ACCC offers hands-on statistical workshops for current employees and students of UIC. Fall 2016 features an SAS workshop series (Spring 2016 featured a SPSS workshop series) There is no charge for these workshops. Please review our new Statistical Software Workshop Guidelines and Registration Procedures.

All statistical workshops are free and registration is first-come/first-seated. Attendees must contact the instructor 72 hours before the class starts to request data and syntax files.

Topics covered in this workshop include an overview of the SAS layout, defining and creating SAS libraries, importing Excel spreadsheets into SAS, SAS statements “DATA”, “INFILE”, “INPUT”; creating SAS variables, row & column include/exclude commands, variable labels, date and currency definition options, PROC statements for “PRINT”, “MEANS”, “FREQ”, and tables-options. This workshop is offered twice this semester and the same material will be covered in each session.

Recommended prerequisite: Completion of a 200-level course in statistics.

Linear Multiple Regression: A start-to-finish example of some of the various procedures and options for univariate screening of data prior to producing a multiple-regression model (e.g., addressing common confounds to producing models, decision-rules for addressing missing-data issues, formulas to partially-ameliorate issues arising from data Missing Completely At Random, regression-selection methods, diagnostic plots, and detection of multicolinearity).

Recommended prerequisites: SAS or SPSS basic skills, comfort with using a syntax programming language, and completion of a 400-level course in statistics.

Taught as a hands-on lesson with a realistic sample dataset and narrative context, this workshop teaches programming protocols for the screening and cleaning of various types of data errors (e.g., missing or invalid primary-keys, alphabetical entries in numeric fields, invalid values, data mis-keyed into incorrect fields). Lessons are useful for all professionals that manage and/or analyze data.

Prerequisites: Completion of ACCC's “Introduction to SAS/SPSS” and “Linear Multiple Regression” workshops, OR equivalent knowledge acquired through academics and/or workplace experience. Comfort with using a syntax programming language.

Analysis of Variance (ANOVA) is one of the most commonly-used statistical tests. This class will cover “One-Way-Between-Subjects”, one of the most basic and frequently used models, employed to assess whether group-averages differ meaningfully due to some attribute(s) specific to one or more groups being studied.

Topics covered in this workshop include importing various genres of research data into SAS or SPSS, mathematical assumptions that must be met in order for ANOVA test results to be valid and reliable, relative effects of mathematical assumption violations on ANOVA testing, relative effectiveness of widely-accepted countermeasures to ANOVA assumption violations, using tabular or graphical means to identify assumption violations, and producing ANOVA graphs.

Recommended prerequisites: Completion of a 300- or 400-level course in statistics, completion of ACCC's “Introduction to SAS/SPSS” and “Linear Multiple Regression"; OR equivalent knowledge acquired through academics and/or workplace experience. Comfort with using a syntax programming language.

Logistic Regression is the third-leg of the mathematically logical triangle, which includes Linear Regression and ANOVA. This class will cover Logistic Binary Regression, which is employed to build predictive models when the outcome of research interest is dichotomous (e.g., “patient died/survived” or “customer purchased/didn’t purchase”). Like the Linear Regression and ANOVA workshops, this class is comprised of hands-on, applied, SPSS or SAS programming exercises. To instill the course with “real world” context and also provide a sense of topical continuity, all instruction will revolve around the same dataset: an authentic medical-trauma research study.

Topics covered in this workshop include: 1) a brief refresher on the mathematics that drive Logistic Binary Regression and of the circumstances under which this is the best procedure to use for predictive-model building; 2) extensive diagnostic data-screening to assess whether mathematical and scientific assumptions have been met sufficiently for a constructed model’s predictions to be valid and reliable; 3) individual case-deletion diagnostics to identify outliers that disproportionately influence the estimation of correlation coefficients during the building of the predictive model; 4) case-by-case qualitative analysis of the model’s predictive failures; 5) Plots, ROC Curve and S-graphs, and, 6) a brief introduction to the topic of “Propensity Score Matching” as a countermeasure for predictive-modeling confounds occasioned by substantial aggregate dissimilarities between Control and Treatment groups.

Recommended prerequisites: completion of 400-level statistics classes, an aptitude for and an interest in multivariate statistical testing; and, a level of competency writing SPSS or SAS programming language which is equivalent to having completed the other four ACCC-sponsored SPSS or SAS workshops.

Due to its particular usefulness as a hypothesis-test for evaluating experimental treatments in human-subjects studies over time, r-ANOVA is widely used for clinical research as well as many other topically diverse areas like agriculture, economics, engineering, marketing, and psychology.

This workshop will cover 1) a brief refresher of the r-ANOVA formula and of the "mathematical mechanics" that it performs, 2) some conditions under which r-ANOVA may - or may not - be the most appropriate hypothesis-test for a dataset, 3) a few basic diagnostic protocols to help determine if, and how much, the mathematical properties of a dataset could weaken or even confound r-ANOVA test-results, 4) practical examples of some types of research confounds that cannot be directly assessed through statistical diagnostics-tests, 5) methods commonly used to partially-compensate for some violations of assumptions, 6) the SPSS programming syntax required to perform metric-diagnostic tests, run the F-Test and alternatives to the F-Test, run post-hoc tests and alternative post-hoc tests, generate graphs for diagnostic-assessments and presentation of r-ANOVA test-results; 7) reporting of test-results and interpretations.

Recommended prerequisites: completion of 400-level statistics classes, an aptitude for and an interest in multivariate statistical testing; and, a level of competency writing SPSS or SAS programming language which is equivalent to having completed the other four ACCC-sponsored SPSS or SAS workshops.

All cameras & recording equipment are prohibited and electronic devices must be turned off during the workshops.

Last updated:

October 31, 2016