Statistical support
Welcome to the Stat Support page of the Swiss Center for Affective Sciences (CISA). Here you will find information on the statistical support offered at CISA, as well as an archive of recommendations, tips, R scripts, and teaching material concerning data analysis. Consultancy services are available to all researchers affiliated with CISA, in two general categories:
Regular support:
- Ambulant consulting service on statistical questions
- Assistance on data analysis (e.g., running an analysis together, interpreting output, visualization)
- Maintaining a library of statistics books as look-up references
- Communicating statistics tips by e-mail
Extended support:
- Regular workshops or seminars
- Large data projects that require specialized methods (e.g., missing data imputation, machine learning analysis, archival data)
- Contributions to research papers (e.g., writing method/results sections, proofreading)
- Contributions to grant applications (e.g., power/sample size analysis)
Conditions:
Statistical support can be contacted any time or day of the regular workweek (). Depending on the complexity of the request and the current demand, questions are typically handled between 1 hour or 5 days. Consultations that fall under regular support do not require any co-authorship for publications. For extended support this may be requested if the amount of input warrants it. Requests may be referred to another statistician if outside of our expertise (e.g., meta-analysis, time series analysis, or neuro-imaging analysis).
Workshops
Workshops are taught on a periodic basis, with the most recent slides and scripts available in the list below. Workshops are co-organized by the Swiss Doctoral School connected to CISA.
- Multilevel linear regression (Ver. 4 -- 01.03.2022):
- Part 1: Repeated measures (M)ANOVA (.pptx)
- Part 2: Random intercept and random slope model (.pptx)
- Part 3: Model selection (.pptx)
- Part 4: Hierarchical, trial-level, and longitudinal data (.pptx)
- Part 5: Miscellaneous topics (.pptx)
- R scripts (.zip)
- Non-parametric data analysis (Ver. 2 -- 17.01.2023):
- Part 1: Assumptions in parametric statistics (.pptx)
- Part 2: Non-parametric statistics - Permutation tests (.pptx)
- R scripts (.zip)
- Logistic regression (Ver. 1 -- 17.11.2022)
- Part 1: Analysis of frequency tables (.pptx)
- Part 2: Logistic regression (.pptx)
- Part 3: Paired categorical data (.pptx)
- R scripts (.zip)
- Machine learning (Ver. 6 -- 22.04.2024):
- Part 1: Introduction and principles (.pptx)
- Part 2: Machine learning models (.pptx)
- Part 3: Practical session (.pptx)
- R scripts (.zip)
- Common misconceptions in statistics (Ver. 6 -- 21.11.2021)
Mailings
The following is an archive of statistics tips distributed by e-mail. Support mails are distributed on a bi-weekly basis via the CISA mailing list.
- Flowchart for the analysis of factorial ANOVA (PDF)
- Normality of residuals assumption in regression (PDF)
- Cautions with ANCOVA in SPSS (PDF)
- Lord's paradox in 2×2 intervention designs (PDF)
- Multicollinearity in regression (PDF)
- Correlation plots (PDF)
- Weak, moderate or strong effect size? (PDF)
- Analysis of paired categorical data (PDF)
- Nonlinear correlation (PDF)
- Missing data and imputation 101 (PDF) and 102 (PDF)
- Hierarchical clustering (PDF)
- Simpson's paradox and causality (PDF)
- Heteroscedasticity (PDF)
- Types of ANOVA (.pptx)
- Sample size: More than just power (PDF)
- Mediated change in intervention studies (PDF)
- Multiple comparison problem (PDF)
- Advantages of within-subjects designs (PDF)
- Partial eta-squared in multilevel models (PDF)
- Common pitfalls in experimental design (PDF)
- Caution on the use of GLMM models (PDF)
- Multilevel regression reporting templates (.docx)
- Regression for ordinal outcomes (PDF)
- Paired correlation problem (PDF)
- Caution on mixed designs with G*Power (PDF)
- Testing the need for a random intercept (PDF)
- How many components in PCA? (PDF)
Page last updated on 05.07.2024
Ben Meuleman (Ph.D.). As statistician, I am the primary consultant on the data analysis of scientific projects at CISA. I assist in design, planning, analysis, and publication of studies to ensure correct practices and help researchers gain insight into their data. I specialize in machine learning, linear regression, longitudinal data, causal inference, and R programming, and teach workshops on these subjects.
Contact
Swiss Center for Affective Sciences
University of Geneva | Campus Biotech
Chemin des Mines 9 | CH-1202 Genève
+41 (0)22 379 09 79