Junior Faculty RFA

Participant Resource Pool

Scientist Seminars

2020

February 2020 – Put Your Suit On, We’re Getting in the Pool: Approaches to Pooling Datasets

Dr. Briana Mezuk discusses what it means to “pool” datasets. The fundamental idea is that you can take information in one study and use it to “fill in the gaps” in another study. There are several variations, or ways of doing this, but they all stem from the idea that the data you have can be used to impute the data you do not. Like all imputation strategies, these approaches can address some sources of bias, but at a cost of precision.

2020 Feb AnC Seminar Notes_Briana Mezuk (Word)

AnC Scholar Seminar PowerPoint_2.19.20 (PDF)

For MCUAAAR training on pooled analysis READ ME (PDF)

April 2020 - Doing All Things: When the World is Going to $#@! Around You

Dr. Amanda Woodward discussed time management and strategies to be productive when there is a global pandemic. She discussed issues such as planning one’s semester, aligning time with priorities, writing tips and overcoming resistance to writing, and dealing with the avalanche of email. Participants also shared their favorite tips and strategies to increase productivity.

Doing all the things – 3-18-2020 mini session  (PDF)

May 2020 - Reproducible Analytic Workflow: Tips and an Application

Dr. Wassim Tarraf shared tips on how to create a workflow for reproducible research. He presented Scott Long’s concept of research workflow which includes planning, organizing, and documenting scientific process, establishing and fostering collaborations, managing and sharing data, analyzing data, disseminating findings, and archiving the process for replication. Dr. Tarraf also walked participants through an example using Git and GitHub.

Reproducible Analytic Workflow: Tips and an Application (PDF)

October 2020 - Resources and Strategies for Conducting Research During the COVID-Era

Dr. James McNally and Dr. Briana Mezuk facilitated a discussion of data resources and strategies for staying productive during a period of time when a) data collection is either prohibited, limited, or altered to be virtual and b) “working from home” has a new set of challenges. Seminar participants were asked to view the NACDA video before the seminar, and bring their questions/ideas/et to discuss during the AnC seminar.

November 2020 - News flash! Author Mentions in Science News Reveal Wide-Spread Ethnic Bias

Mini-Training, November 11, 2020

Tips for Talking with Journalists

Dr. Amanda Woodward is the Associate Director for Academic Affairs and Research, MSW Program Director, and Professor at Wayne State University. She is also a Co-Lead of the MCUAAAR REC. Her Mini-Training had tips on one’s message, making sure you have talking points, general communication tips, and more in-depth suggestions for talking with journalists. There was a lively discussion after the presentation with scientists and faculty.

Scientist Seminar, November 18, 2020

News flash! Author Mentions in Science News Reveal Wide-Spread Ethnic Bias

David Jurgens is an Assistant Professor in the School of Information at the University of Michigan. Participants were asked to read the article found here https://arxiv.org/abs/2009.01896 prior to the seminar. He presented two studies as examples of how to think about text as data. One was based on the article participants were asked to read, and is reflected in the title of the Seminar. The other study was entitled, “Racial Disparities in Police Language Usage”. The research question was, do officers treat white community members with a greater degree of respect than they afford to black community members? Text from body cameras of officers in the Oakland Police Department was used as data in this study.

December 2020 – Mini-Training: Quantile Regression

Dr. Wassim Tarraf is an Associate Professor, at the Institute of Gerontology, Department of Healthcare Sciences, at Wayne State University. His Mini-Training was entitled, “An Alternative Modeling Approach to Uncover Disparities: Quantile Regression”. While researchers typically use ordinary least squares regression, that method has assumptions that are necessarily met, especially when you have bimodal or multimodal distributions, skewed distributions or outliers. Quantile regression is an extension of linear regression and is a modeling technique that can be used when the assumption of least squares regression cannot be met.