Michael J. Clark
Brief Curriculum Vitae
| CSCAR | micl@umich.edu |
| Rackham Building 3551 | personal |
| Ann Arbor, MI | github |
Ph.D. Experimental Psychology, Concentration: Statistics, University of North Texas.
B.Sc. Philosophy & Psychology, Cum Laude, Texas Christian University.
I do a lot of statistical programming, and combined with my consulting experience, this means I’ve covered a lot of ground- statistically, in terms of dealing with many types of data both big and small, and with ideas, just burgeoning or traipsing well established domains. Clients come from a wide variety of backgrounds and the range of their needs is quite extensive. In a given week, I might help someone with some code to scrape the web to obtain their data in an efficient fashion, employ a Bayesian model to incorporate spatial random effects, use machine learning techniques to predict rare events, develop theoretical motivations to eventually be tested with structural equation modeling, or maybe help an undergrad understand some regression basics. Among my own work, clarity, especially via visualization, as well as reproducibility, are the goals I continually strive for. Underlying all of my efforts is a willingness to use whatever means necessary to gain more knowledge about the underlying mechanisms that produce the information (in the form of data) we seek to understand better, and a view of science as software development, continually upgraded, and never static.
Position providing statistical consulting for faculty and students from various disciplines across campus, as well as serving as analytical lead or providing consulting services for specific research projects. I also conduct workshops related to statistical programming and modeling techniques.
Position providing aid at any stage of various research projects for students, staff and faculty from various departments on campus, particularly, but not exclusive to, those of the Social Sciences. Capabilities include providing support throughout the research process in terms of getting ideas off the ground, analysis, reproducibility, code optimization, and visualization and reporting. Also conducted occasional workshops related to programming.
I’ve taught full courses at the University of North Texas, and was guest lecture for courses at Notre Dame and the University of Michigan. For undergraduates and graduates, my primary workload was quantitative courses, but when I was in the field of Psychology I also taught courses such as Advanced Cognitive Psychology, History of Psychology, and others.
In addition, I’ve taught and continue to teach dozens of workshops on numerous statistical and programming topics at for both beginning and advanced audiences. For more details see my website.
As of Summer 2018. Principal Investigator in parenthesis.
Readiness of Trauma Surgeons (Brian George)
Sleep Disorders in Youth after Adenotonsillectomy (Ronald Chervin)
Library Learning Analytics (Felichism Kabo)
George, B. C., Clark, M., et al. (in preparation) Mind the Gap: the Gulf of Perception between Resident and Attending Assessments of Autonomy During Surgical Training.
Fischer, I., et al. (in preparation) The Gap Between Resident and Faculty Perceptions of Resident Performance in the OR: a Multi-Center Trial.
Ahle, S. L., et al. (2019) Residents’ Readiness To Operate Independently May Not be Adequately Assessed by End Of Rotation Evaluations. Academic Medicine.
King, C. et al. (2018). Let’s Connect Community Mentorship Program for Youth with Peer Social Problems: Preliminary Findings from a Randomized Effectiveness Trial. Journal of Community Psychology.
Arango, A., et al. (2018). The Protective Role of Connectedness on Depression and Suicidal Ideation among Bully Victimized Youth. Journal of Clinical Child and Adolescent Psychology.
George, B. et al. (2017). Readiness of US General Surgery Residents for Independent Practice. Annals of Surgery. (link to article; 99th Altmetric percentile).
Archie, E.A., Tung, J., Clark, M., Altmann, J., Alberts, S.C. (2014). Social affiliation matters: both same-sex and opposite-sex relationships predict survival in wild female baboons. Proceedings of the Royal Society: of London Series B. (link; 99th Altmetric percentile).
I also write various documents pertaining to advanced statistical methods and programming. These can be found at my website.
I primarily work within the R statistical environment for most of my own needs, including the development of my own packages, and I have years of experience with it. I use Stan for Bayesian analysis, and Python for things such as web scraping, text analysis, and machine learning. I have experience utilizing a high-performance computing environment and parallel processing in general.
I’ve also had minor exposure to other statistical or numerical programming languages such as Julia and Matlab. Among traditional statistical programs I’ve had more recent experience with Stata, and in the past, SPSS and SAS. I’ve used various programs for more specific needs, e.g. Mplus, BUGS, and JAGS. In addition, I’ve used programs specific to visualization and text among other things.
I have notable experience with the concept of literate programming, using various languages related to it and reproducible research generally. Standard and non-standard document processing programs, editors and related (e.g. Markdown, \(\LaTeX\)). Version control via Git, primarily through GitHub, Gitlab, and BitBucket. For various needs, I’ve also used languages and applications related to web design (e.g. HTML5/CSS, Adobe products such as Dreamweaver, Illustrator, InDesign, and others such as Gimp). I have some minimal experience with other scripting languages such as JavaScript.
Just to provide some more insight into the analysis I can provide, I’ll note the following experience. Aside from traditional methods with generalized linear models, mixed models, and latent variable models, I’ve also extended those approaches to the Bayesian world and expanded upon them there. I’ve examined nonlinear relationships via additive models, gaussian processes etc. I’ve analyzed different types of networks and graphical models generally. I have explored time and space issues via mixed/multilevel/growth model frameworks, survival analysis, and spatial models for both the discrete and continuous setting. I’ve also utilized various machine learning/algorithmic approaches in a variety of settings, including out-of-memory situations and continuous data streams. I’ve also dealt with unstructured data situations such as that found in the analysis of text.
\(\cdot\) Ann Arbor R Users Group
\(\cdot\) Center for Surgical Training and Research
\(\cdot\) American Statistical Association
\(\cdot\) Association for Psychological Science
\(\cdot\) Member of several thesis and dissertation committees, and have been intimately involved with hundreds more in my role as statistical consultant.
\(\cdot\) Member of the Experimental Program Committee reviewing student applications and student progress within the Experimental program at UNT, as well as developing the program itself.
\(\cdot\) Member of the search committee overseeing faculty hires for the Experimental program at UNT.
\(\cdot\) Supervised approximately a dozen undergraduate and graduate students in lab research.
\(\cdot\) Judge for the Regional Science Fair, Behavioral and Social Sciences Fort Worth.
I like learning about new techniques, how to be a better programmer, refreshing knowledge gained in the past, etc., and to that extent I continue to go to workshops and talks, attend the local UseR! group meetings, and have even take the occasional MOOC. Engaging in statistical science means there is always something new around the corner.
Available upon request.