TRAUMA RESEARCH -- Statistics
Statistical Analysis Procedures -- Their Role in Quantitative Research
One of the things you can look into using Wikipedia is the expression: "Lies, damned lies, and statistics" referring to types of "falsehoods". The phrase alluds to the persuasive power of numbers, particularly the use of statistics to bolster weak arguments. It is also sometimes colloquially used to doubt statistics used to prove an opponent's point. This is one issue I hope to discusse briefly on this site, as it is an issue of crucial importance.
What are the goals of scientific research in psychology? Researchers seek not only to describe behaviors and explain why these behaviors occur; they also strive to create research that can be used to predict and even change human behavior.
Key Terms to Know
Hypothesis: An educated guess about the possible relationship between two or more variables.
Variable: A factor or element that can change in observable and measurable ways.
Operational Definition: A full description of exactly how variables are defined, how they will be manipulated, and how they will be measured. This step is crucial and defines GIGO -- Garbage In, Garbarge Out -- one cannot make sense of statistical analyses -- one shouldn't even try! -- before one knows how "data" have been operationally defined, but making this mistake is one of the commonest problems in "real world research", including in the field of developmental trauma research -- for an possible example, see the discussion of the use of Covariance on my Diagnosis page [LINK].
Before a researcher can begin, they must choose a topic to study. Once an area of interest has been chosen, the researchers must then conduct a thorough review of the existing literature on the subject. This review will provide valuable information about what has already been learned about the topic and what questions remain to be answered. A literature review might involve looking at a considerable amount of written material from both books and academic journals dating back decades. The relevant information collected by the researcher will be presented in the introduction section of the final published study results. This background material will also help the researcher with the first major step in conducting a psychology study — formulating a hypothesis.
Initial Steps in Psychological Research -- for other steps see [LINK]
Step 1 – Forming a Testable Hypothesis This is the step wich often determines whether or not a PhD will be successful -- the point of a PhD is to satisfy the criteria of an academic institution. Whether or not it actually contributes "substantially" (to a LARGE) extent is NOT the criterion, and is in fact probably an unrealistic goal. A "testable" hypothesis should be able to be able to stated in a "thesis" consisting of between one and three sentences.
Step 2 – Devise a Study and Collect Data -- crucial to this step are the methods used to collect data; the criteria by which data are selected; and the means by which the data are analysed -- how the data is collected (from whom "data" is obtained -- qualitative or quantitative; the observation schedule, say); the statistical analyses performed) etc.
-- Issues in Application to Trauma Research --
COVARIANCE
-- Understanding Covariance -- A discussion -- [LINK]
In clinical research, data is obtained from "known groups" -- groups (sometimes of 1!) "known", or defined, as possessing some set of characteristics, often "diagnosis". One core issue in trauma research is the debate about PTSD's symptomatic presentation, linked questions about its structure, and the special characteristics of exposure to traumatic stressors -- its nature as a disorder, or "family" of disorders, including Complex PTSD in adult survivors of childhood trauma -- see articles here [LINK]. The discussion about CPTSD in Resick et al. (2012) suggests that the analysis of CPTSD as a candidate for inclusion in the DSM-5 (and by extension the ICD-11) is best examined using the notion of construct validity (Cronbach & Meehl, 1955). This is a debatable proposition because this implies that PTSD is a construct, like intelligence, not a diagnostic entity, disorder, or disease. Indeed, Meehl himself (Meehl, 1995) did not take the view that the classification problem was one of construct validity. Instead, Meehl advocated distinguishing between evidentiary and definitory criteria, and drew upon the examples of diseases in clinical medicine, where constructs are not a part of the discussion: “The...medical model does not identify disease taxa with the operationally defined syndrome; the syndrome is taken as evidentiary, not as definitory” (p. 267). A disease entity in medicine outside of psychiatry is a circumstance of pathology and etiology, and, as Meehl points out, is why one can be asymptomatic, but still have a disease — Magic Johnson and AIDS being one well-known example.
Meehl’s conceptualization led to the development of taxometrics (Waller & Meehl, 1998), one of several approaches (Meehl, 2004) aimed at determining whether category or dimension fits the pathological phenomenon. Hyman’s worry about reification of diagnoses is just an example of the tendency to introduce simplicity or shorthand labels (e.g., hypertension) even where everyone knows there is no category. It is more pernicious, however, in thinking about psychiatric or mental disorders and is the same cognitive habit of using cutpoints to analyze dimensional data.
The literature in PTSD about the phenomenology has, in my view, suffered greatly by the popularity of latent models and easily available software to conduct confirmatory factor analyses and the concomitant confining exploration of the co-variance structure of symptoms to measures that contain only the DSM-IV symptoms themselves, such as the versions of the PCL (Weathers, Litz, Herman, Huska, & Keene, 1993). Any serious attempt to look for subtypes of PTSD, or to understand the full spectrum of symptoms triggered by exposure to a traumatic stressor requires casting a much larger net than the 17 symptoms. Such studies do not look at the structure of PTSD, they look at the structure of the measures of the codified diagnostic criteria. Indeed, I would argue that the idea that there is a structure to PTSD misunderstands that manifest symptoms are probabilistic and a rare symptom not measured redundantly will not survive such analyses, but can nonetheless be among the most pathognomic indicators. Discussions of the criteria for most other disorders in the DSM-5 do not focus on the results of confirmatory factor analyses. Rare, but important symptoms cannot survive such analyses; nevertheless, they may be fully evidentiary in Meehl’s sense. Hopefully, committee deliberations will recognize that the essential core of a disorder or disease must allow not only for atypical manifestations and rare symptoms, but also will recognize that when a non-rare manifestation appears to accompany a limited set of exposures, there may indeed be a difference in kind.
Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological bulletin, 52(4), 281–302.
Meehl, P. E. (1995). Bootstraps taxometrics: Solving the classification problem in psychopathology. American Psychologist, 50(4), 266–275. doi:10.1037/0003-066X.50.4.266
Meehl, P. E. (2004). What’s in a Taxon? Journal of Abnormal Psychology, 113(1), 39–43. doi:10.1037/0021-843X.113.1.39
Waller, N. G., & Meehl, P. E. (1998). Multivariate Taxometric Procedures: Distinguishing Types From Continua. Thousand Oaks, CA: Sage.
Weathers, F. W., Litz, B., Herman, D., Huska, J., & Keene, T. M. (1993). The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility. Presented at the Annual Convention of the International Society for Traumatic Stress Studies, San Antonio, TX.
COVARIANCE -- Understanding & Application