Secondary Logo

Journal Logo

Supplement Article

Logics of Commitment in Vaccine Trials of the ANRS Cohorts 1992–2001

Fillieule, Olivier PhD

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: October 1, 2018 - Volume 79 - Issue - p S59-S68
doi: 10.1097/QAI.0000000000001809
  • Free


In this short chapter, we will confine ourselves to presenting only the rationale for application in a diachronic manner without exploring the details of the rationale for selection. We propose a synchronic and diachronic view of the population of 4259 candidates. Is it possible to define a typical profile? Do these people share any common characteristics? In a second step, we analyze a sample of the application letters by means of a lexicometric analysis1 to explore in more detail the motives (C. Wright Mills, “Situated Actions and Vocabularies of Motive.” American Sociological Review, Vol. 5, No. 6. December 1940: 904–913.) put forward by the applicants, as well as to further explore hypotheses about the rationale for recruitment.

Three objectives underlie this article. First, although the research on the social characteristics of volunteers and on blood and organ donation is developing in France, the recruitment of volunteers for vaccine trials is still in its infancy. We hope here to show how much engagement in trials owes to the same theoretical questions and the same key notions as the forms of volunteering usually studied. Then, the fact of having access to data concerning the entire decade of the 1990s permits an objective view, in an original way, of the effects of the relative normalization of the HIV epidemic in terms of its public image and representations of the disease (Regarding the question of the normalization or institutionalization of HIV, see Refs. 2 and 3.). Finally, the very abundant material contained in the application materials of both the people selected and not selected is a powerful tool to evaluate recruitment methods and communication campaigns in the different stages of the selection process.

The General Selection Process

Over the 10 years scheduled for the study, selection was conducted in a series of stages with little variation. A call for applications was made through various targeted media. This was expanded in 2001 to include general mass media. Until 2000, applicants were invited to send an initial letter to request the application materials and provide their contact information. The Agence Nationale de Recherche sur le Sida et les Hépatites (ANRS) sent the application materials, which needed to be returned together with a cover letter. After an initial selection was made on this basis, the applicants were requested to go to a hospital to undergo biological examinations, clinical examinations, and an interview with a psychologist or psychiatrist. The results of these tests were shared with the selection committee, which evaluated them and decided, if appropriate, to send an application form for inclusion in the network together with the results of the analyses. The people who returned a completed and signed form were included and contacted again to undergo additional medical examinations. The results report was an important stage in recruitment given that the study considered the volunteers as partners.


The application materials made it possible to collect information regarding sex, year of birth, department of residence, and profession. Other details collected are marital status, number of children, whether or not the respondent was living as a couple, and the way in which the applicants learned about the call for applications. Also specified are the year of application, successful completion of a number of stages, whether the respondent dropped out of the process at a given stage, inclusion in the network, and inclusion or not in one or more trials for members of the network (Because of an interruption in the publicity campaign in 2001, the data after this change [cohort “2001b”] is treated in a specific manner distinct from data from the same year but before the change [cohort “2001a”]. In addition, the numbers for 2001b alone are equal to those for all of the previous cohorts combined).

Diachronic Change in Applicant Profiles

In addition to a general vision of all the applicants who responded to the call of the ANRS, we propose an analysis by an agglomerative hierarchical clustering, which makes it possible to distinguish 5 classes of applicant with relatively similar numbers and to present these profiles in a summarized manner (Table 1).

Five Sociodemographic Candidate Profiles
Five Sociodemographic Candidate Profiles

Class 1 essentially comprises singles younger than 35 years, often men with no children, half of whom live outside of Paris, who are students, senior managers, or employees in various sectors. Two other classes comprise almost exclusively people older than 45 years: class 2 includes people who are married or divorced but generally are in a couple, most of them live outside of Paris, have at least 3 children, and are often senior managers or nonmanagerial employees, with 50% in the health care of education sectors but with a notable share of retirees; class 3 includes as many divorced people and singles as married people, with half living solely as a couple, no more than 2 children and working in professional positions, which are slightly lower in the hierarchy of French socioprofessional categories (SPC). The members of the last 2 classes belong to different SPCs and are of middle age (30–49 years old): class 4 comprises people who are either married or living as a couple outside of Paris, with 2 or 3 children, 40% work in health care or education; class 5 comprises single men, half of whom are between 35 and 39 years old, they tend to live alone, half of them live in Île-de-France, they have no children and work more than the other classes in media, the arts, or the private tertiary sector.

Four significant elements stood out in the diachronic analysis: a relative feminization, relative stability in terms of age, a decline in singles, and a diversification in professional profiles (Table 1 and Fig. 1).

Changes over time of class of applicants.

Female applicants are the minority throughout the decade. Their number grows steadily except during a dip in 1996, which should not be considered significant because their number at this time is quite low. Women more or less catch-up with men in 2001 after a new advertising campaign was launched, which seems to have reinforced the feminization of applicants.

We can observe a change toward an older population of participants but at a relatively slow and irregular pace. Only 18- to 21-year-olds are under-represented in the applications. However, this stability hides a discrepancy: out of an equal sample, men are slightly fewer among 20- to 24-year-olds but over-represented among 25- to 39-year-olds, whereas women predominate in the higher age groups. Because the women are older on average, this significant share contributes to the trend of a higher age in the population studied.

The share of the different marital statuses is fairly stable over 10 years. Singles are the clear majority at over 50% of applications consistently, but we note a slight decline over time with a rise in the number of people living as a couple, which supports the hypothesis of a relative “normalization” of the cause.2

From a professional point of view, the decline in senior managers and students, and the growth in nonmanagerial employees and manual laborers illustrate the expansion of the cause and its spread outward compared with the affluent and educated populations of Île-de-France. The growth in SPC and business sectors makes it possible to distinguish precise professional profiles that evolve the most in terms of volume during the years 1995–2001. Without going into details, we are able to identify some clear trends. Among the professions that regress the most, we find secondary school teachers, higher education teachers, and researchers (from 20% of applicants at the start of the decade to 6% at the end of the decade), professions in entertainment, the arts and communications (from 5% to 3%), and senior and middle managers in sales and marketing (from 2% to 1%). However, there is a rise in the number of administrative employees (from 2% to 5%), middle managers and category-B civil servants (from 1% to 3%), nonmanagerial employees and manual laborers (from 3% to 5%), and independent craftsmen (from 0% to 1%). There is a decline in the midlevel professions in the social (from 3% to 2%) and health care services (from 6% to 5%), but growth is reported for senior managers and nonmanagerial employees/health care workers, as well as nonmanagerial employees/workers in the social services. It is these “internal transfers” that allow these sectors, particularly the health care sector, to continue to act as a reservoir of candidates, unlike other sectors that mainly fuelled recruitment in its early stages and then experienced a decline.

Overall, engagement in the trials seems at the end of the decade to be less mysterious and to require a priori less resources to be feasible: steady growth in the number of women, an older population as of 1998, a slight decline in the number of single people and people without children in a demographic context, which is nevertheless favorable for them, higher frequency of couples despite a reverse demographic trend in the general population, decline in senior managers and students but growth in nonmanagerial employees and workers, and expansion of recruitment to include sectors less affected by the disease.


A multiple correspondences analysis can provide an overview of the factors that contribute to the evolution of the composition of the cohorts of applicants. Figure 2 represents the 2 most important factors for an understanding of our population. Axis 1 contrasts young singles with no children, often living alone and often students, with older people, either married or at least living as a couple, and with 2 or 3 children, thereby illustrating a sort of distribution by age group. Axis 2 contrasts the profiles of male senior executives, often living in the Paris region, between the ages of 30 and 34 years and living as a couple, with applicants who are very young and often students. More broadly, there is a contrast between the employed and the unemployed, which includes both young and retired people.

Sociodemographic and lexical map of applicants. The points represent the 44 modalities of the 8 variables taken into account in the multiple correspondence analysis. Discs, triangles, and squares have been distinguished to improve readability. All the variables are said to be active, which means that they contribute to the structure of the map, with the exception of the 5 classes of candidates constructed above: clearly redundant with the other variables, they are confined to the role of illustrative variables, that is, projected on the graph after the construction of it. The size of the points depends on the contribution of each to the construction of the card: for example, not having children or being single are properties much more discriminating than the sex of the candidates. Strokes have been added to link the age, number of children, and time patterns to schematize the main graph structures.

A breakdown by period gives rise to a scattering along the 2 axes: from singles in 1992–1993, to married people in 2001b along the first axis; from senior managers in the Paris region in 1992–1993 to very young people in 1998–2001a along the second axis, with a return of cohort 2001b toward the central zone of the axis in the direction of applicants in a couple and married applicants.

We also find the 5 types from the classification presented above. On axis 1 in particular, young singles with no children and single men between the ages of 35 and 39 years contrast with married applicants and applicants older than 45 years in the years 1998–2001b.


Engagement in the trials is subject to a formal recruitment procedure regulated by the Huriet Act of 1988, which provides for informed consent. This requirement involves a clear presentation of the costs of engagement, the possible risks, and that any hope of compensation would be without basis or illusory. As a result, before any investigation of the justifications for engagement expressed by applicants, it is necessary to specify the elements connected with the costs, risks, and rewards involved in the appeals of the ANRS to have a better understanding of the manner in which the so called “motivation letters” (The letters do not actually indicate the applicant's “motivation,” which implies the idea of an internal drive that may or may not be conscious, but rather his or her “reasons for taking action,” “justifications,” or “reasons.” The reasons combine, to varying degrees according to the timing and the situation of enunciation, disinterested verbalizations with the evocation of expected rewards4,5) try to comply with the assumed expectations of the recruiters.


A cost is defined as anything that the applicant will certainly have to cope with if he or she takes part. In addition to the usual secondary effects connected with vaccines, they also include pseudoseropositivity, the required monthly availability, and the suspension of any blood donation for the entire duration of inclusion and all participation in other types of trial. More generally, the recruitment procedure itself prompts a self-assessment and an assessment of one's health, particularly through clinical and psychological consultations.6

In terms of risk, the information provided is explicit with regard to women of childbearing age and what would expose the unborn child to false seropositivity. We also emphasize the fact that any belief in protection from the virus because of the vaccine would be illusory. Finally, without this being clearly indicated, there is a risk of decompensation of a potential psychological suffering linked, among other things, to pseudoseropositivity.

If we now turn to rewards, which are understood as benefits that individuals intend to receive from their engagement or that are promised to them, we can see that, in addition to the inherent elements of any appeal for donations (generosity and social engagement) and elements that are specific to the fight against AIDS (scope of the pandemic and urgency), the appeals for applications highlight an original retributive element connected with the partnership relationship that the ANRS intends to establish with volunteers.

In addition to a proposed relationship of equals between experimenter and test subject, a partnership also signals a promise that the volunteers will be offered very detailed information about the disease, its evolution, and medical progress.

Rewards considered to be “illegitimate” or illusory were also pointed out. First of all, taking part in a trial cannot provide any protection from AIDS. Second, the issue of compensation was explicitly addressed to clearly indicate that participation should be altruistic, in accordance with the law. Finally, the recruiters implicitly highlighted 2 types of reward, which they consider illegitimate: a desire to make a “sacrifice” to be a “hero” and a desire for compensation due to low self-esteem. The psychological questionnaire included questions that aimed to identify people who felt a social dissatisfaction connected in particular to their professional position.

Finally, for most of the candidates, writing a cover letter was a familiar exercise, especially in a professional context. A number of letters show that a substantial, schooly effort was made in this regard to examine all forms of possible motivation, categorize them, and put them in a hierarchical order. This indicates a concern to fulfill the request using more or less obvious strategies to take into account the variety of possible reasons while taking a distance from reasons that are considered to be less noble or less convincing and highlighting the reasons that would be more likely to secure recruitment. Two consequences ensued from this.

On the one hand, the letters are full of theory because the applicants mention a whole range of “general” considerations as to what constitutes motivation, referring freely to the usual academic analyses of the reasons for engagement. In particular, they contrast altruistic engagement with other types of engagement and seek to satisfy their own personal interest.

In addition, the strategies used seem to be highly dependent on the cognitive and rhetorical skills possessed by the applicants. Thus, we observe that the people who are the least comfortable with the task and who feel that they are under “examination” use the most conventional formulas of deference, typical of official letters, and apologies made in advance for their clumsy letters. Others, however, feel a need to make it clear that they cannot be fooled by the recruiters, and that they differentiate, of their own accord, between reasons, which could easily be used to be persuasive and reasons that might be less expected and less effective but more personal.

This last point highlights a need to consider social resources of individuals when comparing the group of applicants selected for the next stage with the group of applicants rejected in the first stage in the analysis [It is impossible to know the real reasons for rejection. It is not possible to distinguish between applicants whose letters seem inadequate and applicants whose reason for refusal is connected with criteria investigated in the questionnaire (eg, age limits). It is therefore impossible, as Luc Boltanski does based on a corpus of letters of complaint sent to the newspaper Le Monde, to “search for rules that can correlate the properties of texts, the representation of the author contained in them, and the feeling of normality or abnormality that they generate in the reader”.7].

However, the differences in skills, proven by the task of writing a cover letter, and observable through certain stylistic attributes of the letters, undoubtedly play a role in the validity judgements made by the recruiters. It is the importance of this dimension, which allows for a cautious investigation of the 2 subsamples despite the practical impossibility of systematically comparing the letters from the rejected applications and the letters from the accepted applications.


The ALCESTE software program can capture on a quite fine level the vocabulary of motives of applicants (The validity of the use of text analysis is now well-established for the study of responses to open-ended questions.8 The relatively short length of the cover letters that make up our corpus allows them to be assimilated in this type of data without the problems associated with responses to open-ended questions. Although there may be difficulties related to the transcription of the answers,9,10 in this case, the processed data are transcribed without loss of information, after anonymization and possible spelling corrections.) by dividing the corpus into “elementary units of context” (EUC) that are then regrouped according to lexical concordances into “lexical worlds” (The corpus comprises a sample of 1662 application letters chosen randomly and in proportion to the total numbers of applicants for each year studied.). The software makes it possible to suggest more than to define autonomous universes of meaning, hence the need for a return to a reading and a manual analysis of the subcorpus specific to each class. This especially since we have chosen here to present an analysis that forbids to split each letter in as many EUCs as registers invoked. This means that we are classifying “whole” responses, and therefore individuals, and not groups of motives. Thus, most of the time multiple types of motivations overlap, for example, engagement characterized by a personal emotional involvement may be combined with an interest in the health situation in Africa (An initial stage in the analysis which divides individual unity of contexts into the appropriate number of EUCs is described in Ref. 11. It shows that the responses are structured around 2 axes with, on the one hand, an opposition between engagement to fight against AIDS and engagement for the sake of science and, on the other hand, a concern based on a personal relationship or social activism of varying degrees of strength, on a basis which is very similar to the results in Refs. 12 and 13 for the volunteers of AIDES.).

The classification comprises 1648 units, that is, 90% of the corpus. There are 2 main registers of almost equivalent weight (Fig. 3), one characterized by emotional and personal justifications (48%) and the other by rational and general justifications (52%). Of the predominantly emotional motives, 2 classes can be distinguished on the basis of vocabulary indicating a personal and emotional impact of the disease (the “affected,” who make 37% of the corpus) and motives, which share the common characteristic that they situate participation in the trial in a context of practices relating to the donation of the body (the “donors,” who make up 11% of the corpus). The motives expressed in a rational and general register can be divided into 2 classes: class 2, in which engagement is generally connected with a need for civic and civil engagement (the “civics,” who make up 33% of the EUCs) and class 3, which comprises individuals who see their engagement as part of a scientific partnership and collaboration (the “partners,” who make up 19% of the EUCs).

Descending hierarchical classification of motives (entire corpus and indivisible individual unity of contexts).

The people in class 1 mention reasons which refer to a solidarity based on proximity and a generational identity, using a vocabulary dominated by words of kinship as well as words connected with life and death, threat and disease. The most frequent terms refer to a threat which will unfairly punish new generations who will never know the peace of sex without fear of infection.

The people in class 4 describe their application as part of their usual donation practices. What is most characteristic of this class is the specific vocabulary used when referring to these other forms of donation (blood, bone marrow, and platelets). Their justifications show an opposition to other forms of generosity, which they consider to be less self-giving (especially the donation of money), and some of them want to give something back and are generally grateful to medicine. They also express the idea that health is a form of capital, and that it is normal to share it with others. In this sense, the participants proceed either from an explicitly Christian and charitable basis or the donation implicitly fulfils a strategic and protective function in relation to the risk of a deterioration in health.

The people in class 2 feel that it is categorically imperative that they take part. The style of the letters is demonstrative. They list a series of factors, which usually consist of a hierarchy of reasons, with reasons of a universal nature at the top, followed by professional reasons, and finally, personal reasons (eg, the fact that they belong to an at-risk population), which seem to be secondary in their arguments. In this case, candidates speak here on behalf of concerns far beyond their own experience.

Finally, the applicants in class 3 express a desire to have the most accurate information possible, either in connection with their profession or because of an interest in medical matters. What the responses in this class also have in common is the expression of a particular relationship with science and, more specifically, medical research. This relationship is one of both submission and fascination. The idea of a possible partnership is particularly valued. Three types of attitude, which are not exclusive of each other, can be identified. A first group simply demonstrates interest in medical questions, which can take the form of reading popular magazines; for others, the core motives is more focused on reversing the power relations between patients and doctors. The promise of a partnership with cutting-edge scientists is a powerful incentive. Finally, some people see participation in the trials as a way of being closer to professional activities, which they have no access to. This is the case for people employed in medicine and/or social services in a job that is dominated by doctors and researchers, as well as people with family or friends who are connected to this world in some way.

In some cases, this motive is combined with a feeling of being indebted to medicine because of treatment received in the past. This feeling is expressed most frequently in the group of donors.

Based on these results, a relatively complex map can be drawn of the forms of relationship with medicine which drive the motives for taking part in the trials.


The projection of the 4 lexical universes on a plane by means of a factor analysis of the correspondences reveals the dimensions around which the volunteers' motives are structured. The resulting classification shows a main opposition between the emotional and rational registers and an opposition between civic activism and scientific partnership. This spatial representation shows the diversity of resources and strategies that the applicants mobilize to convince the adequacy of their profile to the request formulated by the ANRS.

A perceived need to make engagement in the vaccine trials consistent with the writer's life story unifies all the letters. However, for volunteers who express themselves in a personal register, the desire for engagement is based on their intimate experience, a relationship with people living with AIDS, or a personal sense of threat. These are the same people who, to prove their aptitude, refer to other altruistic engagements or a philosophy of self-giving, supported with Christian references that often refer to what has been learned in school or from their family. Others take an opposite approach and try to convince by means of a rhetoric of urgent action and social need carefully detached from any personal implication. Here, the applicants usually speak in general terms, and when they mention their personal history, it is not to demonstrate their empathy, which goes without saying in the name of a universal feeling of common humanity, but to prove that their application is serious. Thus, the applicants demonstrate their conviction mainly by referring to their jobs, whether in the professional world or an association.

Two factors explaining these strategies can be identified: one is related to class and the other is related to sex. The factor analysis of the social and biological characteristics of the applicants shows this quite clearly (Fig. 4).

Profiles of applicants: social and biological characteristics. Distribution of classified context units by correspondence factor analysis under ALCESTE.

First, most of the “affected” class are women, whereas most of the “civics” class are men. We know how much the ways in which women express their relationship to politics are classically the object of a form of exclusion, in the name of a characterization of politics and more generally of concern for social activism as necessarily calculated and rational.14,15 As such, the expression of interest in engagement in a personal and emotional mode are rejected because of both social and epistemological invisibilization.

As such, while women and men are both willing to engage in vaccine research, the former do so in a different way. We observe the following for women: the relative recurrence of the topic of family and motherhood; an emotional and subjective register combining logos and pathos; and a moral and humanist method of argumentation rather than a directly political one.16 We can also highlight the modest self-presentation of women compared with men who are more forward about their intrinsic merits, whether concerning their courage and intelligence or their close relationship with the world of research and science. It is therefore very logical in the letters from women that we find fewer references to their jobs or skills. The letters of women also express more gratitude and respect toward the recruiters. The men usually have a higher need to place themselves on an equal footing, whether in terms of engagement or their skills.

Second, the issue of class plays a role. An ease with writing, a general familiarity with the presentation of an argument, and a general ability to defend one's participation in a cause are skills that are not distributed randomly in society. It is also not surprising to observe that the 4 groups produced by the analysis in this case are quite distinctive in terms of social class, as measured on the basis of data connected with profession and industry.

Figure 4 clearly shows 2 different groups of people, with one comprising applicants in a high-status profession, often connected with the world of medicine or social services (group A), and the other comprising applicants in professions that are mostly in the social services sector and more distant from the world of medicine and science (group B).

This therefore gives us a better understanding of the basis underlying the self-presentation strategies used. Thus, those “affected,” who try to defend the adequacy of their profile by referring almost exclusively to their own experiences and emotional concerns, are also in the group with an above-average similarity to people from popular categories who are traditionally more accustomed to speaking on their own behalf than in the name of universal principles or abstract duties. Conversely, the group of “civics” comprises, unsurprisingly, the professions which accumulate both economic and cultural capital. In addition, this group typically has a certain close connection with the recruiters responsible for conducting the selection interviews because they often work in a medical or scientific field.

This split is distributed along an axis (in bold in Fig. 4) that contrasts on the one hand, the applicants who go successfully through all the stages of recruitment and, on the other, the applicants who fail at the outset of recruitment based on the content of their application letter.

If we return to the factor analysis that summarizes the synthesis of the sociodemographic analysis (Fig. 2) and we observe how the ALCESTE clusters are distributed therein (circles diamonds), we see that the split is close to the split between the sociodemographic profiles (gray triangles) of young singles without children, often students, and single men between the ages of 30–35 years and, on the other hand, married applicants and people older than 45 years. The first 2 groups tend to use rational arguments and arguments that refer to a scientific partnership between researchers and volunteers, whereas the others prefer a register of civic and emotional justifications.

The interviews conducted with the recruiters indicate that they only seem to approve of the universalist motives of the “civic” applicants and the position of equality or submission adopted by the “partners.” The more rational motives are judged to be more predictive of valid inclusion in the trials, but emotion may be an element of their motives if it remains within a limited extent. Beyond this, it may compromise the relationship with the researchers or generate dangerous decompensations for the protocol and for the health of the subject himself or herself. In addition, the competence of the “civics,” which was apparent from the distance that they were able to maintain from the situation of test subject, is reassuring. “Applicants with a brain,” to use the expression by which several recruiters refer to this skill, are more likely to understand the constraints of the protocols and to comply with them. They will be more able than others to cope with the physical, psychological, and social problems that pseudoseropositivity can induce as they go about their daily lives or when other people question their engagement.


The analysis of the characteristics of applicants at the time of inclusion in the ANRS network between 1992 and 2001 is part of the series of initial, pioneering research on the motivations of the volunteers in the first 2 cohorts of 1991 and 1992,17,18 although based on a very different theoretical basis and methodology. It allows a milestone to be set in the knowledge of the conditions of the recruitment of volunteers in a French context distinct from the context of the countries of the South19 and the United States.20 By allowing the population of applicants to be characterized synchronically and diachronically, a control group was created to which the population of selected volunteers can then be compared. This comparison at all stages of the selection process, combined with interviews with the recruiters and the more detailed statements of a sample of selected volunteers in different cohorts, goes beyond the studies that endeavor to explore the attributes and so called “motivations” of the volunteer populations only. Finally, the results presented here clearly illustrate the reality of how HIV/AIDS has become ordinary and normalized in the wake of the appearance of combination therapies and their effects on the demand for engagement in the population.2

Recruitment Campaigns of the ANRS. The Explicit Selection Criteria

The recruitment criteria consider the important matter of the ability of the volunteers to understand the key aspects of the disease and participation in the trials. This gives rise to the elimination of any person who does not initially have a clear understanding of the nature of AIDS, how an infection is contracted and the trials. “When assessing the immunogenicity of drugs and their tolerance by the human body (and not, in any case, the efficacy of the drugs) it is […] important that they do not perceive their participation in a trial as the acquisition of any form of protection, and that they continue to protect themselves when engaging in practices that expose them to HIV, in particular, sexual behavior” (VR/Volunteer Network, methodology, version of August 2002, p. 4). On this basis, “the selection aims to prevent 4 types of risk”: a clinical risk induced by a state of health incompatible with participation in a trial; a behavioural risk that would cause an increase in exposure to the risk for the volunteer; a psychological risk specific to people “who present with certain pathological features that, for example, motivate them to take part in the vaccine trial on the basis of heroism, a desire for sacrifice, transgression”; and a risk of noncompliance with the trial protocol. The selection procedure must determine as much as possible whether the volunteers are able to take part in experiments of a long duration. The defection of a volunteer could effectively compromise the interpretation of the results (VR/volunteers' network, methodology, version of August 2002, p. 4).


1. Reinert M. Alceste une méthodologie d'analyse des données textuelles et une application: aurélia de Gérard De Nerval. Bull Méthodologie Sociologique. 1990;26:24–54.
2. Rosenbrock R, Dubois-Arber F, Moers M, et al. The normalization of AIDS in Western European countries. Soc Sci Med. 2000;50:1607–1629.
3. Voegtli M, Fillieule O. Constitution, diversification and normalization of a health problem: organizing the fight against AIDS in Switzerland (1984–2005). Contemp Polit. 2012;18:200–212.
4. Fillieule O. Le désengagement militant. Paris, France: Belin; 2005.
5. Agrikoliansky E, Fillieule O. Les rétributions du militantisme. Du concept à la méthode. In: Barrault L, Gaïti B, Lehingue P, eds. Mélanges en l'honneur de Daniel Gaxie. Rennes, France: Presses Universitaires de Rennes; 2016.
6. Linard F, Souteyrand Y. Dimensions socio-culturelles et éthiques des essais. Sociétés d'Afrique et Sida. 1993;2.
7. Boltanski L, Thévenot L. De la justification: les économies de la grandeur. Paris, France: Gallimard; 1991.
8. Lebart L, Salem A. Statistique textuelle. Paris, France: Dunod; 1993.
9. Marc X. Les modalités de recueil des réponses libres en institut de sondage. Journal de la Société Française de Statistique. 2001;142:21–27.
10. Lallich-Boidin G. Quelques difficultés du traitement des questions ouvertes. Journal de la Société Française de Statistique. 2001;142:29–35.
11. Fillieule O. Les logiques de l'engagement dans les essais vaccinaux ANRS. Communication aux Journées scientifiques en sciences humaines et sociales de l'Institut de Recherche sur le Vaccin (VRI). 2015. Unpublished, Research Report.
12. Pollak M, Rosman S. Les associations de lutte contre le sida: éléments d'évaluation et de réflexion. Paris, France: rapport de recherche MIRE/EHESS/CNRS, juillet; 1989.
13. Fillieule O, Broqua C. Raisons d'agir et proximité à la maladie dans l'économie de l'engagement à AIDES, 1984–1998. In: Micoud A, Peroni M, eds. Ce qui nous relie. La Tour d'Aigues, France: Éditions de l'Aube; 2000:283–315.
14. Taylor V. Watching for vibes: bringing emotions into the study of feminist organizations. In: Marx Ferree M, Yancey Martin P, eds. Feminist Organizations: Harvest of the New Women's Movement. Philadelphia, PA: Temple University Press; 1995:223–233.
15. Marx Ferree M, Merril DA. Hot movements, cold cognition: thinking about social movements in gendered frames. Contemp Sociol. 2000;3:454–462.
16. Contamin JG. Genre et modes d'entrée dans l'action collective. L'exemple du mouvement pétitionnaire contre le projet de loi Debré. Politix. 2007;2:13–38.
17. Giami A, lavigne C. Refus d'engagement dans des essais vaccinaux contre le VIH. Psychol Française. 1997;42:291–300.
18. Giami A, Lavigne C. Motivations et représentations chez les volontaires engagés dans des essais vaccinaux contre le VIH (phase I). Psychol Française. 1996;41:173–188.
19. Couderc M. Enjeux et pratiques de la recherche médicale transnationale en Afrique [Thèse]. Aix-en Provence, France: Analyse anthropologique d'un centre de recherche clinique sur le VIH à Dakar (Sénégal); 2011.
20. Fisher JA. Medical Research for Hire: The Political Economy of Pharmaceutical Clinical Trials. New Brunswick, NJ: Rutgers University Press; 2009.

logics of commitment in vaccine trials

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.