcar Cabral central square: preparations for Mardis Gras on the street.
Santo Antão Island:
February, 20: Porto Novo (district of Porto Novo)
the São Vicente/Santo Antão team (left) - a low moment, early morning, after 5 non-stopping days in São Vicente, and after rushing for the ferry, before another full day in Porto Novo;
end-of-the-day checking, checking, checking (right).
February, 21: Fajã Domingas Benta, Chã de Enrique, Lombo Pedranto (district of Ribeira Grande)
Fajã Domingas Benta: fabulous Santo Antão landscape (left), Rónyo interviewing (right).
February, 22: Chã de Pedras - Pia de Cima, Várzea, Pia de Baixo (district of Ribeira Grande)
Chã de Pedras: walking, a major part of this job (left); Rónyo interviewing (right).
February, 25: São Filipe - Fonte Aleixo, Cobon (district of São Filipe)
February, 26: Curral Grande (district of São Filipe)
February, 27: Queimada Guincho (district of Mosteiros)
The End, (the happy Fogo team) on the way back from Mosteiros, a stop at Ponta de Salina.
I was present physically (supervising the distribution of enumerators) at all 30 areas of sampling in the survey. I conducted interviews or observed its conduction (for the improvement of the performance of the interviewers, especially during the first part, until Achada Monte, in the interior of Santiago) at all areas except in 7: Achada Santo António (full day), Tira Chapéu, Achada Fazenda, Assomada - Trás de Empa, Tarrafal - Codje Bicho, São Jorge/Ribeirão Galinha, Chão Tavares/Cruz Grande.
At all areas, approximately the same number of interviews was attempted; the nth houses (as a function of the number of households in the area) were approximately sought; second visits were tried/scheduled when possible; non-respondents' basic characteristics were registed by interviewers.
The in-island transportation of the team (including myself) was made by buses, and assorted collective taxis ("Hiaces") in Praia, by a rented Hiace in the interior of Santiago (Mário), by rented Hiace in Santo Antão (Pedro), and by rented Hiace in Fogo, outside Sáo Filipe (João Francisco). The good-will of these drivers is ackowledged.
Checking of each interview of the day was performed every night, with enumerator/individual recommendations ready the next day. Battery re-charge for all machines was also done every night.
Stage 5: Climbing Mount Fogo
After wrapping up the survey in São Filipe, paying wages, being sure no improvements on the data could be done, breathing, I have decided to commemorate by climbing to Mount Fogo, the vulcano that provides the name to the island, at 2829m - 2h30m from Chã de Caldeiras on the way up (about 1000m in altitute), 30m on the way down, running.
Some basic facts on the country:
Cape Verde (CV) is a 9 (inhabited) -island country off the coast of Senegal in the Atlantic Ocean.
It has around 441 thousand inhabitants (2000, WDI2002).
CV, a portuguese speaking country, turned independent from Portugal, its 500-year colonizer) in 1975;
Its first political regime was soviet-like (as part of a common trend in Lusophone Africa).
Democratization was initiated in 1989, first free elections happened in 1991.
Aid per capita: USD 225 in 1985, USD 307 in 1995 (accompanying IMF and World Bank -sponsored economic reforms).
Growth of the economy from 1987:
GDPpc USD 4299 in 2000, current prices, PWT2002;
Ranked 81 in 134 countries in terms of GDPpcPPP in 2000, 1996 prices, PWT2002.
5.2% average growth GDPpc in 1987-2000, WDI2002.
Some links to information on Cape Verde:
Cape Verde at the World Factbook (CIA).
http://www.caboverde24.com: general portal.
http://www.portaldecaboverde.com: general portal (in Portuguese).
BACK TO THE MAIN PAGE