Elsevier

Social Science Research

Volume 53, September 2015, Pages 300-310
Social Science Research

Scientific consensus, the law, and same sex parenting outcomes

https://doi.org/10.1016/j.ssresearch.2015.06.008Get rights and content

Highlights

  • We demonstrate how citation networks can evaluate scientific consensus.

  • Estimating consensus requires a dynamic temporal conceptualization.

  • Consensus on outcomes for children of same-sex parents arose by the late 1990s.

  • Our estimation procedure is relatively robust to noise in the data.

  • That consensus reflects “no differences” on most examined outcomes.

Abstract

While the US Supreme Court was considering two related cases involving the constitutionality of same-sex marriage, one major question informing that decision was whether scientific research had achieved consensus regarding how children of same-sex couples fare. Determining the extent of consensus has become a key aspect of how social science evidence and testimony is accepted by the courts. Here, we show how a method of analyzing temporal patterns in citation networks can be used to assess the state of social scientific literature as a means to inform just such a question. Patterns of clustering within these citation networks reveal whether and when consensus arises within a scientific field. We find that the literature on outcomes for children of same-sex parents is marked by scientific consensus that they experience “no differences” compared to children from other parental configurations.

Introduction

A central component of the legal debate concerning recognition of same-sex unions addressed whether and how outcomes for children raised by same-sex parents differ in comparison to children raised in other family configurations. Decisions in two cases by the US Supreme Court hinged, in part, on the scientific consensus regarding these questions (United States V. Windsor, 2013, Hollingsworth v. Perry, 2013). Reflecting the centrality of childrens’ outcomes in this decision, during oral arguments, Justice Scalia suggested that “there’s considerable disagreement [a]mong sociologists as to what the consequences of raising a child in a [s]ingle-sex family, whether that is harmful to the child or not” (Hollingsworth v. Perry, 2013: 19). Accordingly, scientific consensus became a central question for scientific papers and amici curiae briefs addressing this case (APA, 2013a, APA, 2013b, ASA, 2013, Pafford et al., 2013, Siegel et al., 2013). For example, a brief submitted by the American Sociological Association (ASA) concludes, “Whether a child is raised by same-sex or opposite-sex parents has no bearing on a child’s wellbeing” (ASA, 2013: 3); and Siegel and colleagues recommend that decisions about parenting competency should be made “without regard to their parents’ gender or sexual orientation” (Siegel et al., 2013: 829). In this recent case and others, such as Wal-Mart v. Dukes (2011), the relationship between social science and the law has been brought to the fore.

Indeed, a significant strain of legal thought relies, in part, on social science for guidance in arguing and adjudicating legal cases (Monahan and Walker, 2009). For example, expert witnesses trained in a social science may attest to the likelihood that a defendant will re-offend during the sentencing phase of a criminal trial (Uggen and Inderbitzin, 2010). Alternatively, more general evidence of the persistence of racial disparities may influence a judge’s decision about a potential violation of the Voting Rights act (for a recent example, see references to sociologist Vincent J. Roscigno’s expert testimony in NAACP v. Husted, 2014). The success of these empirically-based arguments is likely to depend in part on where scientific consensus lies on legal issues confronting the courts (Acker, 1990).

Social science has become widely incorporated into the American legal system. Indeed, every Supreme Court Justice at least since 1986 has cited social science in at least one opinion or joined an opinion that has (Acker, 1990, Monahan and Walker, 1986).1 With this in mind, lawyers have, for some time, tried to wrangle social science to their favor collecting amicus curiae briefs and hiring social scientists as expert witnesses. Social science becomes a key element of many legal cases and the balance of these cases often hinges on arguments about the trust placed in empirical research (Collins and Evans, 2002). Further, this trust is often explicitly linked to whether a particular social scientific issue has reached a state of consensus. Although a good deal of research addresses the formation (e.g. Shwed and Bearman, 2010) and perception of scientific consensus within society generally (e.g. Oreskes (2004) on climate change), less attention has connected the tools of sociology of science to key issues before the courts. Lawyers, juries, and judges are left to determine scientific consensus on their own (Acker, 1990).

Systematic approaches from the sociology of science provide strategies for evaluating scientific consensus and can assist in the legal determination of the value of social science research. These approaches are particularly needed for social science evidence that aims to address issues beyond narrow determinations about issues specific to a case to, as Acker states, “general, empirical propositions about social events or relationships that may be instrumental to legal rule-making” (1990: 26). This evidence most often enters the legal record via amici curiae briefs, which may—or may not—necessarily directly engage scientific consensus on the question(s) before them.

Shwed and Bearman (2010) demonstrate one method for identifying when scientific consensus emerges. This method employs citation networks among the relevant scientific literature to provide a structural overview of consensus within a scientific community as a means of capturing content shifts. Like Shwed and Bearman (2010), we conceive of consensus as a type of closure. In other words, consensus forms when boundaries around a contentious issue are resolved or enclose around one correct answer (Shwed and Bearman, 2010, Collins and Evans, 2002). This type of consensus forms a key dimension of how and whether social scientific research is accepted within the courts (Daubert v. Merrell, 1993). Here, we build upon their method to examine whether the research on outcomes for children of same-sex parents is marked by scientific consensus, and estimate when the attained level of consensus was achieved. These techniques can inform court decision-making, but are also worth answering independent of whether “no difference” should be the basis for adjudicating a legal decision about union legalization (Yurkewicz, 2012).

We adapt Shwed and Bearman’s approach to the case of same-sex parenting outcomes. Our adaptation is consistent with their theoretical and methodological bases in two ways. First, a purely search-term based means for identifying the salient literature has some limitations, which requires the identification of strategies for estimating the robustness of identified patterns to potential noise in the data. Second, we note that in absence of expert “consensus statements,” akin to the one made by the Intergovernmental Panel on Climate Change about anthropomorphic causes of climate change, the method also requires scientific consensus to be interpreted as a process, not solely an end-state. To preview, our analyses reveal that consensus exists on outcomes for children of same-sex parents, and arose early in the twenty first century.

Section snippets

Background

The relationship between scientific consensus and policy is complicated. On one hand, policies can respond to scientific consensus, putting into practice recommendations that stem from empirical evidence. Alternatively, policies may be enacted prior to the identification of scientific consensus, or even agnostic to the empirical basis for such policy recommendations. Either way, if policy is to be linked to empirical scientific evidence—whether in the former situation as the basis of enacting

Data

We begin by identifying a corpus of literature on same-sex parenting, as represented within ISI Web of Science. From this corpus, we extract an analytic sample of 19,430 publications. We identified this literature corpus from a search for all sources that included the following terms in their topic: (same∗sex OR homosexual OR gay OR lesbian) AND parent∗, which did not restrict the dates of publications returned by this search, and thus started from a list of 21,369 publications between 1900

Methods

SB estimate the evolution of scientific consensus based on the number and extent of divisions between scientific “communities” in the field. Communities identify subsections of the network where ties are predominantly within group, and limited across groups (Fortunato, 2010, Porter et al., 2009). This community structure can be summarized with a modularity score (Brandes et al., 2007, Newman, 2006). We compute this modularity score for each of the temporal slices, using the edge-betweenness

Consensus trajectory

Fig. 1 presents the modularity trends for moving windows ending from 1977 to 2013. The pattern in our case is consistent with two examples in SB: research on anthropogenic sources of climate change and carcinogenicity of ultraviolet radiation. In each of these cases including our own, the trend is marked by an early period of high modularity (here occurring in periods that end prior to the mid 1980s), or strong community segmentation, indicative of early epistemic rivalries in the field. This

Discussion

Consensus is not something that should be considered to simply exist or not. Instead, as the temporal evolution approach used here demonstrates, its state at any given point in time should be seen as a cross-sectional snapshot of a process that unfolds over time. For the questions posed here – whether there is consensus on outcomes for children of same-sex parents, when and how it arose, and what it is composed of – this dynamic perspective has three implications for interpreting the answers

Conclusion

Previously the method described by Shwed and Bearman (2010) was primarily used to calibrate when known consensus was achieved (Navon and Shwed, 2012, Shwed and Bearman, 2010), to identify cases where it has not yet arrived (Shwed and Bearman, 2010), or make sense of anomalous patterns in such trajectories (Shwed, 2015). Here, we have shown how the approach can also be used as a systematic means for adjudicating between competing claims when consensus remains contested. The scientific community

References (57)

  • ASA, American Sociological Association, 2013. Brief of Amicus Curiae in Support of Respondent Kristin M. Perry and...
  • Vladimir Batagelj et al.

    Pajek: analysis and visualization of large networks

  • Donald N. Bersoff

    Social science data and the supreme court: lockhart as a case in point

    Am. Psychol.

    (1987)
  • Bielby, Willam T., 2003. Expert Report of William T. Bielby, Ph.D., Betty Dukes, et al. v. Wal-Mart Stores....
  • Vincent D. Blondel et al.

    Fast unfolding of communities in large networks

    J. Stat. Mech.

    (2008)
  • Ulrik Brandes et al.

    On finding graph clusterings with maximum modularity

    Lect. Notes Comput. Sci.

    (2007)
  • Brown v. Entertainment Merchants Association, 2011. U.S. Supreme Court, 564...
  • Cheng, Simon, Powell, Brian, 2015. Measurement, methods, and divergent patterns: reassessing the effects of same-sex...
  • H.M. Collins et al.

    The third wave of science studies: studies of expertise and experience

    Soc. Stud. Sci.

    (2002)
  • Gábor Csárdi et al.

    The Igraph software package for complex network research

    Int. J. Complex Syst.

    (2006)
  • Daubert V. Merrell Dow Pharmaceuticals Inc., 1993. U.S. Supreme Court. 509 U.S....
  • Ben K. Grunwald

    Suboptimal social science and judicial precedent

    Univ. Pennsyl. Law Rev.

    (2012)
  • Mark L. Hatzenbuehler et al.

    The impact of institutional discrimination on psychiatric disorders in lesbian, gay and bisexual populations: a prospective study

    Am. J. Public Health

    (2010)
  • G.M. Herek

    Legal recognition of same-sex relationships in the United States: a social science perspective

    Am. Psychol.

    (2006)
  • Hollingsworth v. Perry, 2013. U.S. Supreme Court 570...
  • David C. Kimball

    Judges are not social scientists (yet)

    Elect. Law J.

    (2013)
  • Lockhart v. McCree, 1986. U.S. Supreme Court, 476 U.S....
  • Wendy D. Manning et al.

    The relative stability of cohabiting and marital unions for children

    Popul. Res. Policy Rev.

    (2004)
  • Cited by (42)

    • Clouding climate science: A comparative network and text analysis of consensus and anti-consensus scientists

      2023, Social Networks
      Citation Excerpt :

      Consensus - a centerpiece of the climate science debate - consists of a type of closure or stabilization that indicates agreement around a particular set of ideas or facts (Pinch and Bijker, 1984). Consensus is dynamic and subject to some change, although change may not occur easily (Adams and Light, 2015). Shwed and Bearman (2017) describe how consensus and contestation follow formal trajectories where contestation resolves into consensus and stabilizes or persists.

    • Family Diversity in Didactic Materials for Early Childhood Education

      2023, International Journal of Sociology of Education
    View all citing articles on Scopus
    View full text