Tamara Last MSc, VU University Amsterdam
12 May 2015

Download this article in PDF format.


The descriptions below provide a summary of the methodology of the Database – the major decisions taken in designing, conducting and presenting the research. A detailed evaluation of the methodology is the subject of a forthcoming scientific paper.



The period covered by the Deaths at the Borders Database is 1/1/1990 to 31/12/2013, reflecting the period under study by the broader research project. As a consequence, a few cases from the late 1980s – in Cadiz region, Spain, for instance – have been missed. Nonetheless, the period covered captures the years in which ‘border deaths’ became a phenomenon in the southern EU border regions.

The definition of a ‘border death’ for the purposes of the Deaths at the Borders Database is: People who have died attempting to migrate irregularly to Europe by crossing the southern external borders of the EU without authorization, whose bodies were found on or brought to the territories of Spain, Gibraltar, Italy, Malta or Greece. This definition excludes people whose bodies were found or brought to territories outside the southern EU Member States, such as Turkey, Egypt, Libya, Tunisia, Algeria, Morocco, Mauritania, or Senegal. It also excludes persons who have gone missing, whose bodies have never been found.

The geographical scope of the Deaths at the Borders Database is shown in Figure 1. Data was collected in Malta and Gibraltar. In Spain, data was collected along the southern coasts from the Portuguese border to Valencia, from the coasts of the Baleares Islands and the Canary Islands, and from the Spanish enclaves in Morocco (Ceuta and Melilla). In Italy, data was collected from the coasts of Sicily (as well as inland in the south of Sicily, for reasons that will be explained in more detail under the country strategies section below), from the southern coasts of Calabria and Sardinia, from the southern and eastern coasts of Puglia up to and including Foggia and the international port of Ancona, and from the major international port of Naples. Finally, data was collected in six designated regions of Greece: the Cyclades and eastern coasts of Evoia, the Dodecanese, Crete, the North Aegean islands, and the Greek side of the land borders with Albania and Turkey (aka Evros). Together these areas were identified as encompassing all locations within the southern EU Member States where irregular migrants have arrived by crossing an external border as well as possible areas where dead bodies may be brought to the territory by the sea.

Figure 1: Map representing where data was collected for the Deaths at the Borders Database

Methodology Graph1


The Deaths at the Borders Database is sourced from the state death management systems in Spain, Gibraltar, Italy, Malta and Greece. A death management system is a term used here to encompass the actors, infrastructure, procedures and regulations involved when a dead body is found, recovered, investigated, recorded and buried. Death management systems leave a bureaucratic paper trail which provides information about who died, when, where and how they died, and what became of their bodies.

We used death certificates primarily to access the information gathered by death management systems in southern EU countries. A death certificate is the official, public record of a person’s death[1], and provides legal proof of death for the purposes of inheritance, remarriage, child custody, ancestry and – as a general rule – burial. Death certificates are made and archived by civil registries (stato civile in Italy, registro civil in Spain, public registry in Malta and Gibraltar, and ληξιαρχεία (lixiarcheía) in Greece).

Pilot studies in Greece, Malta, Italy, Gibraltar and Spain demonstrated that in general death certificates are:

  • the easiest document to access because they are archived in civil registries in each municipality and because they are public records;
  • the most reliable document issued in the management of a dead body in the sense that it is highly irregular for a death not to be registered and virtually impossible for a death to be registered more than once;
  • a useful summary of the investigation into the death and identity of the deceased.

Two other documents were consistently collected as a source of data for the Database: In Spain, cause of death is – since 1994 – recorded separately from the death certificate in a legajo, a document issued by the court of first instance overseeing the investigation into judicial deaths. We were able to access these legajos in most registries in Spain. In the Sicilian province of Agrigento, Italy, the civil registration system alone was found to be an insufficient source and therefore cemetery registers were also searched throughout this province. There were also circumstances in which other documents such as coroner’s reports, coast guard and police reports, burial permits and judicial orders, were made available to the research or sought out to supplement death certificates for particular reasons (see the Common Methodology).

All the information for each entry in the Database can be traced to its original official source(s).


Data was collected from 563 civil registries and additional offices by 12 researchers and Tamara Last between April 2014 and February 2015. Orcun Ulusoy provided logistical coordination, while Tamara Last supervised the progress of the field researchers and compiled the data.

The field researchers were trained in a 2-day Methodology Workshop held in Athens in April 2014, to ensure consistency in collection of data. The data collection stage of the project was closed with a Debriefing Workshop held in Madrid in February 2015, providing an opportunity to share experiences and reflect on challenges and the wealth of qualitative material that had been gathered from interactions with actors in the course of data collection.

Country Strategies

While Malta and Gibraltar have one centralized registry each, there are thousands of registro civil in Spain, stato civile in Italy and lixiarcheía in Greece and so it was necessary to develop country strategies to determine where to go to collect data.

In Spain, we searched all civil registries along: all coasts of the Canary Islands, all southern coasts of the Baleares Islands, all southern coasts of mainland Spain from the Portuguese border to Valencia, and the autonomous cities of Ceuta and Melilla. Strict rules about registration made it highly unlikely if not impossible that sea-border deaths would be registered in any non-coastal jurisdiction.

In Italy, we selected provinces which have received migrants by sea. The standard practice was to then search all civil registries along the coasts of these provinces, but inland civil registries could be included on the basis of a 1-2 day pilot study conducted in the provincial capital with procura, prefettura, coast guards, health authority, police, cemeterial office, etc.

In Greece, six regions were identified as being the first port of call for irregular migrants at some time during the study period. Each region was then allocated to different field researchers, who familiarized themselves with that area, gathering advice from key actors such as coast guards, local officials, and fishermen about which civil registries might be excluded on the basis that there was very little to no chance that border deaths would be registered there.

The Common Methodology

The Deaths at the Borders Database was conceptualized and designed as a single comprehensive database; thus it was important that data was collected in a uniform way across all five countries. As there were so many researchers involved, a Common Methodology was developed to standardize as far as possible the way in which data was collected.

The Common Methodology for the Deaths at the Borders Database consisted of:

  • a set of tools (codebook, database template, and logbook to record data collection)
  • a working definition of a ‘border death’ to guide the researchers in which death certificates to collect
  • a step-by-step manual (Instructions for Field Researchers) for collecting data from archives using the codebook and completing data entry in a standardized database template


Wherever possible researchers searched directly through archives of death registry books so as to be able to search through all death certificates issued from 1/1/1990 to the date on which they visited the registry. They collected the data for each person in a separate copy of the codebook form and, where allowed, took copies of the original in order to be able to double-check the data collected.

Civil registries do not maintain a special index of migrant deaths, nor is it common to indicate on the death certificate that the person was an unauthorised migrant (although this does happen). Therefore, identifying border deaths in death registries is a matter of deduction. Researchers automatically dismissed death certificates stating European nationalities, or if the place of birth or last residence was in Europe; they used the information on the death certificate and a process of elimination to determine whether each case of a non-European, non-resident was a border death or not. Usually this involved checking the place of death, the age and descriptions of the deceased.

Researchers completed a logbook for each civil registry they visited, in which they recorded the technique they had employed to search quickly through the archives of registry books at that registry, any restrictions on data collection from that registry, and any deviations from the Common Methodology.

There were two kinds of deviations from the Common Methodology in response to the logistics of data collection and problems discovered in the field:

  1. Dealing with restrictions which compromised some aspect of the Common Methodology in (a) situations in which information/recording was restricted, or (b) situations in which access to archives was restricted.

Solution: Researchers negotiated compromises that would least impact the data collected.

Effect: Minor, if any, impact on the content of the Database; slightly higher risk of having missed a case.

  1. Finding alternative official sources to death certificates in (a) situations in which access to registries was denied, and (b) situations in which death certificates were suspected of being insufficient for recording border deaths.

Solution: Wherever possible, the most suitable alternative official documents (secondary ‘access points’ to the same death management system) were used to substitute or complement death certificates.

Effect: Differences in the kind of information available as compared with death certificates; and a few geographical gaps in the Database (in Spain: San Sebastian de la Gomera, Las Palmas de Gran Canaria, Palma de Mallorca, Marbella, Villajoyosa, El Ejido; in Greece: Soufli for the period 1990-2000).

Ethical Considerations

Collecting data from death registries and secondary sources exposed a considerable amount of personal and sensitive information. Only the researchers who collected the data and the members of the research team at VU have access to the original source of the data. Any copies of original documents will be destroyed upon completion of the project when there is no longer any scientific use for them. Deceased persons are excluded from the scope of EU data protection law, but the public version of the Database has been anonymized and all references to document numbers removed to protect the privacy of the deceased and their families.

Although death registries are public records, the exact standing of researchers in terms of their right to access civil registry archives is not crystal clear and local officials have different perspectives on this issue. The field researchers were instructed to follow the procedures they were asked to follow by each registry, and to respect and accommodate as far as possible the concerns of the local authority they were dealing with. As there was variation in the interpretations of rights of access, however, the field researchers were only expected to follow the regulations and conditions asked of them by the particular registry they were dealing with.


Upon completion of data collection, all cases collected were reviewed to (a) check the data recorded against the copies of original documents, wherever possible; (b) to determine whether each case fit within the project’s definition of a border death; and (c) to determine the degree of certainty that each case was a border death.

A common list of criteria was developed and tested on all national datasets for classifying cases under three categories: confirmed, likely, and possible. Two other categories (unlikely, and excluded) were deleted from the public version.

Confirmed cases (certainty level 1), include: cases confirmed to be border deaths in the documents or by the local authorities involved; cases in an incident we found involving more than two deceased persons, with nothing to suggest that it is not a border death; or cases matched with an incident recorded in UNITED’s List of Deaths or by the Fortress Europe blog.

Likely cases (certainty level 2), include cases in which (a) the nationality/race is associated with irregular migration in that region and/or the age is between 18-40 years; (b) the place of death is appropriate for a border death (including vague places strongly associated with border deaths, such as simply “Tarifa”); and (c) the cause of death is common for border deaths. Drowning, hypothermia, suffocation, dehydration, starvation are common causes of death for migrants at sea borders, while hypothermia, suffocation, dehydration, starvation, vehicle accidents near border-crossing points and violent deaths are common causes of death for migrants at land borders. In Italy the cause of death was rarely recorded in death certificates, so in Italian cases for the purpose of establishing certainty the cause of death was assumed to be drowning if the body was found in the sea.

Possible cases (certainty level 3), include: cases that fit some of the criteria of likely cases but some vital information is missing, unusual or vague; Spanish cases that met the criteria of likely cases but were not investigated by a judicial body and the place of death is not a hospital; cases in which only the place of death is known and the place is at sea/on a beach/very close to a land border, and there were no details to suggest that it is not a border death; cases in which the place of death is a hospital, the nationality or presumed ethnicity/race is associated with irregular migration in that region, and the date of death is less than a month after an incident recorded in UNITED’s List of Deaths or the Fortress Europe blog in the same geographical area as the hospital; and cases of unidentified persons with insufficient or no information to determine whether it is a border death or not.


[1] Death is one of the three ‘vital events’ of a person’s life[1] registered by local state authorities since the second half of the 19th century.