Karen Bandeen-Roche, PhD is a statistician / gerontologist who currently chairs the Department of Biostatistics at the Johns Hopkins Bloomberg School of Public Health. Her statistical specialty areas of research are in latent variable and multivariate outcome modeling. Her gerontologic research aims to better understand the causes and course of physical disability, cognitive decline, and frailty in older adults, so that their adverse implications can be delayed or avoided. She is an ASA Fellow and a Marvin Zelen Leadership in Statistical Science Award winner, and she has contributed extensive service to promote the statistical profession through leadership in scientific review panels and our professional societies.
+ Full Transcript
John Bailer: I’d like to welcome you to today’s Stats and Short Stories episode. Stats and Short Stories is a partnership between Miami University and the American Statistical Association. I’m John Bailer, Chair of the Department of Statistics at Miami University, and I’m joined by my colleagues Rosemary Pennington and Richard Campbell, Professors in the Department of Media, Journalism, and Film. We’re fortunate to be joined today by guest Karen Bandeen-Rouche, Hurley Dorrier Professor, and Chair of the Department of Biostatistics at John Hopkins University. She’s a statistical leader in geronmetrics, the measurement of constructs of aging. She joins us today to talk about establishing valid measurements of hard-to-measure concepts. Karen, welcome.
Karen Bandeen-Rouche: Oh, it’s good to be here, thank you.
Bailer: It’s a delight to have you. I’d like to start with, I’ll call it a simple question, but you may say, not so much. What are the dimensions you consider when evaluating the validity of measuring a concept, well say, such as frailty?
Bandeen-Rouche: Well, so, frailty turns out to be a great example of the sorts of things that validity pertains to. Taken generally, those things we often call constructs, they are theorized attributes, say, of a person that can’t be measured directly, but can be assessed using indirect measures. And so, a very first preliminary is to be sure that one has defined whatever the concept or the construct is as a very first step. And so, if I were to do that for frailty, the present day definition of the physical frailty phenotype is a biologic syndrome of decreased reserve and resistance to stressors, that results from cumulative declines across stress response, physiological systems, as well as those that are responsible for energy production, and then together cause vulnerability to adverse outcomes. So, that provides a theoretical lattice, if you will, around which we can verify that, measures that we choose to indirectly get at frailty, actually appear to be succeeding and doing so. And so, a first thing that we look at is face validity, and this is the least standard, and basically, just asks whether the measures look right. And so, if one had measures of frailty indirect measures and somehow, they were identifying 80-year-old marathon runners as being frail, that would be a measure without much face validity.
So, the next thing is content validity, and content has to do with whether one has chosen measures that appropriately reflect, in some sense, the whole construct domain, and also exclude irrelevant measures. And so, if I use the definition I said a minute ago as an example, the physical frailty measured phenotype includes, weakness, to get at the idea of dysregulated systems that may result in muscle wasting, exhaustion, and low physical activity to reflect energy deficits, weight loss, again, is systemic dysregulation, and gait speed is an integrated measure across those things. And so, these measures were chosen very deliberately to match the definition of the construct that was theorized in this case.
The next aspect is criterion validity, and that has to do with whether the measures one have chosen, associate in a logical or intended way with other variables that one would designate. So, for example, in the case of frailty, the whole reason that geriatricians are interested, is as a means to identify people who are vulnerable to bad things happening to them so they can do something about it. And so, measures of frailty ought to predict bad things happening to older people. So that would be criterion validity. And then finally, construct validity has to do with whether the measures behave as they are predicted to behave in statistics jargon, whether their distributions are consistent with behavior predicted by the theory. And so, an example in frailty would be, if frailty really is a biologic syndrome, then that, in some case, implements emergence of frailty as a critical mass of symptoms, and that is something that it can at least be partially evaluated by examining distributions of the multiple measures that comprise frailty, also might have to do with close relationships to predicted dominance or downstream outcomes.
Richard Campbell: Karen, your work is very complex, and one of the things that I like to ask our guest is about how their work is represented in journalism, because a lot of the ways that people know about frailty and aging is through the work of journalists and the stories they tell. Can you talk a little bit about what journalists maybe can do better or how they’re doing, in general, uncovering both aging, and frailty, and your work as well?
Bandeen-Rouche: I think the main thing that comes to mind, for me, is just the need on both sides to work together patiently and deliberately. And so, when I have had little frustrations working with journalists, it would be that something that I would have said ends up being misinterpreted in the attempt to bring it to a lay-level. And conversely, I’m sure that journalists become frustrated with me in terms of using too much jargon and being able to communicate at a lay-level. And so, it’s just on both parts, the willingness to invest the work together that arrives at the happy medium and that sort of a communication.
Campbell: Very good.
Bailer: Oh, that’s – well thank you for that. It’s been our pleasure to have Karen Bandeen-Rouche join us on Stats and Short Stories. Thanks, Karen.
Bandeen-Rouche: Thanks so much. I really enjoyed it.
Bailer: Oh, us too. Stats and Stories is a partnership between Miami University’s Department of Statistics and the Department of Media, Journalism, and Film and the American Statistical Association. Stay tuned, and keep following us on Twitter, or Apple Podcasts, or other sources for your favorite podcasts. If you’d like to share your thoughts on our program, send your email to, StatsAndStories@MiamiOH.edu and be sure to listen for future episodes where we discuss the statistics behind the stories and the stories behind the statistics.