Return after flight : exploring the decision making process of Sudanese war-displaced people by employing an extended version of the theory of reasoned action

Research Issues and Approach

Forced displacement and the response of the international community is one of the most pressing challenges of contemporary times. Whether in dealing with refugees or internally displaced people the international system has been struggling to prevent forced migration, address its consequences and find ‘durable’ solutions (chapter 1). This study concentrates on repatriation policy and practice, which are considered to lack responsiveness to the needs, initiatives and strategies of the displaced as the prime actors (chapter 2 and 3). The assumption addressed in this study is that a better understanding of displaced people’s return planning, and the beliefs, values and motivations underpinning them, is fundamental to developing effective strategies that result in appropriate protection, humanitarian assistance and reintegration support as well as sustainable development for people displaced by conflict.

Displaced people are not helpless and passive recipients of ‘well-intended’ aid. Inappropriate generalisations about refugees and other war related displaced people have encouraged ignorance of war-displaced people as conscious and active human beings, strategising their return options and making sense out of often highly complex and dynamic return environments. Displaced people are therefore to be regarded as actors possessing agency, i.e. they are capable of processing social experience and devising ways of coping with life, even under the most extreme forms of coercion. This study therefore emphasises the importance of human agency in informing both repatriation policy and practice.

The research presented in this thesis has:

1. Explored the factors that have greatest influence on the return decisions of war-displaced people

2. Looked into the ability of an actor oriented model to account for, and explain, the dynamic nature of return beliefs and issues informing the decision to return

3. Examined the application of the extended Theory of Reasoned Action in terms of increasing the understanding of prospective return behaviour as compared with more traditional models

4. Tested the inclusion of the vulnerability concept in enhancing the sensitivity of the model in volatile and hazard prone contexts

5. Looked into the appropriateness of an actor oriented formal model in informing both policy and practice.

Method

The Theory of Reasoned Action (Ajzen and Fishbein, 1980) was selected as the actor-oriented model of choice to explore war-displaced people’s return decision-making. Most reviews in the field of behavioural research have drawn attention to the TORA as one of the principal theoretical constructs for both the study and prediction of behaviour. The theory has been applied to a wide variety of behavioural domains proving its robust nature, predictive power and its ability to perform in different contexts (section 4.3 presents the rationale for its application to this study). This research presents the first time this theoretical construct has been applied to study the systematic identification of return beliefs and issues taken into account by people displaced by war when making decisions regarding return and repatriation.

The information gathered through applying the TORA allows for a detailed explanation of displaced people’s return decision-making (see figure 4.1). At a general level return behaviour is determined by return intent. The principal theoretical assumption underpinning the approach is that attitudes and social normative pressures inform behavioural intent. A third variable, Perceived Behavioural Control was later included in the model to expand the TORA’s ability to address behaviours that are not under the full volitional control of the subjects under study, i.e. the Theory of Planned Behaviour (TPB) (Ajzen, 1985, 1988 and 2005). This study suggests a fourth determinant, Perceived Vulnerability, based on the assumption that it would enhance the sensitivity of the model when dealing with risk prone behavioural contexts. The third level of analysis accounts for attitudes and subjective norms in terms of beliefs about the consequences of return and about the normative expectations of relevant referents. Each successive level of analysis in this sequence provides a more comprehensive account of the factors determining repatriation (reference is made to figure 4.1).

One of the components (section 1.3) of this research is to apply the extended TORA model in exploring the return intent of a group of Sudanese refugees sheltered in three camps in western Ethiopia. Another, and related component, is to test the appropriateness of the extended TORA in different forced migration situations by exploring return intent of a group of internally displaced people in the Sudan. The application of the Theory of Reasoned Action within both displacement contexts is described in section 4.4, 6.1 and 9.1. The findings are presented in chapters 6, 7 and 9.

Research Findings

Research Question 1: What factors have greatest influence on the decision to return of Sudanese refugees and IDPs? Are these structurally different?

Both the refugee and IDP study demonstrate the influence of attitude over social norms. Return intent, and ultimately return behaviour, was found to be primarily governed by displaced people’s own experience and perspectives regarding return, as opposed to social influences or pressures. The Perceived Level of Control and Perceived Vulnerability, though strongly correlated with general return intent for the refugee study and weakly for the IDP study, are less influential compared with attitude (both variables are strongly correlated with each other indicating that vulnerability expresses a significant perceived lack of control). These findings underline displaced people’s disposition to decide return based on their personal experience and evaluation of the consequences of return (sections 6.4 and 9.3).

Various descriptive variables were defined to distinguish between differences in displaced people’s backgrounds and contexts. Generally speaking, refugees’ return planning and strategies were found to be most significantly influenced by:

· Camp location
· Tribal background
· Gender
· Type of critical event initiating flight
· Flight pattern and length of displacement experience
· Knowledge regarding whereabouts of family in the Sudan
· Prominent challenge faced by being a refugee
· The type of conflict thought to present the main challenge towards future life in the Sudan.

The internally displaced study found a much more limited number of variables which registered significantly differences in return intent:

· Having or not having an income from a permanent job or daily labour
· The expressed level of vulnerability in displacement
· The type of benefits associated by residing in or near Khartoum.

The difference in the number of diagnostic variables partly reflects the diverse make-up of the refugee study, which was comprised of eight ethnic groups and many different return contexts, as compared with one particular tribal group and return area for the IDP study.

Analysis of the refugee data found that, in general, return intent is influenced by the following barriers or blockages presented in decreasing order of influence (section 6.5, appendix 6.6):

· Will be conscripted into an armed group
· Face hardship and difficulties
· Our children will be abducted
· Will be repeatedly displaced
· Will be arrested
· No modern health care services provided
· No educational services provided
· Our children will not have access to schooling
· To rebuild livelihoods better not wait for UN repatriation
· Loss of chance for resettlement in western country
· Insecurity makes it difficult to have enough to eat.

Influential drivers for return were found to be:

· Food-secure within a short period of time (less than six months)
· Can rely on traditional ways of peacemaking
· Collect and consume wild foods to survive
· Large household helps to become food secure.

Either, a change in the level to which these beliefs are seen to be true or untrue, or a change in the attributed value to these prospective outcomes, affects the intention to return and, therefore, ultimately return behaviour.

The IDP study found the following blockages presented in order of decreasing significance (appendix 9.6):

· Experience serious food shortages and face hunger
· Find that we have become different from those who stayed behind
· Face serious shortage of safe and reliable drinking water
· Difficulties in constructing our tukuls or shelters
· Life will be different from what it used to be before the conflict: ‘life has changed’
· If provided, aid is best distributed by our own leaders
· Will be suspected
· Face repeated displacement
· The international community not providing food aid means we might die
· Face insecurity in our home area
· Half food rations for a year is better than full rations for 6 months.

Influential drivers affecting the return decision were found to be:

· Children will have access to school
· Burial in my home area
· Rely on our traditional ways of peacemaking to make and maintain peace
· Depend on forest foods and fish for survival following return
· Come to own goats or cattle (re-build wealth status)
· Half food rations for a year is better than full rations for 6 months.

The seventy salient beliefs emitted by the refugees, and the fifty-one by the IDPs, were grouped in a logical way to represent attitudinal dimensions of return behaviour referred to as belief domains or systems. A total of eleven domains were defined: safety & security, health, education, food security, mode of return, peace makers, peace spoilers, culture and integration, processes of change, community based support systems, and humanitarian aid.

For both the refugee and IDP study influential barriers informing general return are found in a number of belief domains: safety and security, health, educational and food security. Influential drivers in the traditional community based peace making, food security- and aid-domains. Interestingly only the IDP study found that loss of identity and processes of change are influential in deciding return. It appears that prolonged exposure to an urban lifestyle has formed a deterrent towards return, as the displaced feel that they have changed and therefore life in Abyei will never be the same. However, on the other hand the desire to be buried in one’s homeland forms an influential driver towards return. Both influential beliefs reflect an underlying schism between the older generation who grew up in Abyei and a younger generation having hardly any memory of their home area and a rural livelihood. In a different way this generational divide is also found amongst refugees and is expressed in the popularity of the resettlement option to America, Canada or Australia amongst the younger refugee generation. For them the potential re-settlement option forms a strong barrier against return to the Sudan.

It was shown that clustering of all salient return beliefs provided an interesting insight in the general cognitive structures underlying the various belief and value domains that inform decisions to return. Associations of influential beliefs acting as barriers, or drivers with less influential ones, provided an understanding of the complexity and dynamics of the beliefs taken into consideration when deciding return (figure 6.1 in section 6.6 and figure 9.1 in section 9.4).

Research Question 2: Can a structured actor-oriented model account for, and give insight into, the dynamic nature of return beliefs and expectations of home that inform displaced people’s decision to return?

The findings identified through the data analysis of the refugee and IDP studies underline the importance of recognising the diversity in backgrounds and return contexts over time that exist within war-displaced populations. Application of the TORA construct identified how such differences are reflected in the attitudes, social norms and perceived level of control and vulnerability. The degree to which household heads plan to return, and the type of return beliefs and issues informing the decision to return, showed a very significant variation across different segments of both the refugee and IDP population, as well as across the various return options presented (for the refugee study: return in the 2002/3 dry season, the period 2003-5 and when there is ‘peace’; for the IDP study return in the 2003/4 dry season, following the signing of the peace agreement and return with Abyei presenting a safe and secure home area). The decision-making process of displaced people regarding return was thus found to be both dynamic and context specific.

Findings of both studies highlighted the social diversity and diversification of the communities regarding migrational decision-making. Household categories were found to express a diversity of return responses. Generally speaking, and in line with expectations, the intention to return was found to increase over time with peace signalling a strong increase in that intent. This highlights the importance of conflict resolution to the return decision-making of people displaced by war and conflict. Independent of the return option pursued, the return decision of both refugee and IDP household categories was found to be governed foremost by displaced people’s experience and knowledge regarding displacement, and expectations of the return process and context.

Analysis of the refugee data demonstrates that prior to peace the security situation and its implications for returnees’ livelihoods are seen as serious obstacles for return. However, provision of aid and services in relatively stable and peaceful areas was found to directly result in a higher return intent, particularly so for those households planning to return to their home village. Findings of the IDP study demonstrated the same pattern, with those planning to return before the signing of a peace agreement preferring the SPLA controlled southern part of their home area. Interestingly, if peace is seen to exist, the provision of aid and services in areas of return no longer affects displaced people’s decision to return: the focus shifts towards social integration issues including community support systems.

Both the refugee and IDP studies found that the nature and number of influential return beliefs informing the return-decision change over the return options from barriers to drivers and from beliefs dealing with fulfilment of basic needs and security to socio-culturally connotated beliefs. If peace exists the influential return beliefs informing the decision of both refugees and IDPs form all drivers towards return. The one common influential belief for both groups is the perceived level to which one can rely on traditional mechanisms to manage conflict within their own and across tribal territories (section 6.7 and 9.5).

The Subjective Norm (perceived social pressure) registered a strong correlation with stated Intent to return for the refugees studied, independent of the return option pursued. Although, perceived social pressure was also found to strongly influence IDPs’ return decision in the immediate or proximate future, it became a weaker factor once the location (Abyei) was considered a peaceful and safe environment to return to.

Regarding the refugee study, both the Perceived Level of Control and Perceived Vulnerability were found to strongly influence the return decision independent of the return option. In general, refugees’ decision to return prior to ‘peace’ reflects both an underlying pre-occupation with threats and hazards in what are perceived as potentially volatile post return environments, as well as a lack of control over the return process itself. Following ‘peace’ return is associated with resilience and increased levels of control. The IDP study found that the perceived level of control and vulnerability only informed the decision to return prior to the peace agreement. Following the signing of the agreement and with Abyei being safe and secure both variables do not significantly inform the return decision. Generally speaking this reflects refugees’ tendency to take issues of vulnerability and control into account when deciding return independent of the return moment itself. For the internally displaced Dinka Ngok this depends on the peace process, following the signing of a peace agreement vulnerability and perceived control no longer influence the decision to return as such.

As a structured actor oriented model the TORA was thus found to be able to explain the dynamic nature of return beliefs and issues in informing displaced people’s decision to return.

Research Question 3: Will the application of an ‘actor-oriented’ approach lead to a greater understanding of prospective return behaviour than currently applied models on which return management and policy are based?

A majority of the studies on displaced people’s attitudes towards repatriation reflect the importance of displaced people’s decision-making in the return or repatriation process (section 4.2). Yet most formal repatriation programmes are conceived by governments and agencies with little reference to, and understanding of, the return decision making process of war-displaced people, their return strategies and in many cases pro-active stand regarding return.

Rather than explaining return behaviour as emanating from the external context (socio, economic or political) the TORA explains behaviour by exploring the psychosocial antecedents informing the return decisions of displaced people. The TORA, as a method, encapsulates the concept of human agency by exploring displaced people’s knowledge-, belief-, value- and power-domains which inform the decision to return. Displaced people are not the passive receivers of ‘well intended’ aid but participate and interact with other actors involved in the return process and reconstruction efforts. The beliefs and issues that were found to influence their decision to return reflect both their background and the specific nature of return contexts (chapters 5, 6, 7 and 8, 9).

The findings of both the refugee and internally displaced people studies demonstrate that the descriptive utility of the extended Theory of Reasoned Action is a powerful and informative conceptual model. Certainly when compared with the traditional ‘push-pull’ migration theory as well as assimilationist models of migrant integration that still dominate much of the discourse on migration (section 2.1). The TORA directly accounts for the experience of migrants, a failure of former models, for which they have been criticised (section 10.2). The TORA also presents a participatory approach (the elucidation of salient return beliefs and issues across a displaced population during the first stage of the interview process). Lack of participation, due to insufficient time or space, has been a common criticism of repatriation policy and practice. Additionally the TORA’s findings are supported statistically, while more socio-anthropological oriented studies are sometimes critiqued for their biased outcomes, reflecting the personal views and opinions of the researcher, and their lack of representation.

Research Question 4: Does the inclusion of the notion of vulnerability enhance the sensitivity of an actor-oriented model in volatile or hazard prone environments?

The addition of the ‘Perceived Vulnerability’ variable as an extension of the TORA drew attention to displaced people’s understanding of a potential risk associated with the proposed return. The variable makes it possible to analyse the context specific nature of risk and ability to cope, thus providing an insight into their perception of vulnerability.

The refugee study found that the decision to return is very strongly informed by the Perceived Vulnerability variable, independent of the return option pursued (section 6.9). In general, the nature of Perceived Vulnerability expresses vulnerability associated with beliefs in the security and safety-domain before, and following peace resilience associated with beliefs related to the food security-, cultural-, community support- and aid-domains. The case study on Nuer and Anyuak return (chapter 7) illustrated that prospective return areas may present hazardous environments even with return pursued when there is ‘peace’. From a vulnerability perspective the imminent importance of influential Nuer beliefs, such as fear of hostile tribes, and lack of control by a central authority over small firearms at such a point in time highlights the volatility of prospective areas of return. All of these form influential deterrents regarding return.

The IDP study demonstrated that the return decision is strongly informed by the Perceived Vulnerability variable for ‘immediate’ return only, i.e. return in early 2004. The decision to return following the signing of a peace agreement and with Abyei presenting a peaceful return environment was not informed by the Perceived Vulnerability variable. However, some individual influential return beliefs were found to be associated with aspects of vulnerability, notably beliefs in the safety-security domain and cultural domain respectively (section 9.5 and table 9.3b and 9.3c).

The importance of the Perceived Vulnerability variable is that it draws attention to critical return issues and explains them in terms of the context specific nature of the hazard. Seen as a predictive warning utility the Perceived Vulnerability determinant enables to uphold the voluntary nature of return and to inform appropriate protection and aid programming over and following return, prior to return itself. The Nuer and Anyuak case study also demonstrated its use as an indicator of whether return strategies are based on risk reduction or increased coping capability. Nuer return strategies were found to be primarily informed by the level to which they perceive themselves able to cope with threats posed in post-return environments. The return strategy of the Anyuak was found to be based on a combination of risk reduction and increased coping capability. The case study highlighted the potential of such differentially informed strategies to be dysfunctional in that they are likely to generate conflict in frontier areas where Nuer and Anyuak interests overlap (see section 7.6).

Perceived Vulnerability, as a variable of return intent, is numerically expressed in a Vulnerability measure based on a scale ranging from very ‘vulnerable’ to very ‘resilient’. The ‘Perceived Vulnerability’ scale is useful for measuring and comparing expressed vulnerability or resilience levels both within and across displaced populations. The nature of expressed vulnerability or resilience can be explored by looking into the associations with the reasoned return beliefs and expectations of home. This ability makes the extended TORA a potent tool in community-based vulnerability and capacity analysis when considering return behaviour.

By using the Perceived Vulnerability reading as a descriptor variable, the return planning and strategies of household categories expressing different levels of vulnerability or resilience in return can be studied. Following the signing of a peace agreement, the IDP study found that households expressing vulnerability in return (as opposed to households who express themselves to be resilient or highly resilient) are prepared to take on the biggest risk, while at the same time they were found least likely to cope with that risk. Further analysis established that the return decision of households expressing vulnerability in return is emotively informed, and does not reflect an analysis of what may constitute critical return beliefs and issues. The findings of the IDP study also indicated that, with increasing levels of resilience, the return decision reflects a critical consideration of reasoned return beliefs and expectations of Abyei. In contrast, independent of the level of vulnerability associated with return, refugee household’s decisions to return when there is ‘peace’ were found to be significantly informed by an analysis of reasoned return beliefs and issues.

As a descriptor variable the Perceived Vulnerability determinant proved to be a highly differentiating variable, explaining the level to which refugee or IDP households plan return, the particular sets of return beliefs and expectations of the return context influentially informing the return decision, the nature of vulnerability or resilience expressed in return as well as how the different household categories feel about return (in terms of return beliefs and expectations of home).

Destitution amongst vulnerable IDPs is often more pronounced when compared with refugees. Refugees, independent of the level of expressed vulnerability, have recourse to an established system of international protection and assistance which IDPs lack. The importance of the finding that most vulnerable IDPs make an emotively as opposed to reasoned return decision highlights the fact that situations of severe deprivation can induce instances of behaviour that may be counter to the subjects’ underlying beliefs. In a sense one can say that the voluntary nature of return, a key Guiding Principle on Internal Displacement is itself compromised by pronounced levels of destitution in displacement, which can place the most vulnerable particularly at risk on return.

The expansion of the TORA with the ‘Perceived Vulnerability’ variable was found to enhance its descriptive utility in exploring the potential for disaster in post –return environments. Both this ability and the findings of the refugee and IDP studies suggest that, in contexts where the elements of threat or fear are associated with the decision in question, this extended version of the TORA will enhance the explanatory power of the model for the study of risk-prone behavioural decisions.

Research Question 5: Can a formal theoretical model be suitable for general application to inform repatriation policy and practice?

Is the extended TORA as a formal theoretical model suitable for general application to inform repatriation policy and practice? Findings of the refugee and IDP studies suggest a positive answer to this question on three different accounts.

First, the value of the extended TORA construct as a model for isolating specific return beliefs and issues. These need to be taken into account in policy and strategy formation in order to inform direct interventions in the field of protection, provision of aid and re-integration support. Also, its descriptive utility is a valuable tool in managing more effectively the nexus between spontaneous return and organised repatriation.

Secondly, as an extension to the TORA the Perceived Vulnerability construct enables the exploration of both the importance and nature of vulnerability associated with key migrational decisions. This makes the TORA construct a potent tool for mapping social or community vulnerability in volatile or hazard prone contexts. With vulnerability being a central concept in disaster studies, and a more recent focus as an aspect of world development, the construct of Perceived Vulnerability is a potential tool for the development of disaster policy and practice.

Finally, this research has drawn attention to the importance of the level to which the emotively expressed attitude towards the return decision is informed by the analysis of reasoned return beliefs and expectations of home as an indicator of the voluntary nature of return. This is an important finding, as many a repatriation programme or assisted return has been initiated based only on an emotively informed return response. This has led to misguided interventions resulting in a waste of resources at best, and at worst placed returnees at risk. The findings of both refugee and IDP studies suggest that for voluntary return to take place in safety and dignity, the decision to return should to a significant degree reflect a reasoned analysis of pertinent return beliefs and issues.

Based on the findings of this study recommendations are made regarding areas where the extended version of the TORA can be applied to inform the development of both policy and practice. To mention the most promising (chapter 10):

· Its role in multi-way information systems to enhance communication between the major stakeholders in the management of population return

· Compilation of a checklist of beliefs derived from other research into critical return beliefs as a check to make sure that potentially critical return issues are taken note of

· Rather than exploring general return planning and strategies, particular belief and value domains may be focused upon, e.g. exploring in-depth the security and safety domain with the purpose of informing protection programming

· The need to simplify and deprofessionalise the models’ application in order to make the construct more widely available and cost-effective, e.g. to facilitate its re-deployment to account for changes of return intent over time

· Rather than focusing on the institutional interest and need to understand displaced people’s decision-making, attention should be given to the models’ utility in serving ‘direct’ community interest, e.g. as a tool of community learning regarding critical return factors.



Conclusion

By applying the extended TORA an understanding of war-displaced people’s return decision is gained which is instrumental to guide humanitarian aid and development interventions that support and build on displaced people’s return strategies. The extended TORA can contribute to the development of both repatriation policy and practice, which enables a genuine and more participatory approach to planning and decision-making. This would result in interventions which take account of the complexity inherent in return and repatriation. Honouring displaced people’s agency allows for taking into account their critical return beliefs and issues. Ultimately, this comes down to the question of respect. Respect for the complexity of what ‘voluntary return’ constitutes in today’s world and, even more important, respect for the people who are forcibly displaced. This means insuring a return to their places of origin in safety and with dignity, not as a favour, but as a fundamental human right.War-affected displacement and its responses by the international community is one of the most pressing challenges of contemporary times. Whether in dealing with refugees or internally displaced people, the international system has been struggling to prevent forced migration, address its consequences and find durable solutions. Repatriation policy and practice, pursued as the durable solution to war induced displacement, are being critiqued for lacking responsiveness to the needs, initiatives and strategies of the displaced as its prime actors. A better understanding of displaced people’s return decisions, and the beliefs, values and motivations underpinning them, is seen as instrumental in developing strategies to deliver effective protection, humanitarian assistance and reintegration support.
This thesis explores the return decision-making process of war-displaced communities by taking an actor oriented approach. The Theory of Reasoned Action (Ajzen and Fishbein, 1980; Ajzen, 1985, 1988 and 2005) was used to analyse the psycho-social antecedents of displaced people’s decision to return. A ‘Perceived Vulnerability’ variable was suggested as an extension of the model, based on the belief that it would make the model more sensitive to potentially hazard prone or risky behavioural contexts. Two major surveys were undertaken in order to explore the decision-making process of a group of Sudanese refugees and a group of internally displaced people.

Both surveys found that the dynamics of return are informed by the changing nature of return beliefs and expectations considered influential by war-displaced people when deciding return. Such issues were found to reflect the social diversity and internal differentiation of war-displaced communities, as well as the diversity of return contexts. The ‘Perceived Vulnerability’ variable enhanced the descriptive value of the model in hazard prone or risky contexts and was found instrumental to analyse the context specific nature of risk and ability to cope with those risks thus providing an insight in displaced people’s perception of vulnerability. The study’s findings suggest that the application of a formal model, such as the Theory of Reasoned Action, can be suitable for informing repatriation policy and practice. As an actor oriented approach, the theory may prove helpful in developing flexible and de-centralised approaches that support and build upon displaced people’s return strategies, as appropriate to complex and dynamic contexts.

Keywords: Forced Migration, Humanitarian Aid, Resettlement, Repatriation, Migrational Decision Making, Refugees, IDPs, Sudan, Ethiopia, Theory of Reasoned Action (TORA), Theory of Planned Behaviour (TPB), Vulnerability, Risk Management