This chapter provides an overview of the research design. It gives a brief description of the aims of the study, criteria applied to sample selection, the main
characteristics of the participants, the format of the interviews, and the analysis of the data.
Aims of the study
This study has three main aims. The first objective is to explain the social mechanisms that prompt cohabitation among low income groups. I analyse union formation from a twofold perspective. On the one hand, I pay attention to how it is influenced by social conditions, and on the other hand, I study how individuals deal with those conditions when making partnering decisions. A second aim is to account for the experience of cohabitation from the
perspective of those actually cohabiting. The focus is to find out what kind of arrangement cohabitation is, in particular by comparison with marriage. A third objective is to suggest a tentative explanation as to why
cohabitation has recently increased in Chile.
Sample selection and recruitment
The sample selection criteria were set to collect a sample typical of cohabitees in Chile. The first requirement was that participants should have been living together with the same partner for at least one year, so as to include people with at least a minimum of stability in their relationship. In addition, participants should have had at least one
child born of that relationship, as the focus of the study is on cohabiting people with children. Previous research
has suggested a negative impact of cohabitation on children (see the Introduction), and most cohabiting couples in
Chile have children (Rodríguez Vignoli 2005; Herrera and Valenzuela 2006).
Participants and their partner also had to be single, never married, and aged between 20 to 29 years old. These criteria aimed to select people who did not face legal
restrictions on marriage, as the legal age to marry without parental consent is 18 years. In order to address the issue of gender, both women and men were targeted.
Another criteria was to include people from urban low income groups, as cohabitation is most frequent in urban areas (Irarrázaval and Valenzuela 1993; Rodríguez Vignoli 2005), and among more deprived groups (Irarrázaval and Valenzuela 1993; Rodríguez Vignoli 2005; Herrera and
Valenzuela 2006; Salinas 2009). The majority of the Chilean population is urban, 87 percent according to the last
census (INE 2003)). Greater Santiago was selected as the geographical location of the study. This is a conurbation formed by the most populated boroughs in and around the city of Santiago. It has a population of 5.4 million inhabitants, which represents 36 percent of the national population (INE 2005). Greater Santiago is formed by all the municipalities of the province of Santiago, plus the municipalities of Puente Alto and San Bernardo (see
illustration 3.1).
Cerro Navia and La Pintana, the two most deprived municipalities of Greater Santiago, were chosen to conduct the fieldwork. In these two boroughs around 90 percent of households are classified as middle-low income, low income, or poor (Adimark 2003). They are also similar in terms of the Human Development Index (PNUD and Mideplan No date). Recruitment of participants from the low socioeconomic group (D) was favoured, as it is the largest group. Although there are no representative figures on
cohabitation at the level of municipalities, available data suggests that cohabitation is pronounced in La Pintana and Cerro Navia (Mideplan 2006a). In addition, the
Greater Santiago. La Pintana is located in the south of the city, and Cerro Navia in the north-west (see illustration 3.1). Therefore both sites are alike in variables that are relevant for studying cohabitation, yet they are
geographically distant.
Municipalities of the Greater Santiago district
Methods of socioeconomic classification use the household as the unit of analysis, which is inadequate for cohabiting couples. People who cohabit usually live intermittently with different relatives. Therefore, instead of the
household, the cohabiting couple was used as the unit of analysis for the purposes of socioeconomic classification. The educational attainment and occupation of the main
provider was assessed using the World Association of Market Research system adapted to the Chilean situation (Adimark 2000). The matrix used for socioeconomic classification is attached in Appendix 1. It should be noted that in every couple at least one partner, woman or man, was involved in paid work, and acted as the main provider. In addition, it was required that the main provider had no more education than completion of secondary education. This also ensured that they belonged to a low income group. Also, evidence shows that cohabitation is more common among people with primary or secondary education, in contrast to people with higher education (Herrera and Valenzuela 2006). Box 3.1 presents a summary of the selection criteria.
Box 3.1: Participants’ selection criteria
Twenty-four people were recruited, evenly distributed in
terms of gender and municipality of residence. The majority (19 of the 24) were from the low socio economic group,
group D, with the rest classified as middle-low (group C3). They had been cohabiting for 5 years, on average, and had an average of 1.4 children. Appendix 2 summarises the main characteristics of the interviewees.
The recruitment of participants, men in particular, was problematic. I used two strategies to recruit
participants. First, I contacted local institutions, such i. To have been living together with the same partner for at least one year
ii. To have at least one child born from that partnership iii. Single (never married)
iv. Aged 20 to 29 years v. Women and men
vi. Classified as middle-low, low, or poor in terms of socioeconomic group (C3DE)
as town councils, doctors’ surgeries, charitable organizations, and government social intervention programmes. Yet these institutions provided only a few people. A second, and more effective strategy, was to get in touch with possible participants through personal
contacts (friends and relatives). I used my own my research assistant’s social networks to get in touch with people who met the selection criteria. Thereafter, people who had
already participated in the study helped in recruiting new participants. These people were given a monetary incentive of roughly £5 for every new participant they recruited. The same amount was given to each interviewee in appreciation for having taken part in the study.
Data collection
Information was collected about the life course of each of the 24 participants. Participants also provided some
limited information about their partner and parents. Appendix 2 summarises the data given by the interviewees about their partner and parents. The life histories were gathered in two separate interview sessions, both involving verbal interviewing and the use of questionnaires (see box 3.2). The first interview was focused on interviewees’
family history and life history. Interviewers turned to the subject of cohabitation in the second session. The complete interviewing schedules and the questionnaire are attached in Appendix 3.
Box 3.2: Instruments applied for collecting participants’ life histories
Instrument Focus
1. Semi-structured interview Family history (grandparents and parents) Life history (from birth to interview) 2. Semi-structured interview Relationships, partners, & children
Opinions about cohabitation & marriage 3. Structured questionnaire Socio-economic data (interviewees, partner, and
The life history is a qualitative in-depth interview, which seeks to capture the salient experiences in a person’s life and that person’s definitions of those experiences (Taylor and Bogdan 1998:88-89). I refer to this technique as life history and not life story or narrative, to emphasize that the attention is on experiences and meanings rather than on narratives (Thompson 2009:39). Semi-structured interviews were conducted using interviewing schedules that were mostly descriptive and non-evaluative. The aim was to facilitate personal reconstructions and to avoid forced rationalizations. Interviewing guidelines were revised and changed several times to improve them: because of this, not all interviewees were asked exactly the same questions. Interviews were conducted in a flexible style, topics being introduced in different ways and at different times, to avoid disturbing the flow of the interview.
To control for the possible negative effects of the interviewers’ gender and age, my male research assistant interviewed the men, and I interviewed the women. The
majority of the interviews were conducted in settings that met the requirement of privacy, though settings varied according to gender. Most women’s interviews were carried out in the homes of the interviewees, whereas most men were interviewed at the field recruiters’ home. A few interviews were conducted in other settings, for example cafés, fast- food restaurants, or parks.
Participants were given an information sheet and were asked for their informed and voluntary consent. Translated versions of these documents are included in Appendix 4, below. Anonymity and confidentiality were assured, and participants were informed that the research would be published as a doctoral thesis, and eventually as an academic book or articles. They were also told what they would be asked to do. In particular they were told that interviews included sensitive issues such as their sexual life (Singer 2003:197).
All interviews were recorded digitally. On average, both sessions taken together lasted 2.7 hours. Women’s interviews were slightly longer than men’s, on average 3 and 2.5 hours, respectively. The fieldwork lasted 5 months and was conducted between September 2008 and January 2009.
Data analysis
Each interviewer transcribed his or her own interviews, to maximise accuracy. I also checked every transcription made by my research assistant against the corresponding audio recording. Transcriptions represent the interviewee’s
original narration: thus they include local idioms and are in Spanish, the language in which the interviews were
conducted.
NVivo and SPSS data analysis software was used.
Initially I carried out detailed coding of six interviews, balanced in terms of gender and place of residence. The specific interviews were selected for quality and variety. This first coding produced more than two hundred codes. These codes were then grouped, merged, and some deleted so as to produce a coding tree with ten core codes. These core codes were: background, partnership, life stages, gender roles, parenthood, personality, people, feelings,
sexuality, and expectations. This coding scheme was applied to the rest of the interviews. The data analysis also
involved counting the frequency of relevant events or characteristics. Appendix 5 gives a brief description of the qualitative codes and of the main data included in the quantitative database.
The research results are presented through a combination of quotations and counts. Quotations from interviews were translated into English. I undertook the translation, which aims as far as possible to preserve the particular voice of each interviewee. The names of the participants, and of their partner and parents, have been changed.
The selection of interviewees aimed to represent the kind of cohabitation most typical in Chile, and thus the sample is rather homogeneous. It presents many features in common with other qualitative studies of family life in the Chilean lower class (Raczynski and Serrano 1985; Olavarría 2001; Valdés 2005; Valdés, Saavedra et al. 2005). Thus it is reasonable to say that the sample represents an adequate illustration of families of the lower strata. By extension, it seems probable that it is also a good sample of the most common type of cohabitation found in Chile. As cohabitation is an informal arrangement, recruiting participants was particularly difficult. Government institutions could not be used to reach cohabitees, especially as people may be afraid of losing some of their benefits if they are found to be cohabiting. The following chapter presents initial findings, focused on exploring how cohabiting people relate to formal institutions such as formal employment and social welfare. These findings clearly show how elusive