Researchers looking for advice or insight on the best way to perform statistical analysis or design studies find help at a U-M office focused on their needs.
The Center for Statistical Consultation and Research provides free consulting on data analysis and statistical methods to all U-M researchers. Staff members versed in the latest statistical and data collection methods can assist with the design, analysis, interpretation and communication of research studies in which data and computation play a role.
more information
“CSCAR is a small unit, but we are a friendly and knowledgeable group that can help university researchers with a variety of statistical questions,” says Brenda Gillespie, research scientist and interim CSCAR director.
The most common questions are from researchers looking for the best statistical analysis method to answer their research questions using data already collected, Gillespie says. Other questions come typically at the planning stages of a study, as researchers seek an appropriate sample size or study design.
“Some questions are specific to statistical software syntax or availability. Many questions are related to data analyses for dissertations or theses, manuscripts being prepared for publication, or grant proposals,” she says. Answers to these questions can be straightforward. Or, they may require digging through books or articles, trying software options or bringing the problem to a CSCAR staff meeting.
“Sometimes more than one option is available, and we consider the pros and cons of each,” Gillespie says.
Lorrie Carbone, a clinical social worker conducting research related to childhood epilepsy, along with Dr. Renee Shellhaas, pediatric epileptologist, turned to CSCAR for help and ultimately set up a contract to have CSCAR perform statistical analysis of questionnaires filled out by teen patients and their parents.
“We met in person with our CSCAR consultant, Missy Plegue, a couple of times and did most of the interactions via email,” Carbone says. “Ms. Plegue helped us create reports of initial results right after the group intervention was complete and a final report when the four-month results had come in. She has been very responsive to our questions about the reports.”
Plegue, a senior statistician, also helped with a poster presentation and a journal submission related to the research.
“CSCAR has been very helpful for me as a clinician who started to do research after 25 years of clinical work,” Carbone says.
She has also taken two CSCAR workshops to learn about creating databases and general statistical concepts. Workshop offerings include Statistics Review, SAS, SPSS, Stata, Linear Regression, Logistic Regression, Structural Equation Modeling, Analysis of Complex Survey Data, and Analysis with R.
CSCAR has recently added staff with expertise in genetic statistics, big data, social science applications and questionnaire development.
“These areas complement our expertise in other areas such as regression modeling including mixed models, survey sampling, meta-analysis, qualitative analysis, clinical trials and multivariate methods,” says Gillespie.
The field of statistics is rapidly developing, and the quantity of data captured is growing by leaps and bounds, she says. “CSCAR is poised to meet new challenges, and help guide university researchers using all available tools.”
