The MPSI Methods Sessions
What to expect from the MPSI Methods Sessions:
From August 15th to 19th, our fantastic quantitative methodologists will offer five full-day hybrid (in-person and online) courses on advanced topics in statistical modeling, measurement, and inference. All sessions will be based in the freely available R statistical software environment.
The training has five courses:
Monday, 8/15, 9:00 am – 4:00 pm, Townsend 222
Scale Development (Dr. Wes Bonifay)
In this session, Dr. Bonifay will provide an overview of the entire psychometric test (survey/ scale/ questionnaire) construction process. This course will emphasize modern validity theory – as recommended by the APA/AERA/NCME Standards for Educational & Psychological Testing – in every phase of this process. Attendees will learn 1) to design a test blueprint that will guide and facilitate item writing, 2) to incorporate valuable feedback from substantive experts and members of the target population of test-takers, 3) to use cutting-edge statistical methods (e.g., item response theory) to analyze and refine the initial test data, 4) to establish test fairness and investigate the consequences of test use, and 5) to evaluate, finalize, disseminate, and maintain the test. No previous test construction experience is necessary.
Tuesday, 8/16, 9:00 am – 4:00 pm, Townsend 222
Practical Multilevel Modeling (Dr. Francis Huang)
Multilevel modeling (MLM) as an analytic technique for the analysis of clustered data (e.g., students within schools, patients within hospitals) has grown over the years. The workshop will introduce applied researchers to basic MLM concepts using R. Attendees will learn: 1) how to construct various types of multilevel models (i.e., unconditional, random intercept, random slope models), 2) when and how to use different forms of centering in order to properly specify models, 3) model binary outcomes, 4) deal with pesky non-convergence issues, and 5) conduct multilevel regression diagnostics. Attendees should already have a good grasp of standard regression techniques. Attendees maybe those who are new to MLM or those who may already be familiar with MLM using other software.
Wednesday, 8/17, 9:00 am – 4:00 pm, Townsend 222
Meta Analysis (Dr. Bixi Zhang)
In this session, Dr. Zhang will introduce meta-analysis and how meta-analysis can be conducted in statistical computing software. The topics will cover effect sizes, pooling effect sizes (fixed effects model and random effects model), measures of heterogeneity, subgroup analyses, and power analysis in meta-analysis. Advanced topics will also be introduced briefly in the session (e.g., robust variance estimation in meta-regression, SEM meta-analysis), which are related to the dependent effect size issue. The attendees will learn 1) methodologies behind each topic; 2) how to use R to do a meta-analysis; 3) interpretations of the results and related plots; and 4) recent development in meta-analysis. The session is designed for researchers who are interested in conducting meta-analyses and learn how to use R language to analyze their searching results of studies with statistical models. Familiarity with the R language will be helpful but not required.
Thursday, 8/18, 9:00 am – 4:00 pm, Townsend 222
The Basics of Bayesian Statistics (Dr. Sonja Winter)
This session is designed for researchers who would like to better understand Bayesian statistics and who are interested in incorporating Bayesian methods into their research practices. Using common statistical models (e.g., t‑tests or linear regression), Dr. Winter will cover 1) the key principles of Bayesian statistics, 2) how the Bayesian approach differs from the frequentist approach (e.g., p‑values and confidence intervals), 3) how to use Bayesian methods to estimate parameters and test hypotheses, and 4) how to report results from a Bayesian analysis. No previous experience with Bayesian methods is required. Attendees will be introduced to several (open-source, free) software programs. Familiarity with the R language will be helpful but not required, as all examples will be based on pre-written code made available to attendees.
Friday, 8/19, 9:00 am – 4:00 pm, Townsend 222
Non-Gaussian Causal Inference (Dr. Wolfgang Wiedermann)
This session introduces modern causal inference approaches for both, observational (non-experimental) and randomized controlled trial data. The first half of the session focuses on learning causal mechanisms (i.e., empirically deriving statements concerning cause and effect) from observational data alone through introducing principles and best practice applications of a recently proposed statistical framework called Direction Dependence Analysis. In the second half of the session, principles of distributional (causal) treatment effects (DTEs) are introduced. DTEs extend average (causal) treatment effects (ATEs) to intervention effects that manifest in changes beyond means of constructs and are modelled using a distributional regression approach called Generalized Additive Models for Location, Scale, and Shape (GAMLSS). Under the DTE principle, intervention effectiveness is allowed to manifest in any features of the outcome distribution, i.e., changes in variance, skewness, kurtosis, as well as ceiling/floor effects. GAMLSS model building guidelines will be presented using real-world data applications. Basic familiarity with the linear regression model and the R statistical programming environment are assumed.
In-Person: $250 per course
Online: $200 per course
Training Location and Travel Information
- The Broadway Columbia Hotel (.5 miles)
- Website: https://doubletree3.hilton.com/en/hotels/missouri/the-broadway-columbia-a-doubletree-by-hilton-hotel-COUTBDT/index.html?SEO_id=GMB-DT-COUTBDT
- Address: 1111 E Broadway, Columbia, MO 65201
- Phone: (573) 875‑7000
- The Tiger Hotel (.4 miles)
- Hampton Inn and Suites (.9 miles)
- Website: https://hamptoninn3.hilton.com/en/hotels/missouri/hampton-inn-and-suites-columbia-at-the-university-of-missouri-COUUMHX/index.html?SEO_id=GMB-HP-COUUMHX
- Address: 1225 Fellow’s Place Boulevard, Columbia, MO 65201
- Phone: (573) 214‑2222
- Residence Inn (2.8 miles)
- Website: https://www.marriott.com/hotels/travel/couri-residence-inn-columbia/?scid=bb1a189a-fec3-4d19-a255-54ba596febe2
- Address: 1100 Woodland Springs Ct, Columbia, MO 65202
- Phone: (573) 442‑5601
- Columbia Regional Airport (11 miles)
- American Airlines
- St. Louis International Airport (107 Miles)
- Most Major Airlines
- Kansas City International Airport (130 Miles)
- Most Major Airlines
- Airport Shuttle Service from St. Louis and Kansas City Airports: MoX
- Parking is $7 a day for parking lots and $9 per day for garages.
- Link for visitors permit: https://mu.nupark.com/portal/Account/VisitorLogin?ReturnUrl=%2Fportal%2F