A developmental case methodology to address new clinical questions

Abstract

Standard research designs, literature reviews, and case study methodologies do not meet the need for development of new clinically relevant hypotheses. To the contrary, these standard forms of publication have tended to promote conformity and stifle divergent research. This paper describes a method of producing empirically based clinical hypotheses that is more open-ended than standard methods, but also more empirical than clinical opinion. A developmental ordered set of cases, each with standardized assessments and set within its own family system across two or sometimes three generations can approximate the advantages of a sequential, cross-sectional longitudinal research design. A review of the literature suggests the need for more divergent approaches to research. This approach is described here and applied to sexual signs and behavior in children in the accompanying paper.

There are several ways to write a publishable academic paper. In Sexual Signs and Behavior in Children ¹ , Andrea Landini and I do not properly use any of them, although we borrow ideas from all of them.

This is intentional. We do not think that the standard forms fit our intention to offer a divergent, but plausible, hypothesis about the function and meaning of sexual behavior directed to and used by children. Our reasons have to do with both the limitations of the usual forms of writing and the suitability of our novel methodology to clinical applications.

Limitations to the usual forms of writing for clinical application

Reviews of the existing literature

There are many sources of information. Information generated by previous researchers and theorists is highly valued and it is standard to review published studies before presenting one’s own ideas. Moreover, it is implicitly assumed that the current work will extend or challenge the previous work, thus refining the knowledge base. In Sexual Signs and Behavior in Children, we do not review the literature because the literature focuses on the pathological qualities of such behavior and does not query about possible positive functions. Reviewing such papers would force our paper to begin by highlighting the very ideas we wish to change. We fear the review would not only take space away from articulation of a different perspective, but might also reify in readers’ minds the ideas we wish to challenge. We note that very few divergent papers have been published in the last half century (Kozlow, 2023); instead, research has valued shared understanding and conformity and is often funded by corporate, governmental, or bureaucratic sources that have an interest in the outcomes (Bauer, 2004; Forschur, et al., 2023). That is, minority perspectives and contrarian ideas are persistently excluded from both funding and publication yielding an effect that is similar to censorship.

In Sexual Signs and Behavior in Children Landini and I want to propose something more radical than modifying existing understanding of sexuality with children. To do this, we chose to begin with a clean slate. We tried to observe and describe sexual behavior with and by children, as much as possible without preconceptions. This led us away from the predominant negative understandings of sexual behavior with children to its adaptive value to very threatened children and parents.

Hypothesis-testing with group designs

“Empirical” information, as defined by hypothesis-testing with group designs, is highly valued. But such studies present two problems for clinical application. First, when the hypothesis to be tested has little or no existing research, testing it in a group design can waste money if the hypothesis proves to be null. Before investing funds in sound empirical research, a foundation that suggests the probable value of the ideas is needed. For new ideas, this poses a problem.

Second, the results of group studies are too far removed from the personal realities that they purport to represent to know their meaning for clinical application. Behind the main and interactive effects reported statistically lie real people whose visceral feelings and sometimes desperate acts are not captured by the constructs, variables, and statistics that researchers use. For example, what does it mean for my client “Mary Smith” that “warmth” was correlated at .39, with a .05 probability of accuracy, with “couple stability?” What does that look like for the other dyads in the study and how did it feel for “baby #89?” How different was that from the experience of “baby #94?” A standard deviation does not address that for any specific individual. If babies #89 and #94 did not experience maternal warmth, should one intervene? How? The same way with both babies and their mothers? Because the causes of psychological states and behavior are not known with precision, empirical studies reporting group psychological and behavioral data are very limited regarding which treatments are appropriate for individuals.

Case studies can address clinical questions, but one case is woefully inadequate to address the much broader topic of either variation in any characteristic, for example, sexualized behavior among human children, or the universal aspects of a characteristic, such as developmental trends in children’s sexual behavior. We wanted to be rooted in unique experiences, while at the same time addressing universal human processes. We needed a new design suited to discovering new developmental processes.

A novel method for developing clinically meaningful hypotheses

Developing a solution

Years ago, when I was publishing my first work, I was already using a hybrid design, but I wrote it in such a way that its unique characteristics were not visible. My first two important publications (Crittenden, 1981, 1985) were drawn from samples of mother-infant dyads whom I knew very well. I had worked clinically with them for many months and sometimes even years. I had been in their homes, met the other family members, and observed them struggling with the adversities of living in impoverished communities. I had seen their babies develop. Crucially for being able to develop hypotheses, I had many, many videotapes of their dyadic behavior in a standardized free play procedure that I later named the CARE-Index (Crittenden, 1981). I was not blind to their identity and history; indeed, I could look at a printout of the data and pick out individual data points that looked incorrect. This reduced the data entry error in my dataset, while concurrently creating the possibility of bias. To counteract the possible bias, the coders who coded the videotapes were totally blind, in the research sense of not knowing the hypotheses or dyads. They knew only what they observed. The result was sound empirical data that could be analyzed and reported statistically, but which also could be read line-by-line by me with images and histories of specific dyads springing to mind. When I discussed the findings of these two studies, I was discussing individual differences among known people, set in their life contexts. Very few authors of empirical studies know their data in this personalized way. As a consequence, they tend to focus on constructs.

I wanted then and still want now to use clinical information to better understand the process of adaptation—so as to better meet the needs of individual distressed people. I note that this person-informed clinical approach is used by most clinicians, including John Bowlby in his trilogy (Bowlby, 1969, 1972, 1980), but it suffered from the central problem of all clinical reports: the lack of standardized and verifiable data. Mary Ainsworth followed this lead in her report on mother-infant dyads in Uganda, but with more systematic data collection (Ainsworth, 1967). In her longitudinal study of 32 dyads in Baltimore (Ainsworth et al., 1978/2015), Ainsworth combined detailed clinical narratives with standardized data collection. In Sexual Signs and Signals in Children, Landini and I carry this “personalized research with standardized assessments” approach further. Rather than purposefully collecting a new sample about poorly understood sexual behavior, we gathered together an array of cases with sexualized behavior that we have known over a 30-year period. We arranged them by child age, described their functioning in a standardized procedure, and asked what functions the sexual signs and behavior might serve for each dyad and how could this be seen in the videotaped or verbally transcribed data that we had. Then, having addressed each case uniquely, we asked whether there were any processes that connected the cases and whether these processes changed as the children developed.

An example of applying clinical experience to traditional research

Possibly because I worked clinically for a decade before I returned to university to be trained as a researcher, I often saw things differently than my more purely academic colleagues who studied the same risk populations. Whatever the reason for the difference, I decided early to favor the images and words of people over published, laboratory studies based on constructs and variables. These two sources of information sometimes led in differing directions. For example, when sitting together looking at the same videotapes, my academically trained colleague Mary Main saw disorganization where I saw alternative organizations adapted to unique life threats with known family members in specific neighborhood contexts.

The most crucial drawback to such a clinical approach is the introduction of personal biases. By announcing my perspective, readers have the opportunity to evaluate its shortcomings. Such bias is also present in empirical studies (in the choice of hypotheses, variables, and interpretation of the results), but the bias is hidden behind an impersonal academic writing style, thus giving the inaccurate appearance of objectivity.

Observant and insightful clinicians sometimes discover important conditions and processes earlier than academicians. The current interest in adverse childhood events provides an example. Bowlby was discussing family discord and parent separation as early as 1949 and 1944, respectively (Bowlby, 1944, 1949). Ainsworth and I wrote about danger as a critical cause of psychopathology as early as 1989 (Crittenden and Ainsworth, 1989). Felitti, in a huge national database study, announced that adverse childhood life experiences were tied to later psychopathology in 1998 (Felitti, et al., 1998). By then the DMM had moved on to address individual differences in how brain maturation interacted with parental protection and comfort to yield differing dispositional representations of the same adverse event (Crittenden and Spieker, 2019).

My point is that new methodologies are needed to capture clinically generated, developmental information (Crittenden et al., 2021). Others have addressed the issue of using case material in ways that are compatible with our approach in Sexual Signs and Behavior in Children (e.g., Miles and Huberman, 1994; Stiles, 2009) and with a broader view to the history of case methodology (e.g., Maxwell, 2012, 2021; Yin, 2017). Our approach differs somewhat in its emphasis on the developmental ordering of multiple cases to inform theory. Landini and I have arrived at this developmental, case-based approach that we use in Sexual Signs and Behavior in Children, after three decades of grappling with how to present case-generated information in a scientifically sound way. We particularly hope that the developmental perspective, including multiple generations in each case, can inform the development of clinically applicable theory. Although we do not assert that our application of these ideas is fully mature, we do propose that it points the way to more clinically relevant research.

The method that we offer might be considered an adaptation of cross-lagged longitudinal study based on case studies. It requires (a) selecting a disparate set of interpersonal cases (seeking to represent the full range of variation, especially divergent cases), (b) ordering the cases developmentally, with particular attention to clustering by periods of neurological maturation, (c) having concrete evidence, especially on validated instruments, that others can view, (d) describing the observations sequentially in behavioral terms that all viewers could accept, (e) analyzing sequences of behavior in terms of conditions and constructs that seemed to function causally, then looking for three or more similarly functioning sequences before considering the interpretation relevant, (f) only then, using external data (the backstory) about personal and community/cultural history to test and refine the interpretations and (f) drawing developmentally plausible conclusions across cases that (g) could have clinical implications for customized interventions. A crucial aspect of our “cases” was that they were family units, in our case mother-child dyads, who reciprocally shaping one another’s behavior and development and that their family experience was at least partly known by us, thus providing a “backstory” to the observed behavior. Moreover, every case that we included in Sexual Signs and Behavior in Children had been seen, analyzed, and discussed by many mental health professionals; this was necessary to reveal and eliminate the implicit biases among the viewers.

Universities have been thought to be the “refuge for lone geniuses,” but that is becoming less true (Editorial, 2023, p. 651). Instead, conformity, with incremental change, rules. Of course, seminal works should be read and referenced. But sometimes a fresh perspective is needed. With this defense of our novel methodology for generating clinically relevant hypotheses, developed in a clinical context and without external funding, we present a set of novel ideas about sexual signals and behavior used with and by children.