Research Article: Methodologic approach to sampling and field-based data collection for a large-scale in-depth interview study: The Social Position and Family Formation (SPAFF) project

Date Published: January 15, 2019

Publisher: Public Library of Science

Author(s): Diana Romero, Amy Kwan, Lauren Suchman, Bryan Sykes.


Over the past several decades there have been dramatic shifts in demographic patterns pertaining to family formation, with declining and delayed marriage and childbearing, and increased cohabitation in the United States and other Western industrialized nations. These trends in family demography have been predominantly studied using large-scale datasets, which have identified total population and subgroup trends over time, including differences by age, gender, racial/ethnic, economic, educational, religious, and other characteristics. However, there is limited knowledge and understanding of how individuals across different levels of social position, as well as other important characteristics, make decisions around forming families. This lack of qualitative data on contemporary attitudes regarding family formation has hampered our ability to more completely understand the factors driving behaviors pertaining to the large-scale (ie, international) shifts in demographic trends. The Social Position and Family Formation (SPAFF) project is an in-depth interview study that used quantitative data to guide recruitment of a large sample for qualitative interview data collection on factors influencing different aspects of family formation among heterosexual females and males (18–35 years) in the context of individuals’ social position. This methodological paper describes the use of a ‘quantitatively-informed’ purposive sampling approach in a large metropolitan area to collect qualitative data (through in-depth interviews) from a large sample (n = 200), utilizing web-based tools for successful community-based recruitment and project management.

Partial Text

The field of family demography has been extensively studied using large-scale datasets from both observational surveys [1,2] as well as surveillance or administrative datasets [3]. These research approaches have been able to identify trends over time, including important differences by subgroups (e.g., age, gender, racial/ethnic, economic, educational, religious, etc.). Indeed, from these type of research activities, we know that over the past several decades, many Western industrialized countries have witnessed a dramatic shift in the phenomena associated with family formation, in particular with regard to marriage, cohabitation, and childbearing. In the United States (US), marriage rates have been on a steady decline, divorce has decreased, cohabiting unions have increased markedly, and childbearing overall has decreased [4–11]. Alongside these trends has been increasing attention directed at rates of childbearing outside of marriage, which has risen from 4% of all births in 1940 to about 40% in each year from 2007 to 2013 [12–18]. The implications of these demographic trends are broad-based, however, they are most salient for individuals at the lowest ranks of income distribution. Although strong persistent associations exist between low socioeconomic status and single parenthood [19], these findings from predominantly quantitative research still leave us uncertain of the specific mechanisms that may be at work to explain these associations.

The SPAFF project is a study that set out to conduct IDIs with a diverse sample of 200 women and men between 18 and 35 years of age. This required developing a sampling strategy that would increase the likelihood of recruiting participants in different neighborhoods with a range of characteristics and who would, therefore, reflect individuals across the socioeconomic (SES) spectrum and the demographics of the broader NYC metropolitan area. Put differently, we sought to minimize the potential bias associated with the correlation between individuals’ characteristics and the neighborhoods they live in or frequent (e.g., more vs. less affluent neighborhoods, or those with more members of a particular ethnic or religious group). In addition, in order to efficiently carry out such a large qualitative data collection effort, we designed the study to incorporate various web-based tools to assist with the purposive sampling, recruitment, data collection and management, and analysis phases. Below we describe the various elements of the study design and field-based methods.

The results of several different quantitative analyses of the sample are presented below. First, we provide recruitment data and a demographic description of the total sample interviewed. Next is the analysis of the study sample vis-à-vis the larger population from which it was drawn. Finally, we present the results of an analysis comparing study participants with those screened but who did not participate to detect potential participation bias. This is followed by a summary of the specific qualitative thematic analyses that are currently underway.

The complementary goals of this research project were to recruit a community-based sample for a large, IDI study that reflected key demographic characteristics of the target population. In order to do so effectively, we found that combined use of a host of web-based tools–from demarcating neighborhoods and identifying recruitment sites, to managing the field staff and collecting data–was integral to our success with recruiting a large, diverse sample for qualitative data collection.

The SPAFF study utilized a research design involving both ‘quantitatively-informed’ purposive sampling and qualitative data collection, supported by combined use of various web- and field-based tools to address research questions pertaining to family-formation decision making. To best understand the dramatic changes in the configuration of relationships and childbearing patterns over the past half century [5,14,15], it was clear that much more extensive, qualitative data were needed. This sentiment was strongly articulated at a working group of family, reproductive and sexual health researchers and practitioners convened by the NIH National Institute of Child Health and Human Development [44]. To some researchers, large-scale qualitative data collection may be considered generally unnecessary and possibly antithetical to the paradigm of qualitative research. Yet, specific circumstances have been described in which larger sample sizes are valid and even necessary to address the research question(s) at hand [45]. Several of those conditions pertain to this study, including heterogeneity of the population; number of key selection criteria (including “nesting” of criteria, also referred to as stratification); groups of special interest that require intensive study; and, resources available. We felt that these conditions applied to the SPAFF study; thus, to best explore factors related to family-formation decision-making there was a compelling reason to undertake such a large-scale qualitative effort.




Leave a Reply

Your email address will not be published.