Measuring statelessness: A study of the Pemba - 2016
Population and Housing Census [hh/popcen]
The survey of the Pemba was an attempt to reach all households in Kenya with links to Pemba in Tanzania. It was conducted in the two counties of Kilifi and Kwale on the coast, north and south of Mombasa, respectively. According to information from village elders familiar with the Pemba community in Kenya, most of the Pemba population resides in these two counties. While there are some Pemba residents in Lamu, the security situation prevented data collection there. Further, a few Pemba are believed to live in the city of Mombasa and elsewhere in the country. But due to lack of further information, no data were collected in Mombasa or elsewhere.
The objectives of the full survey, conducted in August 2016, were:
1. To establish the number and characteristics of the Pemba living in Kenya, including their arrival period in Kenya, nationality and their problems;
2. To make recommendations for the issuance of the documentation that is required for those who apply for citizenshiop by registration
Kind of Data
Census/enumeration data [cen]
Unit of Analysis
v2.1: Edited, anonymous dataset for licensed distribution.
Individuals: Demographics and citizenship status.
Household: Demographics and ownership of personal documents
Housing, Land and Property
Livelihood & Social cohesion
Pemba of Kenya
Kwale and Kilifi counties, Kenya.
The total number of households with links to Pemba in Tanzania, in Kilifi and Kwale counties.
Producers and sponsors
Norway Refugee Council
Kenya National Bureau of Statistics
A household mapping exercise was conducted in Kilifi and Kwale to identify Pemba households and to make it easier to locate them on the ground. The mapping was done from 4 to 12 August 2016 by a team from UNHCR Kenya office and KNBS.
The mapping in each village commenced with a visit to the chief's office, who put the team in touch with the village chair. The team explained the purpose of its visit to the village chair and began the mapping exercise. The importance of involving the chiefs and village chairpersons is that they are well connected, recognised and trusted by residents in their communities. The same procedure is followed by KNBS when they are mapping for sample surveys and censuses.
The team established physical boundaries of the area to be mapped, located the boundaries on the map and then identified and listed the Pemba households within the enumeration boundary. A Pemba household, in this context, is one identified by the informants as having at least one person with origins or links to Pemba. The links may include a person's spouse, parents or grandparents, who migrated to Kenya from Pemba or where a person has migrated from Pemba to Kenya.
The mapping team was followed by the village chair to the Pemba households, where the UNHCR and Haki Centre staff listed number of persons in each, while the KNBS staff marked the location of the household on the map. The entrances of identified Pemba households were marked in chalk with the letters HCR and a number starting at 001 to make it easier to find the houses during the enumeration. Since it seems to be generally well known where the Pemba live it was not considered stigmatising to mark their doors. During the feedback forums with the Pemba after the survey, there was no mention of stigmatization due to marking the door with chalk.
The maps were from the 2009 national housing and population census, purchased from KNBS. The team made lists with information about the location, number and size of each household. The mapping team visited 17 villages in Kilifi and Kwale (see Table 1 in Section 2.7). All villages visited were identified before the mapping exercise by key informants as locations being home to the Pemba of Kenya. The key informants were Pemba elders in different sub-counties previously identified for providing background information on the Pemba arrival and history in Kenya. In each sub-country, the chief, the assistant chief or the village chair also accompanied the team. In Kwale, 358 households were identified with 2,220 persons, and in Kilifi, 86 households with 558 persons.
The rate of non-response was low. Of the 452 households visited, visits to only 23 households can be categorised as non-response. A lot of effort was made to revisit non-responding households, using interviewers living nearby. Out of the 23 non-responsive households, 12 were not at home or there was no adult at home. There were 2 interrupted interviews, 7 refusals and 2 with no links to Pemba. In one household the respondent was not mentally stable enough to be interviewed, according to the enumerator.
Dates of Data Collection
Data Collection Mode
Data Collection Notes
Most of the enumerators had been recruited during the pilot study and the majority were retained. A few new enumerators were recruited for the main survey. Haki Centre, a non-governmental human rights organization based in Mombasa, played an important role in the recruitment. Some of the requirements were that the enumerators should be fluent in English and Swahili, that they had completed secondary school education, and that they had an interest in the overall objectives of the survey. Preference was given to interviewers living in Pemba villages. The recruitment also considered gender and diversity.
Before the pilot study, the enumerators were given a basic training on statelessness and why it occurs in Kenya, including the background of the Pemba of Kenya. A two-day training session for enumerators was held in Mombasa before the main survey (see the program in Annex 9). The session began with an introduction of the instructors and the participants and a discussion of administrative and practical arrangements with respect to the interviewers and supervisors, such as working time, allowances and transport arrangements. During the training, a presentation of the scope and purpose of the survey was given. The roles of the interviewers and supervisors were discussed and a list of tasks for interviewers and supervisors were distributed.
The questionnaire was developed before the pilot survey and revised during and after the pilot survey, based on the experience gained. The pilot survey was used to test the questions and to check for inconsistences and misinterpretations due to unclear concepts and definitions. The testing process also revealed some important themes that had been left out.
The structure of the questionnaire was altered, including the order of the questions and the introductory pages, to facilitate administration of the questionnaire.
Finally, the questionnaire was translated into Swahili. Both the English and Swahili versions were used in the survey, even though the English version was preferred by almost all interviewers. The two versions of the questionnaire are attached in Annex 4 and 5. Enumerators used the English questionnaire to frame the questions in the local and less academic version of Swahili.
The data were imported into a Statistics Analysis Software (SAS) file and validated. Several errors were identified during the validation process, both on how the data had been recorded by the interviewers in the field and how the data had been entered by the clerks.
There were particularly many errors in the entry of the variable “Relation to the household head” (Q.2). There were also many errors in the entry of the age of the household head, which was mostly due to errors in recording the right codes. A substantial amount of time was spent cleaning the data after the data had been entered, which included consulting many paper questionnaires. The quality of the survey data was significantly improved after the data entry revision. The data were analysed using both SAS software and Excel spreadsheets.
UNHCR (2016). Measuring statelessness in Kenya: A 2016 study of the Pemba. UNHCR Microdata Library https://microdata.unhcr.org