beneath the surface iom

8/6/2019 Beneath the Surface IOM

1/96

1

Beneaththesurface.Methodologic

alissuesinresearchanddata

colleconwithassistedtrac

kingvicms


2/96

This is the rst paper in the IOM and NEXUS Instute Human Tracking Research Series.

This publicaon was made possible through support provided by the United States

Department of State, under the terms of Grant No. S-SGTIP-09-CR-0070. The opinions

expressed herein are those of the authors and do not necessarily reect the views of the

United States Department of State.

_______________

IOM is commied to the principle that humane and orderly migraon benets migrants

and society. As an intergovernmental organizaon, IOM acts with its partners in the

internaonal community to: assist in meeng the operaonal challenges of migraon;

advance understanding of migraon issues; encourage social and economic development

through migraon; and uphold the human dignity and well-being of migrants.

_______________

Publishers: Internaonal Organizaon for Migraon

17 route des Morillons

1211 Geneva 19

Switzerland

Tel: +41.22.717 91 11

Fax: +41.22.798 61 50

E-mail: [email protected]

Internet: hp://www.iom.int

NEXUS Instute

1701 Pennsylvania Avenue N.W. Suite 300

Washington, D.C. 20006

United States of America

Email: [email protected]

Internet: www.NEXUSInstute.net

_______________

2010 Internaonal Organizaon for Migraon (IOM)_______________

All rights reserved. No part of this publicaon may be reproduced, stored in a retrieval

system, or transmied in any form or by any means, electronic, mechanical, photocopying,

recording, or otherwise without the prior wrien permission of the publisher.

04_11


3/96

Beneath the surface.

2010

Methodological issues

in research and

data collection

with assistedtrafficking victims

Rebecca Surtees and Sarah Craggs

A research cooperaon between

IOM and NEXUS Instute


4/96


5/96

3




kingvicms

Foreword by NEXUS Institute

In the late 1990s and early 2000s, during deliberaons of the seminal United

Naons An-Tracking Protocol and the early draing of naonal laws that

followed, only a handful of reports addressing contemporary forms of slavery

had been wrien. This has changed. Today hundreds of studies on human

tracking are available with more published each month.

The expansive growth in the body of literature on this subject within this relavely

short meframe is a posive development and is welcomed by all who have

dedicated themselves to ending modern slavery. Government leaders, donors,

service providers and other professionals, faced with the myriad of complexies

and variaons of human tracking, connue to stress the importance of

research to their work and their desire for more data collecon and research.

However, even though policymakers and praconers want and need research

they can use, many conclude, with some dismay, that only a small percentage

of the numerous reports available provide sucient empirical rigour and qualityof analycal insights needed to help guide more informed decisions capable

of producing more eecve an-tracking results. It is clear that despite a

substanal investment of me and resources to produce more and more studies

around the world, few end up being useful to government ocials and an-

tracking professionals in their work.

Implementaon of more eecve pracces to end modern slavery will require

beer underlying research, analysis and evaluaon. Achieving this, however,

will not be possible in the absence of a foundaon of sound data systemacallycollected to serve specic pracce-based research objecves.

This report focuses on approaches to collecng data about human tracking

that underlie a large segment of research produced since the UN Protocol

and, in doing so, reveals some of the key reasons that research generally has

not provided a clearer path to more eecve acon for policymakers and

praconers. It examines how current approaches to the collecon and use

of data about human tracking, while helpful for certain purposes, fall short

of what will be needed to achieve a new generaon of higher quality research

and analysis capable of helping to produce transformave results in addressing

human tracking.


6/96

4

Beneathth

esurface.Methodologicalissuesinresearchanddatacolleconwithassistedtrackingvicms

The NEXUS Instute is a policy and research centre dedicated to helping an-

tracking leaders develop and implement best pracces to improve our collecve

ability to prevent human tracking, assist and protect vicms of tracking as

well as end the impunity of the trackers. The issues spotlighted in this report

by the authors, Rebecca Surtees and Sarah Craggs, serves as a valuable starng

point for a wider discussion among governments, donors, academic instuons

and others who support, conduct or use research concerning human tracking.We encourage governments and others to consider in greater depth how best

to operaonalise the important ndings of this report. Implemenng new

methods to systemacally acquire data that can reveal a fuller picture of human

tracking and enable pracce-based research will provide a rmer foundaon

of understanding to act upon to achieve an end to slavery in our me.

The NEXUS Instute appreciates our valued collaboraon with our partner IOM

on this project. NEXUS is especially grateful to the U.S. Department of States

Oce to Monitor and Combat Tracking in Persons for supporng and makingthis project possible.

Stephen Warnath

Founder and Chairman

NEXUS Instute


7/96

5




kingvicms

Foreword by IOM

In 2000, as countries met to inaugurate the United Naons An-Tracking

Protocol, the IOM oce in Prisna took the important iniave to launch an

internal human tracking case management tool. Its aim was to manage the

assistance provided to tracking vicms as well as to improve the knowledge

base through the collecon of informaon about tracked persons and

the tracking process. The tool, today known as the IOM Human TrackingDatabase, was soon adopted within the South-Eastern Europe region and during

the years that followed, was rolled-out globally within IOM and to our naonal

counterparts.

At the same me, there was an increased interest in the research potenal of

the IOM database, given the detailed and unique degree of primary data being

collected. This move towards assessing the research potenal of the database

and the associated data set on assisted vicms allowed for new partnerships to

be formed, including the partnership with NEXUS Instute.

As the an-tracking community reects upon the ten years that have passed

since the signing of the of the United Naons An-Tracking (Palermo) Protocol,

there is a need to assess the impact of our research and data collecon eorts.

While progress has been made, there remains a real need to further advance our

understanding of human tracking and thus our associated research and data

collecon eorts, by ensuring that a more precise and encompassing picture of

tracking is presented. This requires us to not only reect upon the cases that

have been assisted by IOM but to equally learn from those who have eitherdeclined IOMs assistance or failed to be idened. This further requires eorts

to be undertaken in all corners of the world to ensure that all tracked persons

regardless of sex, age or type of exploitaon are granted the protecon they

need.

This paper notes that there is a fundamental need for more accurate, quality,

and in-depth data, informaon and research on all aspects of the tracking

phenomenon if we are to successfully and eecvely combat human tracking

and end the exploitaon of migrants. On this note, we encourage governments,

policymakers, researchers, internaonal organisaons, and other actors to take

note of the pernent and mely ndings herewith presented in this report; and


8/96

6

Beneathth


to bolster their eorts to support the producon of improved research and data

in order to beer understand and combat human tracking.

IOM has been working to end the exploitaon of migrants and to combat human

tracking since 1994. During this me, the organizaon has assisted tens of

thousands of tracked persons around the world. We hope that by sharing our

experiences and lessons learned we will collecvely be able to protect moremigrants in vulnerable situaons, and to end their exploitaon.

IOM equally appreciates our valued collaboraon with our partner NEXUS on this

project. IOM would further like to extend our connued appreciaon and thanks

to the U.S. Department of States Oce to Monitor and Combat Tracking in

Persons for their generous support to the database over the years and to this

project in parcular.

Irena Vojackova-Sollorano

Director

Department of Migraon Management

IOM Headquarters


9/96

7




kingvicms

Acknowledgements

This research paper is the rst in a joint IOM/NEXUS research series which will

seek to augment and enhance the current knowledge base on human tracking

by conducng targeted research as well as to consider and test methods and

approaches to tracking research in dierent sengs and in response to

dierent situaons. The paper (and the project more generally) are generously

funded by the US Department of State Oce to Monitor and Combat Tracking(G/TIP). The idea for the joint project was conceived over a number of years

of discussion and cooperaon between the authors as well as IOM and NEXUS

Instute and, as such, it is important to acknowledge the role of Richard

Danziger, then Head of IOMs Counter Tracking Division in Geneva (now Chief

of Mission in Sri Lanka) and Stephen Warnath, Chairmen and founder of NEXUS

Instute in Washington. G/TIP has also been very supporve of this research

cooperaon and partnership between IOM and NEXUS; their support must also

be acknowledged here.

This paper benets from the experse of a wide range of respondents from the

elds of research as well as service provision. Our sincere thanks to all those

who parcipated in this research study. We would also like to sincerely thank

the tracked men and women who shared their experiences of being involved

in research with us. Their ideas and suggesons are an important part of this

paper and provide important insight into the issue from the vicm perspecve

including what might be done dierently (and beer) in future.

The inputs and suggesons of our seven peer reviewers has enhanced andstrengthened the paper. Our thanks to Anee Brunovskis (Senior researcher,

Fafo in Norway), Elzbieta Gozdziak (Research Director, Instute for the Study

of Internaonal Migraon, Georgetown University), Frank Laczko (Head of IOM

Research in Geneva), Anh Nguyen (Deputy Chief of Mission/Senior operaons

coordinator, IOM Kyiv), Ana Eva Radice (Head of unit, IOM Brussels, former

Chief of Mission, IOM Skopje), Tal Raviv (Regional Project Development Ocer,

IOM Nairobi) and Lisa Rende-Taylor (Chief Technical Advisor, United Naons

Interagency Project in Bangkok).


10/96

8

Beneathth


In addion, within IOM and NEXUS here are a number of people without whom

this study would not have been possible. Our thanks to Anvar Serojitdinov,

Project Ocer (Migrant Assistance Division, IOM Geneva), who has been an

important member of the project team and undertaken a wide range of roles and

acvies. Thanks also to former IOM interns, Elizabeth Murphy and Sharmaake

Sabrie. Finally parcular thanks to Richard Danziger and Stephen Warnath for

their expert guidance and support as well as their inputs and direcon in thedevelopment of this paper.

Rebecca Surtees Sarah Craggs

Senior Researcher Researcher and Programme Coordinator

NEXUS Instute IOM Geneva

December 2010 December 2010


11/96

9




kingvicms

Table of contents

Foreword by NEXUS Instute ............................................................................. 3

Foreword by IOM ............................................................................................... 5

Acknowledgements ............................................................................................ 7

Table of contents ................................................................................................ 9

Acronyms and abbreviaons ............................................................................ 11

Execuve summary .......................................................................................... 131. Introducon ................................................................................................ 15

2. Methodology and data collecon ............................................................... 19

Interviews with key informants ................................................................... 19

Survey of IOM missions ............................................................................... 20

Analysis and review of the IOM tracking database .................................. 20

Literature review ......................................................................................... 21

Peer review process .................................................................................... 21

Research limitaons .................................................................................... 21

3. Considering the IOM approach ................................................................... 23Data from assisted vicms methodological strengths and limitaons ..... 27

4. A global approach? Data quality and comparability across dierent

terrains ........................................................................................................ 29

The tools used ............................................................................................. 30

The quesons asked, the variables used ..................................................... 32

The language used ...................................................................................... 35

The professionals involved .......................................................................... 38

How data is managed .................................................................................. 40

Missing data, missed meaning? .................................................................. 41Comparability across contexts..................................................................... 42

Standardised but not comparable? ............................................................. 44

5. Who is collecng data? The role of researchers and service providers ...... 45

Blurred roles, blurred boundaries? ............................................................. 47

Rigour, objecvity and independence ......................................................... 50

Resources and me ..................................................................................... 51

Dierent role, dierent goals? .................................................................... 52

Mixed roles? ................................................................................................ 54


12/96

10

Beneathth


6. Being representave? Challenges in obtaining representave samples

of tracking vicms .................................................................................... 57

Are assisted tracking vicms representave? .......................................... 57

Is an organisaons caseload representave? ............................................. 61

Representavity in tracking research and data collecon ....................... 64

7. What quesons are asked and why? Assumpons, biases and agendas

in tracking research and data collecon .................................................. 67What assumpons? What biases? .............................................................. 67

Agenda and ideology ................................................................................... 70

The implicaons? ........................................................................................ 71

Vicms perspecve .................................................................................... 73

8. Where do we go from here? Some thoughts, some recommendaons ..... 77

Data quality and comparability ................................................................... 78

Relaonships and boundaries between service provision and research..... 78

Representavity in tracking research and data collecon ....................... 79

Assumpons, biases and agendas in research quesons ............................ 80Going forward? ............................................................................................ 80

9. Bibliography ................................................................................................ 83


13/96

11




kingvicms

Acronyms and abbreviations

AT An-tracking

BiH Bosnia and Herzegovina

BNRM Bureau of the Dutch Rapporteur on Tracking in Human Beings

CT Counter-tracking

CTM Counter-Tracking Module

EC European CommissionECPAT Global network of organisaons and individuals working

together to eliminate child prostuon, child pornography and

the tracking of child for sexual purposes.

EU European Union

FSU Former Soviet Union

GAO United States Government Accountability Oce

GO Governmental Organisaon

G/TIP U.S. Department of States Oce to Monitor and Combat

Tracking in PersonsHQ Headquarters

ILO Internaonal Labour Organizaon

IO Internaonal Organisaon

IOM Internaonal Organizaon for Migraon

MoI Ministry of Interior

NGO Non-governmental organisaon

OSCE Organisaon for Security and Co-operaon in Europe

SEA South-East Asia

SEE South-Eastern EuropeSW Support workers

THB Tracking in human beings

TIP Tracking in persons

UN United Naons

UNAIDS United Naons Programme on HIV/AIDS

UNDP United Naons Development Programme

UNHCR United Naons High Commission for Refugees

UN.GIFT United Naons Global Iniave to Fight Human Tracking

UNIAP United Naons Inter-Agency Project on Human TrackingUNICEF United Naons Childrens Fund


14/96

12

Beneathth


UNODC United Naons Oce on Drugs and Crime

UNOHCHR United Naons Oce of the High Commission of Human Rights

US United States

USAID United States Agency for Internaonal Development

VoT Vicm of tracking

WHO World Health Organizaon


15/96

13




kingvicms

Executive summary

Recent years have seen increased aenon to research and data collecon on the

issue of human tracking. Many early tracking studies sought to understand

the nature and scope of human tracking. More recently, tracking studies

have considered a wide range of topics including methodological and ethical

issues when conducng research and data collecon. There has also been a

spate of data collecon iniaves as a means of tracking and, by implicaon,beer understanding tracking in persons. Nevertheless, there is some debate

about the current quality and integrity of research and data collecon in the

tracking eld.

This paper draws upon one parcular research and data collecon approach

the IOM human tracking database as a means by which to discuss current

data collecon and research eorts and, equally, as a lens to draw some lessons

and suggesons for future research and data collecon iniaves.

As the paper discusses, much of the current knowledge base on tracking

is drawn from assisted tracking vicms (that is, vicms who have been or

are currently being assisted). And research and data collecon with assisted

tracking vicms, like that undertaken by IOM and other service providers,

has many strengths. It can shed light on a range of issues including risk and

vulnerability factors, the needs of dierent groups of tracking vicms (e.g.

men, women, children, vicms of labour tracking and sex tracking); the

gender dimensions of tracking; details of the tracking process and, albeit it

to a lesser extent, the perpetrators involved, their modus operandi, the routesused and so on. This approach also allows counter-tracking professionals to

idenfy emerging trends and paerns in real me, which, on the one hand,

means the data is current and, on the other hand, allows the data to inform

policy and programmac response.

Nonetheless, as with all research methods, the approach suers from some

challenges which must be understood for the research ndings to be used

eecvely and appropriately. This paper outlines some of these issues, centring

around the following four themes:

A global approach? Data quality and comparability across dierent terrains

Who is collecng data? The role of researchers and service providers


16/96

14

Beneathth


Being representave? Challenges in obtaining representave samples of

tracking vicms

What quesons are asked and why? Assumpons, biases and agendas in

tracking research and data collecon

In exploring these topics, we outline some of the methodological issues which

arise when collecng data about assisted tracking vicms (and their trackingexperiences) through service providers and in the context of an-tracking

assistance programmes. While it is not the aim of this paper to present an

analysis of actual data contained within the IOM database, key examples from

the database are used to illustrate these themes and issues.

This paper equally seeks to draw some lessons for future research and data

collecon iniaves. In sum, while IOM and other actors have made important

steps in drawing upon assisted vicm data for research purposes including for

trend analysis and to idenfy emerging issues there are opportunies to furtheradvance the analysis by ensuring that a more rounded picture of tracking is

presented. This means being mindful of key issues like, issues of data quality and

comparability, the context of data collecon, the representave nature of the

data and biases and assumpons in the research process. This equally involves

moving beyond an analysis of data collected from only one source and one group

of vicms to include other data sources, employing mulple methodologies.

It is crical that service providers and researchers increasingly explore and

present both the strengths and limitaons of data and research in order thatpolicymakers and praconers can make informed decisions about which data

they use and how in the design, implementaon, monitoring and evaluaon of

their an-tracking responses. It is hoped that by making explicit some of the

limitaons of data collecons with assisted vicms, and more specically what

can be learned from the example of the IOM database, that research drawn

from these data can be read and understood in context, including what this

informaon does (and does not) tell us about tracking.

This paper is the rst in a series of research papers being prepared jointly byIOM and the NEXUS Instute and funded by U.S. Department of States Oce to

Monitor and Combat Tracking in Persons (G/TIP). The overall objecve of the

themac research series is not only to seek to augment and enhance the current

knowledge base on human tracking but also to consider and test methods

and approaches to tracking research in dierent sengs and in response to

dierent situaons.

Key words

An-tracking data collecon; assisted tracking vicms; bias; casemanagement; comparability; data quality; ethics; IOM human tracking

database; methods; representavity; service providers; tracking research.


17/96

15




kingvicms

1. Introduction

Recent years have seen increased aenon to research and data collecon

on the issue of human tracking. Many early tracking studies sought to

understand the nature and scope of human tracking. More recently, tracking

studies have considered a wide range of topics including methodological and

ethical issues when conducng research and data collecon. In addion, there

has been a spate of data collecon iniaves, albeit dierent in scope, nature,and locaon, as a means of tracking and, by implicaon, beer understanding

tracking in persons. These eorts are also intended, in many cases, to help

account for focus and expenditures on an-tracking eorts.

Nevertheless, there is some debate about the current quality and integrity of

research and data collecon in the tracking eld. Within research and policy

circles, there have been a number of recent iniaves and publicaons which

have sought to assess the current state of tracking research, including the use of

dierent methodologies and approaches.1

Moreover, many instuons, agenciesand universies have been engaged in discussions of and projects on research

and data collecon, including the establishment of naonal rapporteurs on

tracking or equivalent mechanisms.2 And so, a decade since the signing of the

Palermo Protocol,3 researchers, praconers and policy makers are increasingly

taking stock of what is known about human tracking and what sll needs to be

known to more eecvely undertake tracking research and, by implicaon, to

design and implement more eecve an-tracking programmes and policies.

This necessarily involves a consideraon of research methodology and pracce

as a means by which this greater understanding can be realised.

1 See, for example, Aghazarm, et al 2008; Andreas & Greenhill 2010; Andrees & van der Linden 2005; Brennan

2005; Brunovkis & Surtees 2010; Clawson et al. 2006; Cwikel & Hoban 2005; David 2007; De Cock 2007;

GAO 2006; Goodey 2008; Gozdziak & Bump 2008; Gozdziak & Collet 2005; Gragmena & Laczko 2003; IOM

et al. 2009; Kelly 2005 & 2002; Laczko 2005; Laczko & Gozdziak 2005; Rogers et al 2010; Savona & Stefanizzi

2007; Steinfa 2003, Steinfa et al. 2002; Surtees 2007c; Tyldum 2010 & 2009; Tyldum & Brunovskis 2005;

Tyldum, Tveit & Brunovskis 2005; UNIAP 2008; Young 2009. In addion, there have been numerous naonal,

regional and internaonal seminars, conferences and meengs to discuss research methodology and data

collecon eorts.2 As a non-exhausve list, this includes: David 2007; Doridge 2010; Folden et al. 2007; ILO 2009; IOM et al.

2009; Machado et al. 2007; OSCE 2009; Surtees 2009, 2007a, 2005; UNODC 2009; Vermeulen & Paterson

2010; Vermeulen at al. 2006; Weiner & Hala 2008. Many naonal enes, such as naonal rapporteurs orequivalent mechanisms, also produce annual reports on the state of tracking in the respecve country

(e.g. BNRM 2010).3 Ocially the United Naons Protocol to Prevent, Suppress and Punish Tracking in Persons, especially

Women and Children, signed in 2000.


18/96

16

Beneathth


To this end, this paper aims to discuss some of what we have learned from our

involvement in dierent an-tracking research and data collecon iniaves,

with a view to moving forward. In parcular, we use our joint experience with

one parcular research and data collecon approach IOMs Counter-Tracking

Module (CTM)4 as the starng point for a discussion of current tracking

research and, equally, as a lens to draw some lessons and suggesons for futureresearch and data collecon iniaves. The CTM is IOMs global database

on vicms of human tracking. It collects data about individual tracking

vicms assisted by IOM and/or its partner organisaons, including informaon

on tracked persons background, recruitment, transportaon, tracking

exploitaon and assistance.5

The paper will consider how, to date, the IOM database has (and has not) been

eecve in terms of tracking research and data collecon and, equally, what

methodological, praccal and ethical issues arise from this parcular approach.While it is not the aim of this paper to present an analysis of data contained

within the IOM database, examples from the IOM data sets will, where relevant,

be included. We will also consider what might be done to improve tracking

research and data collecon, suggesons which are specic to the IOM database

and also have broader relevance. It should be noted that many organisaons

employ approaches similar in scope and purpose to the IOM approach. It is,

therefore, hoped that this paper with its discussion of our experience of and,

at mes, our frustraon with the IOM database and other related data collecon

iniaves can contribute to the dialogue on an-tracking research and datacollecon. It is also intended that by making explicit the strengths and limitaons

of the IOM approach (and by implicaon, those of many other organisaons

and instuons), research drawn from it can be read and understood in the

appropriate light.

Certainly research and data collecon is vital in terms of beer understanding

human tracking and, equally, in designing and implemenng an-tracking

responses. The authors, and the instuons we represent, have been involved

in various studies, projects and iniaves with precisely these objecves.

6

At the

4 The database is referred to as the CTM within IOM. In the context of this paper we refer to the IOM traf-

cking database.5 The database is installed in 72 IOM missions globally. As of the end of December 2010, the system contained

data on nearly 16,000 registered IOM beneciaries in more than 85 source and more than 100 desnaon

countries. See also secon three for addional discussion of the IOM tracking database.6 We have each worked with the IOM tracking database at various stages of its development and in dier-

ent roles and capacies. The rst author has approached the database from a largely outside perspecve,

conducng independent research using dierent IOM data sets at dierent stages and in dierent ways. In

some cases, the data sets were exclusively IOM; in other cases, IOM data was part of a larger body of prima-

ry data. She has also conducted extensive eld research with tracked persons in Asia, Europe, the FSU and

Africa. The second author has worked with the IOM database from the inside cleaning and validangthe data, providing training and guidance to eld missions, data analysis for country and themac reports

and contribung to various arcles and documents based on IOMs approach to data collecon. She has

also conducted eld research in Europe, the FSU and North America. In addion, both authors have been

involved in various research discussions, meengs and seminars and have parcipated in dierent (naonal

and regional) data collecon iniaves, including those which have drawn on, learned from and also built

upon the IOM tracking database model.


19/96

17




kingvicms

same me, research and data collecon must be undertaken with cauon, care

and adequate aenon to aendant methodological and ethical issues. And,

as a number of researchers and studies have observed, there are substanal

problems with much research on human tracking and data collecon iniaves

which have, in many cases, been ill conceived and/or poorly implemented.

Nevertheless, there is much that can be learned from these past eorts and

iniaves. Equally, there is a need to move beyond simply taking stock of andcriquing exisng research and data collecon and oer some suggesons and

entry points in terms of how to go forward with this work.

This paper is not intended as a denive exploraon of methods in research and

data collecon with tracking vicms. Rather, it is a starng point for discussion,

illustrated largely through the discussion of IOMs very specic methodology and

approach to research with assisted tracking vicms, including the aendant

strengths and weaknesses. In doing so, it is also intended to encourage other

organisaons, instuons and individual researchers to openly discuss andexplore some of the issues and constraints they face in their research and data

collecon eorts as a mean of advancing both the methodological approach in

this eld and to enhance the rigour of tracking research and analysis. Finally,

the paper is intended to provide some guidance to policymakers and praconers

in terms of the strengths and limitaons of the current body of knowledge on

tracking, not least in terms of how it can and should be read, understood and

used in the design and implementaon of policies and programmes.

This paper is the rst in a series of research papers being prepared jointly byIOM and the NEXUS Instute and funded by U.S. Department of States Oce to

Monitor and Combat Tracking in Persons (G/TIP). The overall objecve of the

themac research series is not only to seek to augment and enhance the current

knowledge base on human tracking but also to draw upon IOM database and

data sets to consider and test methods and approaches to tracking research in

dierent sengs and in response to dierent situaons. Each paper in the series

will be a stand-alone study of a specic aspect of tracking and/or tracking

research. Preliminary analysis of data contained within the IOM database has

led to the selecon of research topics which will include, but are not limitedto, an analysis of research on trackers and tracking operaons, research

on tracking into the construcon industry, tracking and exploitaon of

shermen and seafarers and tracking for forced marriage. At the same me,

taken together, the studies will provide a strong foundaon for considering

not only the potenal of the IOM database toward providing fresh insights on

human tracking but, equally, what is (and is not) working in the eld of an-

tracking data collecon and research, including ideas and recommendaons

for ways forward, both in terms of what is known, what research is needed and

how to go about undertaking such studies.


20/96


21/96


22/96

20

Beneathth


Interviews were also conducted with 18 service providers from IOM and NGOs

who, in addion to their assistance and case management work, also collect

data about the tracking vicms they assist. This included 12 IOM sta from

seven countries in Europe, Asia, the FSU, Africa and Lan America who work or

have previously worked on tracking data collecon and research. In addion,

six NGO sta was interviewed from three countries in Europe, the FSU and Asia.

Service provider interviews focused on their experiences of conducng researchand data collecon with their programme beneciaries some with the IOM

forms, others according to their own tools and methods.

We also interviewed 21 tracked persons who were currently being or had

previously been assisted within an an-tracking assistance programme and, in

that context, had been involved in data collecon process as respondents. Five

tracked persons from Albania and one from Ukraine were interviewed in 2009;

15 tracked persons from Ukraine were interviewed in 2010. This included

both men and women and vicms of both tracking for labour and sexualexploitaon. The intenon was to learn from tracked persons how they had

understood and experienced the research and data collecon process, including

how they felt about this type of data collecon, how they felt about the tools

used and quesons asked, any concerns they had about the data collecon and

what could be done dierently (and beer) in future.

Survey of IOM missions

IOM counter-tracking focal points currently working with the IOM trackingdatabase and/or the accompanying screening and assistance interview forms

were globally surveyed according to a standard research quesonnaire. The

quesonnaire focused on current data collecon processes and protocols within

the mission and, where relevant, the nature and extent of any variaon on or

deviaon from the standard approach outlined in IOMs tracking database

protocols. Responses were received from 23 missions in Europe, North America,

Asia, the FSU, the Middle East, Africa and Lan America. This accounted for

about a third of the 72 missions which currently use or have used the CTM; there

are approximately 250 IOM missions worldwide.

7

We also drew on informaonand feedback shared previously by IOM missions in the context of an internal

evaluaon of the IOM database undertaken in 2006-2007 (Sgter 2006) as well

as in past and ongoing discussions with IOM eld missions.

Analysis and review of the IOM tracking database

This included a review of the database, checking for full data and high levels

of non-response, coding errors, outliers, inconsistencies, logic checks and an

7 Not all of these missions implement or are currently implemenng - an-tracking acvies. Further-

more, while IOM is present in 460 eld locaons, this does not necessarily refer to an oce as a physical

premise but to the presence of IOM sta.


23/96

21




kingvicms

assessment of the open elds and narrave informaon. While some of this was

done in the context of this paper, this exercise is also done regularly within IOM

(at the global and local level) as part of data maintenance and quality assurance.

In addion, queries were made in the database to answer specic quesons in

relaon to non-response, local adaptaons and changing paerns.

Literature review

A broad review was undertaken of tracking research methodology as well as

research methods and approaches which potenally overlap with the tracking

eld for example, methods used in studying migraon, prostuon, hidden

and elusive populaons and sgmased/marginalised populaons. This included

reviewing journal arcles, books, organisaonal reports and presentaons.

Peer review process

The paper was reviewed by seven peer reviewers, with extensive knowledge and

experience in tracking research including, in some cases, detailed knowledge of

the IOM database and methodology. This included four internal peer reviewers

(personnel who had previously or currently work for IOM) and three external

reviewers from research instutes, research projects and universies.

In addion, the paper was reviewed internally within IOM and NEXUS Instute

project teams by Richard Danziger, then Head of the IOM Migrant Assistance

Division and Stephen Warnath, Chair and founder of NEXUS Instute.

Research limitaons

This paper draws on a wide range of experiences and experse in considering,

the current state of tracking research and data collecon, through the lens

of the IOM approach. This includes not only a review of current research and

exisng IOM data sets but also interviewing a range of researchers and service

providers working on the issue as well talking with tracking vicms themselves.

Nevertheless, there are some limitaons in terms of the data collected, whichare discussed briey below.

Geography. There was a geographical bias in that more respondents were from

Europe, North America, the FSU and Asia than from Africa, Lan America and

the Middle East. To some extent this is a consequence of our own working

environments and contacts. However, it is also, at least partly, because a large

amount of tracking research is focused on or conducted by researchers in or

from Europe, Asia and North America (Aghazarm et al. 2008; Laczko 2005). One

survey of IOM tracking publicaons found that, between 2000 and 2004, 44per cent originated from within Europe and 35 per cent from within Asia-Pacic

(Laczko & Gozdziak 2005).


24/96

22

Beneathth


Moreover, IOM missions in SEE and FSU have the longest experience with the

IOM database and, as a result, the largest data sets. While there are pockets of

experience in other regions, it is by no means comparable. As a consequence,

the ndings and issues idened may be more reecve of these sengs than

less represented regions. Future papers in the series will seek to pay aenon to

this geographical bias by considering less studied countries and regions, to the

extent that this is possible.

Representavity amongst respondents. The intenon was to use the IOM

database as a lens to discuss methodological issues faced by those undertaking

research and data collecon with assisted tracking vicms. There are

consequently a limited number of interviews with organisaons which work

directly on data collecon in dierent sengs and do not partner with IOM.

While it was possible to conduct brief eldwork in Ukraine and Albania, this was

an add on to other eldwork projects, which means that there is insucient

aenon to (dierent and diering) eld perspecves and organisaons not

associated with IOM and its partners. There is also a geographic bias in that the

eldwork took place in one country in Europe (Albania) and one country in the

FSU (Ukraine).

Representavity amongst tracked persons. The paper includes the experiences

and opinions of tracked persons, both men and women and vicms of both

labour and sex tracking. However, because eldwork took place only in two

countries in the Europe and FSU region, this informs the extent to which vicmsexperiences are representave and/or globally relevant.

Language barriers. Telephone interviews were primarily conducted in English,

although Russian language interviews were possible due to language skills of

the project research assistant. Interpretaon was available in the context of

eldwork. Nevertheless, there is a bias toward researchers and service providers

who are procient in the English language. The literature review also reects this

bias toward English language sources which means that some valuable research

has been missed. Lack of funds for translang research prevented the inclusionof these foreign language studies.


25/96

23




kingvicms

3. Considering the IOM approach

IOMs tracking database collects single case data about assisted vicms as a

means of facilitang case management in the shorter term and facilitang data

analysis for research purposes in the longer term.Data is collected with two

disnct tools a screening interview form and an assistance interview form.8

Screening interviews, undertaken upon rst contact with the tracked person,

assess whether an individual was tracked according to the denion providedin the Palermo Protocol9 and also their immediate protecon needs. Assistance

interviews, undertaken once the individual has accepted IOM assistance,

document the vicms background, recruitment and transportaon, tracking

experience and their assistance and/or re/integraon needs, including what

service are (and are not) provided. The IOM database encodes data about

individual tracking vicms according to these standardised interview forms.

This standardised data is further supplemented by qualitave data from

interviews with tracking vicms, which allows for the documentaon of details

that fall outside of the standardised elds and adds depth to the informaoncollected. Interviews are undertaken by service providers, either an IOM sta or

an NGO or GO partner with whom IOM works in providing direct assistance.10

The IOM tracking database aims to consolidate into one database tracking

cases assisted by IOM missions and their partners globally. By mobilising the

links between eld missions (in origin, transit and desnaon countries), the

database aims to establish a standardised model for data collecon and data

sharing (allowing for local and cultural adaptaon through the inclusion of

open elds) while reducing the risk of double counng cases. The database alsofacilitates the cross-border transfer of data in a secure manner (Sgter 2006: 9).

IOMs rst counter-tracking data collecon system was started in 2000 in

Kosovo, UNSC resoluon 1244-administered Kosovo11 by IOM counter-tracking

sta who required a system to collect informaon about tracking vicms they

were assisng as well as returning and referring for assistance in countries of

8 Please refer to the IOM Direct Assistance Handbook for a copy of the screening form and a detailed descrip-

on of IOMs approach to screening and assistance. See IOM 2007.9 See Chapter Two of the IOM Direct Assistance Handbook for more informaon on how IOM determines

tracking status. See IOM 2007.10 The interview process and nature of data collecon will be discussed in more detail in subsequent secons.11 Hereinaer referred to as Kosovo/UNSC 1244.


26/96


27/96

25




kingvicms

while others required amendments in the system.15 In 2004, revisions were made

to the database and the old data set, which has been collected since 2000 (with

a very small number of cases having been idened in 1999), was transferred

and integrated into the new IOM tracking data management system.

In January 2006, the then IOM Counter Tracking Division16 hosted a meeng

on how to improve data collecon and how to make the database a more usefulresearch tool. The meeng was aended by relevant HQ and eld sta and

representaves from external instuons like NEXUS Instute and ILO. This was

in response to ongoing inputs, and oen frustraons, from eld missions about

the ease of the tools use and/or its suitability for their purposes. While some

issues were technical17 or linked to resources, others were linked to the approach

itself and included issues such as quesons which were deemed inappropriate

in some sengs, unsuitability for all forms of tracking and all types of vicms,

the large number of quesons and inconsistent or incomplete data. This meeng

was followed by an internal evaluaon of the database, nalised in October2006, to idenfy reasons for none or paral usage of the database (including

challenges in dierent operaonal environments) and to consider the research

potenal of IOM tracking data (Sgter, 2006). Another meeng, aended by

sta from IOM (HQ and eld missions) and external experts (NEXUS Instute and

ILO) was held in Geneva in January 2007 to discuss the ndings and to explore

ways forward.18

Both the evaluaon and subsequent meengs idened the need for further

change within the IOM tracking database, to make the tool more global,adaptable and user-friendly. This included resolving current technical constraints;

ensuring sustainability of the tool within IOM; and accommodang regional

and cultural specicies in data sets. As a result, the screening and assistance

interview forms were revised, with aenon to removing biases. For example,

an earlier version of the database assumed - based on the situaon in SEE at

the me - that all vicms were essenally deceived or coerced into tracking.

Yet this was out of synch with the situaon in other regions where many

tracking vicms were (and sll are) willing labour migrants (oen accessing

formal migraon channels) but ending up tracked. Quesons were also addedto address gaps and somemes a lack of precision in the exisng data sets

for example, about vicms place/family of origin; educaon; past employment

experience; tracker/recruiter proles and dierent types of exploitaon.19

15 Please see secon four for a more detailed discussion of the contextual challenges to global data collecon.16 Now the Migrant Assistance Division.17 While some frustraons and problems connue to be of a technical nature and lessons can be learned from

a discussion of the technical implementaon of the data collecon system, this papers focuses on more

substanve themac and methodological consideraons.18 In late November/ early December 2010, IOM convened an addional meeng with counter-tracking and

assisted voluntary return and reintegraon subject maer experts. The aim of the meeng was to begindiscussing the feasibility of moving IOMs operaonal databases, including the IOM tracking database,

to the internet. Sta also began discussing ways in which to further improve the tool. The process will be

further expanded during 2011/ 2012.19 See secon four for a more detailed discussion of cultural and contextual challenges to global research and

data collecon approaches.


28/96

26

Beneathth


The revised forms were piloted at IOM Headquarters in Geneva, using a number

of cross-cultural tracking scenarios to test for global applicability and regional

and cultural specicies. They were then translated into French, Spanish, Arabic,

Chinese, Kurdish, Japanese, Portuguese and Russian. At the same me, changes

were reected in the technical system, which was piloted at IOM Headquarters

and two eld missions. This resulted in some minor changes and bug xes and

the revised database was then launched globally.

The database is today installed in 72 IOM missions globally, albeit with diering

levels of data collected from mission to mission. As of the end of December 2010,

the system contained data on approximately 16,000 registered IOM assisted

vicms in more than 85 source and more than 100 desnaon countries.20

While the IOM tracking database represents in some ways a unique tool (both

methodologically and technically), IOM is by no means the only organisaon

which uses this approach. Indeed it is similar in many ways to the approach ofmany NGO and GO assistance organisaons around the world which collect data

about the tracked persons whom they idenfy and assist. In some cases, data

collecon mechanisms have been based on the IOM model, drawing on IOMs

past successes and challenges in this eld (IOM et al. 2009; Surtees 2009 & 2007;

Weiner & Hala 2008).21 Most service providers, whether from NGOs, IOs or GOs,

maintain comprehensive case informaon about the tracked persons they

assist both about their tracking exploitaon and their assistance needs. Many

also use this informaon for research as well as case management. There may

be a wide variaon between service providers in terms of what data is collectedand how. Yet this methodological approach that is, accessing vicm-centred

data from service providers as a means of understanding tracking is common

in most countries and constutes the primary source of data about tracking in

the current knowledge base. Moreover, a large number of tracking research

projects are based on vicm-centred data, oen accessing respondents through

assistance programmes and, thus, overlapping in terms of informants with the

informaon collected by assistance programmes (Brunovskis & Surtees 2010).

20 A number of missions also collect standardised data sets based upon the global data collecon approach but

outside of the database. This is due to technical reasons (where the mission is unable to connect to the host-

ing server); or the existence of a previously developed data collecon tool (prior to the roll-out of the global

database). This situaon is currently being resolved and, in 2011-2012, all relevant IOM assistance data will

be captured within the one database. In the interim, IOM requests that all missions operang outside of

the database share a minimum core data set with Headquarters bi-annually. This data is then validated and

compiled.21 IOM has also partnered with various governments globally to support their data collecon mechanisms,

sharing IOMs method and approach and providing technical assistance. For example, this has included

cooperaon and/or technical assistance and training on data collecon in Bahrain, Belgium, Canada, Chad,

Cote DIvoire, Costa Rica, Ecuador, Egypt, Greece, Kosovo/ UNSC 1244, Nepal, Portugal, Romania, SierraLeone, Somalia, Sri Lanka, Syria, Turkey, United Kingdom, United States of America, and Zambia. IOM has

further provided technical assistance at the regional level to the European Union (IOM et al. 2009). Local

iniaves are similarly oen in place at the IOM mission level.


29/96

27




kingvicms

Data from assisted vicms methodological strengths and limitaons

Data collected directly from tracked persons is an important source of

informaon and has proven central in eorts to understand and combat human

tracking. As one CT professional remarked, there is a need to bridge the gap

between research and data collecon and operaonal work:

Research and data are paramount for our work and I really do not

understand why people and managers that are out in the country and in

the eld do not understand this because it is really mandatory. Without

data especially on counter tracking () theres nothing you can do.22

Research and data collecon conducted within the assistance framework, such

as the IOM approach, can shed light on a range of issues including risks and

vulnerability factors, the needs of dierent groups of tracking vicms (e.g. men,

women, children, vicm of labour and sex tracking); the gender dimensions oftracking; details of the tracking process and, albeit it to a lesser extent, the

perpetrators involved, their modus operandi, the routes used and so on.23

Moreover, as will be discussed in more detail in secon 5 (Who is collecng

data? The role of researchers and service providers), this approach (which

involves data collecon by service providers) potenally migates some of the

ethical concerns associated with research with tracked persons, like extensive

interviewing (and re-interviewing), insensivity in the interview process, the risk

of oung respondents through research and so on. The relaonship betweenservice providers and respondents may also translate into more comprehensive,

detailed informaon; data can also potenally be collected over me allowing

for a fuller picture of the individuals experience and needs than a one o

interview may allow.

The sheer volume of data which can be collected in this way potenally allows

for a broader picture of tracking than smaller studies may permit. Research

and data collecon with assisted tracking vicms also allows counter-

tracking professionals to idenfy emerging trends and paerns in real me like new forms of tracking, new routes, emerging means of control, dierent

desnaons, new proles of vicms. This means that data is current and is able

to inform policy and programmac response.

However, as with all methods, research and data collecon with assisted vicms

is not without its problems. And, in going forward, it is important to take into

account and accommodate methodological issues, not least because these

22 Interviews with respondents were recorded, with the consent of the respondent, and then transcribed. The

quotes included in this paper are verbam and have not been edited.23 For a discussion of the methodological limitaons of collecng data about perpetrators from assisted vic-

ms, see for example, Antonopoulos & Winterdyk 2005; Cwikel & Hoban 2005; Goodey 2008; IOM et al.

2009; Kelley 2002; Lazos 2007; Levenkron 2007; Surtees forthcoming 2011, 2009 & 2008b; Troshynski &

Blank 2008.


30/96

28

Beneathth


impact an understanding of tracking and, by implicaon, the ability to prevent

and combat it. This paper lays bare some of the strengths and also the constraints

in working with this approach (both within the IOM tracking database and

through other assistance frameworks). The discussion centres around four main

themes including:

A global approach? Data quality and comparability across dierent terrains Who is collecng data? The role of researchers and service providers

Being representave? Challenges in obtaining representave samples of

tracking vicms

What quesons are asked and why? Assumpons, biases and agendas in

tracking research and data collecon

While the focus of this paper is explicitly about the IOM tracking database, it

is of relevance and importance for all service providers who undertake research

and data collecon in a similar way. In addion, many of the methodologicaland ethical issues raised will resonate with researchers more broadly who, while

perhaps not using this specic approach, may also access tracked persons

within the assistance framework, and, regardless, must consider and address

issues such as data quality and comparability, boundaries between service

provision and research, representavity and assumpons, biases and agendas in

research quesons. Finally, and as importantly, the issues raised and discussed

here are crical for those who rely on data and research to support programme

and policy decisions to combang human tracking. Providing guidance about

what tracking data does (and does not) reveal can go a long way toward abeer understanding of the issue and the policies and programmes which can

serve to prevent and address it. It is these issues that are discussed in the next

secons.


31/96


32/96


33/96

31




kingvicms

has been used over a predened, coded eld and, further, if this is an issue of

training or local adaptaon that is, who is perceived to be an acquaintance in

dierent sengs. Language may also be an issue with open elds, where terms

are listed in the local language or loosely translated rather than according to set

terminology. This is parcularly the case in mulnaonal or mullingual data

collecon aempts, which are oen required in the case of human tracking

due to the transnaonal nature of the phenomenon.25

The above strengths notwithstanding, there are issues with the feasibility (and

desirability) of a standardised approach. Some issues relate to the data collecon

process, while others are a funcon of the rigidity of the model.

While the IOM approach collects standardised data, the data collecon process

is not done in an enrely standardised way. In pracce, many interviewers

use an open ended, unstructured interview technique (oen done by service

providers in the context of case management) and then ll in the interview formbased on memory or the notes they have taken (if the interviewee consented

to note taking). Other interviewers use a more structured format, working

directly with the IOM interview forms, asking each queson on the form in

sequence, using the standardised terminology. Just as dierent methods may

yield dierent response rates (Dillman et al. 2009; Reddy et al. 2006; Schwartz

2000; Yu & Cooper 1983; Weiner & Hala 2008), dierences in implementaon of

one method may yield dierent response rates.26

Moreover, when data is collected by a partnering organisaon (whether NGO,GO agency or other), IOM has even less control over and informaon about

the extent to which data collecon is standardised with data collected by IOM

missions or, for that maer, from case to case within the partner organisaon.

Given that in many countries IOM works with more than one partner organisaon

(and with varying sta within these organisaons), it is reasonable to conclude

that there are threats to standardisaon which result from data collected by

dierent partners even in one seng. Such a consideraon will equally need

to be taken into account by other actors aempng to collate, compare and

analysis data collected from a variety of sources.

25 The issue of language is discussed in more detail towards the end of this secon.26 Research on sensive issues has increasingly considered what methods are likely to yield the most reliable

results. Some recent studies have found that data gathering procedures which avoid social contact (e.g. self

administered quesonnaire, whether wrien or computer based) may be preferred by some respondents

and result in greater disclosure (OLeary 2006; Reddy et al. 2006; Weiner & Hala 2008: 24). However, the

nature of this type of data collecon linked as it is with case management does not necessarily lend itself

to this approach. Rather, service providers work and talk on a day to day basis with beneciaries and, thus,

the interview process is essenally part of this regular way of working and interacng. It is also how case

relevant informaon is collected and counseling provided. Moreover, research has found that the interviewmethod may allow for greater comprehension of quesons and builds rapport between interviewer and

respondent in ways that lead to greater disclosure (Schwartz 2000: 815; Weiner & Hala 2008: 24).


34/96

32

Beneathth


Equally relevant is that while standardisaon may be enhanced by the use

of generic, standardised data entry templates, they oen lack exibility

for instance, to accommodate new paerns and unancipated trends and

behaviours. The oen nuanced and complex nature of tracking is dicult to

capture through predened elds, which means that data collectors may nd

themselves trying to choose the best (predened) answer from a menu of oen

less than adequate responses, as dierent respondents both within and outsideof IOM observed:

[Generic tools] may not be designed to capture the subtle nuances

unique to the cases. Recordings of these nuances give us a more

accurate picture of evolving and changing indicators and local trends.

You cant take that kind of very complex set of circumstances [as seen

in tracking] and convert it into a more specic and simpler set of

quesons that you then ask and repeat and carefully make sure thatyoure asking in very much the same way of everybody, rather than a

kind of opened ended conversaon.

A lot of data comes from drop down boxes so the variables are

predened and the predened variables dont always cover the actual

scenario the predened naming is one of the problems. [in] a lot of

[places] where the variables are dened, it might be beer just to leave

[the eld] open because then you might get a beer picture in terms of

research.

Moreover, working with predened categories may sll leave room for dierent

interpretaons where insucient direcon and training is provided to those

working with the research tools (Sgter 2006). One simple example relates to the

use of the answer opon partner when asked about the vicms relaonship to

people in the recruitment process. There has been confusion as to whether the

word partner denotes a purely professional relaonship (i.e. business partner)

or a purely personal relaonship (i.e. inmate other). As this answer opon is

used in relaon to quesons about dierent people involved in recruitment, itis important to have clear denions otherwise the validity of the coded answer

response may be called into queson.

The quesons asked, the variables used

What quesons are asked and how are vital in terms of the validity and reliability

of the responses received. Poorly designed quesons for example, those which

are poorly worded or use jargon, have ambiguous meaning, are insensive or

uncomfortable for respondents can lead to unreliable or invalid ndings in that

respondents may misinterpret or misunderstand the quesons or be unwilling

to answer in a truthful manner (Di Lillo et al 2006: 419-423; Hamby et al 2006:

515-516; Mitchels 2004: 37-38; Golafshani 2003: 604). As one researcher


35/96

33




kingvicms

interviewed in the context of this paper stressed, there needs to be a process

and criteria to ensure that the queson will be understood in all places where

it is used, that people will have the same probability of answering truthfully to

these quesons.

One example is the use of the term tracking, a term which may be unfamiliar

to respondents, understood dierently by dierent respondents or one whichtracked persons do not associate with their experience. The use of the

tracking term, then, will (dierenally) impact how quesons are understood

and answered. A respondent who answers no to a quesons about whether

they have been tracked may sll have been tracked but may be unfamiliar or

uncomfortable with the term, may see him or herself as failed a migrant rather

than tracking vicms, etc. If the same respondent was instead asked a series

of quesons relang to deceiul recruitment pracces and exploitaon, which

cumulavely constute a tracking experience, the overall assessment may be

yes. Clarity of meaning for all terms and concepts discussed with respondentsis, therefore, essenal.

Moreover, it has been noted that labelling for example, using problemac

or sensive labels such as rape or prostuon or tracking or illegal migraon

may undermine disclosure of vicmisaon. Labelling may be parcularly an issue

amongst some sub-populaons for example, certain culture groups, sexes, age

groups, naonalies, ethnicies, etc. (Weiner & Hala 2008: 21). For instance,

one of the authors has, in past research, explored how the term vicm and/or

tracking vicm can be unpalatable to some vicms, leading some to declinebeing idened and assisted (Brunovskis & Surtees 2007; Surtees 2008a&b).

This highlights the importance of how quesons are framed and issues queried

and how even standard quesons may yield dierent response rates given the

individual or social dynamics which inevitably come into play in such research. It

also signals the need for reliance on behaviourally specic quesons to minimise

labelling and capture issues of and associated with vicmisaon (Weiner & Hala

2008: 21).

Lack of consensus around key quesons and issues have the potenal to seriouslycompromise data standardisaon, thus negavely impacng comparability

and validity. Even subtle or seemingly insignicant changes to wording and/

or response categories can aect the level of self reporng and, thus, data

quality. A 2006 internal evaluaon of the IOM tracking database found that

data validity could be augmented by improving some of the quesons and by

adding addional quesons to obtain a more comprehensive understanding of

the tracking process (Sgter 2006: 13, 25).

Consider again quesons about the roles of persons involved in tracking.

Earlier versions of the database were based on the assumpon that the tracker

was synonymous with the recruiter. When the vicm relayed informaon in

relaon to the tracker, it was encoded into the system as being related to the


36/96

34

Beneathth


recruiter only. This led to informaon on transporters, harbourers, exploiters

and other actors being collapsed into one category that of recruiter. This

potenally caused an informaon gap about these specic categories and, at the

same me, distorted an understanding of who are (and are not) trackers and,

equally, how they do (and do not) operate. Later versions of the database have

included addional quesons which dierenate between the oen diering

roles played by dierent actors throughout the tracking process. Wherethere is relevant data, this allows for cross-variable analysis such as the cross-

tabulang the gender of the vicm against the gender of the actor involved and

the associated role (s/he, they) played.

Certainly some informaon is more easily understood in a range of dierent

sengs for example, more objecve categories like age and marital status.

However, even what might appear to be quite universal and objecve categories

like educaon levels require careful consideraon and denion when

working with vicms, interviewers or case managers who originate fromdierent countries (and thus dierent educaon systems) and who may have

a dierent understanding of what constutes primary, secondary and terary

educaon. As one researcher observed, many seemingly direct categories are,

in fact, quite complicated and diverse when considering all of the various global

constellaons and opons which might arise and, thus, need to be captured in

a global database:

Things like the structure of the family, you need to allow for some quite

strange structures if you are dealing with cultures all over the world.

Similarly, somemes responses will mean dierent things for dierent people

and in dierent contexts. For example, one queson in the database asks how

vicms were referred for assistance, one response for which is self-referred.

In some situaons, this self-referral was quite straighorward i.e. the vicm

escaped tracking and sought the assistance of IOM. However, for others,

self referral is a more complex chain of events, somemes involving a range of

triggers like seeing an adversement, being told about it by a friend, seeking out

assistance only following ones return home, self-referring only in a situaon ofcrisis, etc. For example, amongst Ukrainian men tracked for labour we found

a specic paern of self-referral, whereby they mainly self-referred aer

receiving a recommendaon from a friend who trusted the NGO or IOM and

had visited and received assistance. And usually this self-referral took place only

aer some me, when they had exhausted all other opons for support.

Other data are even more problemac in terms of ensuring standardisaon.

Oen mes an assessment of the vicms economic situaon is based on their

individual assessment (e.g. very poor, poor, average and well o) rather than

an objecve measurement (for example, actual income, income relave to

minimum wage, etc.). This is equally the case with issues such as past experience

of abuse and violence. Percepons of violence and abuse are very individual


37/96


38/96

36

Beneathth


In this vein, IOM has recently worked with eld missions to standardise interview

forms across language. The latest versions were launched in 2008 and have since

been professionally translated from English into French and Spanish. A number

of IOM eld missions have similarly translated and adapted the forms into a

variety of other languages (such as Arabic, Japanese, Kurdish, Portuguese and

Russian) and also dialects to meet their local and cultural needs; while other IOM

missions work mainly with the English versions of the forms, hiring interpretersfor the interview process, when necessary.

However, having forms in a local language only partly addresses language barriers.

Because of the transnaonal nature of both tracking and an-tracking eorts,

data collecon is oen done through translaon or interpretaon at least once

and perhaps at mulple stages (e.g. at origin, transit or desnaon) or in

dierent ways (e.g. a translator is involved in the interview process, translaon of

the interview informaon for entry into the database). A typical scenario might

be a situaon in which a Russian language speaker from the former Soviet Union(FSU) is tracked into prostuon in South-East Asia (SEA). At idencaon

the interview will be conducted by a local NGO or IOM mission sta, which

most oen means interviewing in the local language of the desnaon country

through interpretaon. Data will inially be recorded in the case le in the

local language but then translated into English when the informaon is entered

into the database, which is an English language tool.27 Should the vicm then

be returned home, later informaon gathering will be conducted in their own

language but then translated back into English for the purpose of data entry.

The English language skills of the data entry sta at origin and desnaon may vary quite considerably. The data will then be processed and analysed

by someone who, at minimum, can work with English language data sets but

for whom it may not be a rst language and who may not have knowledge of

the language (or culture) of the respondents. Beyond navigang the dierent

languages is that sta is likely to have diering levels of uency and literacy in

the various languages (as well as in English, their common language), which can

impact the quality of the informaon collected.

Not only is the process of working through many languages cumbersome butthere is further the queson of what gets lost in translaon. One researcher

stressed how generaons of translaon compromise data sets, drawing on her

own recent experience with a mullingual process:

The interviews were conducted in the language of the individual. It was

something that I really did insist on. But, of course, then it was translated

into [the naonal language]. I dont speak [the naonal language].... The

27 The ocial languages of IOM are English, French, and Spanish. Up unl 2008, the database was available inall three of these languages. As the tool was updated, funding constraints only allowed for re-development

of the English version of the tool. While the accompanying screening and assistance interview forms are sll

available in French and Spanish (along with other language), the actual database interface is in English only.

IOM intends to address this issue in 2011-2012.


39/96


40/96


41/96

39




kingvicms

However, a major limitaon remains the fact that such training materials and

guidance notes are only available in English as are many training events and

the technical support system. Further, not all missions have had equal access to

these training opportunies nor are they systemacally implemented. Thus, it is

likely that some IOM oces in certain regions may be less informed and skilled

when using the research tools and entering data into the database itself. Thelack of training opportunies also reduces the number of occasions to receive

interacve and detailed user feedback and, thus, to enhance and improve the

research tools and database itself. Also important is that these improvements

are not able to address issues in the data sets which preceded greater aenon

to training, which is something to be acknowledged in data analysis.

Other factors beyond an interviewers skills and experience can play a role in the

research dynamic. Characteriscs including, but not limited to gender, age, class,

social status, economic background, educaon and ethnicity may play a role inthe interview process including what quesons are answered and consequently

how data is collected. Sex, for example, may aect disclosure of vicmisaon

(Schwartz 2000: 825-26), oen making women prefer female interviewers

parcularly when they have been vicmised by men (WHO 2003: 14). It may also

be a funcon of cultural and social mores. This was illustrated in the responses

of two former tracking vicms (women in their 20s from Albania) when asked

whether they would feel comfortable in being interviewed by a male researcher

or by a researcher with a male translator. Both were vicms of sex tracking and

stated a clear preference for women to be involved:

If there is a female, I would speak really freely because I would think I

am speaking to my mom or my sister. But if there was a man I wouldnt

have come at all.

Its a bit uncomfortable if there is a man, because if there is a woman I

feel more free.

By contrast, tracked men interviewed in Ukraine were also asked if they hadany concerns about the sex of the interviewer and more specically about being

interviewed by women. Their responses were more ambivalent, as the quotes

from two tracked men below illustrate:

Q. Would it have been dierent with males, having a male or female

interpreter?

A. [.] really doesnt maer speaking with men or women it doesnt

maer much.

It doesnt maer much to me but, of course, it is more pleasant to

speak with women. A male translator is probably beer because he

understands men [.] a man can understand what I mean for certain

things.


42/96


43/96


44/96

42

Beneathth


Some missing data may be more a funcon of instuonal arrangements.

There may be reluctance - or an inability from partnering organisaons

to share certain types of informaon for example, because of concerns with

condenality, legal issues associated with data protecon, lack of safe data

sharing channels, issues of ownership of the data or because of a more

limited role played by IOM in providing assistance. In many sengs, IOM no

longer directly implements assistance programmes and, thus, much current datacollecon relies on agreements and cooperaon between partner organisaons

and IOM.

In some instances, it may be me and resource constraints which limit the

amount of data collected. Service providers oen dont have me to collect

and/or enter a lot of detail about each beneciary. Where data collecon is too

ambious (e.g. too many data elds or quesons), there is likely to be a high

degree of missing data, as one NGO worker noted of the IOM interview forms:

The whole process of collecng this level of detail has problems. It takes

a long me to ask the quesons and then ll in the informaon. And it

is stressful in many cases to be asked such specic quesons, especially

about the tracking experience.

Missing data might also be a consequence of the methodological approach taken.

Consider, for example, interviews conducted with males in the FSU. A desk-top

analysis of relevant (and de-personalized) IOM database data in advance of in-

country eldwork revealed that many males were noted as having self-referredto IOM or partnering NGOs. However, aer employing a mixed method approach

which involved interviewing males, we leant that many men had in fact only

self-referred aer receiving a recommendaon from a friend or colleague who

trusted the NGO or IOM. In this example, the data received in a eld research

seng added to and enhanced the ndings from the IOM database.32

Thus, while there is a need to ensure suciently complete data for standardisaon

and comparability, data collecon iniaves also need to balance what needs

to be known against the degree of informaon that can be realiscally (andethically) collected.

Comparability across contexts

An addional queson is to what extent tracking data is comparable across

dierent naonal and cultural boundaries. Quesons and concepts (and, by

implicaon, interview responses) may not be understood in the same way

between countries, cultures, dierent sta, interviewers and interpreters and/

or at dierent points in me (e.g. Van Liempt & Bilger 2009:3; Surtees 2007c;

Bre-Davies 2007: 82; Mitchels 2004:37-38).

32 The ability to draw upon the IOM assistance data in advance of the eldwork did however help to provide a

stronger baseline understanding and entry point for analysis, helping to rene the research focus.


45/96


46/96


47/96


48/96

46

Beneathth


other seng because these are not the experiences that are useful to

statutory services in terms of prosecuon; because of shame/sgma

around sexual violence; and because for women who have been tricked,

deceived and abused, it is oen dicult to trust anyone. We believe that

the women we support tell their support workers of these experiences

precisely because they are in a place of safety where they are believed

(POPPY Project 2004: 1).

There is also an element of protecon and condenality when data is collected

within the assistance framework because some forms of respondent recruitment

for example, community sampling may serve to idenfy tracked persons

to their families and communies. While oung tracking vicms to the

community may dier in signicance and fallout for example, relave to

the form of tracking or cultural context in many cases, it has a very real

and serious implicaons, including social ostracism, rejecon by ones family,

discriminaon, reduced economic opons, limited access to social services,threats and retribuon from tracking and even physical and sexual violence by

family or community (Brunovskis & Surtees 2010). This approach also potenally

provides a layer of protecon in that, should any assistance needs surface during

or following the interview, service providers are aware of and able to respond

accordingly (Brunovskis & Surtees 2010, 2007; Surtees 2007b).

Data collected by service providers is also praccal in that it potenally provides

a great deal of informaon about a large number of tracking cases in an

economical fashion, tapping into an organisaons exisng data colleconprocess. And it is dynamic. When vicms are assisted over a period of me,

data can be updated, expanded and even corrected as a fuller picture of the

individuals experience and needs emerge, as is oen the case with IOM data.

When contact with vicms is ongoing, even beyond the formal assistance and

reintegraon process, this may provide access also to longitudinal data, rather

than a snap shot approach (for example, single interviews or surveys at a set

point in me), which is common in tracking research.34 And because vicm

assistance organisaons are on the frontline of the issue, this can mean being

able to idenfy trends in real me like new forms of tracking, new ro

beneath the surface iom

Documents