Statistician for psychology researchers: psychometrics (EFA/CFA/IRT), SEM, mediation, reliability and reviewer responses. PhD theses and Q1/Q2 in R, SPSS, Mplus.
Statistics for psychology research that defends your thesis and convinces the reviewer
It's my own field. PhD in Psychology (Cum Laude, Universitat de València): doctoral theses, Q1/Q2 articles and instrument validation in clinical, social, educational, work and health psychology. Full psychometrics (EFA, CFA, IRT, invariance), SEM, mediation, multilevel models, network analysis and reviewer responses.
If your project fits any of these, there is method. The N you have is almost always enough for a defensible analysis.
You collected your thesis data and now you hesitate between SEM, regression, multilevel or a mediation model. You need to choose well before burning time in the wrong software.
Measurement invariance, bootstrapped mediation, a Bayesian model, multiple-comparison control… The reviewer asks for it, you have a major-revision deadline and you've never done it.
You translated or built a scale and need the full validation: factor structure (EFA + CFA), omega reliability, invariance across groups, IRT and convergent/discriminant validity.
Exploratory and confirmatory factor analysis (lavaan), fit indices, omega reliability, IRT and DIF to validate your instrument end to end.
Configural, metric and scalar before comparing means across sexes, countries or time points. Essential if you compare groups on a questionnaire.
Latent variables, path analysis, measurement and structural models, robust estimators (MLR, WLSMV) and SAM for small samples.
Mediation, moderation and moderated mediation with PROCESS or lavaan, bootstrapped indirect effects and APA 7 reporting.
Nested data (students in classrooms, repeated measures), latent growth curves (LGM) and latent class growth (LCGA).
Network models in psychopathology (qgraph, bootnet) and latent class/profile analysis (LCA/LPA) to identify subgroups.
No clock running on your tab. We agree on scope, timeline and a fixed price upfront: no inflated hours, no surprise invoice at the end. Before delivery, every analysis goes through a careful review on my side: I rerun the code, re-check the assumptions and review tables and figures line by line. That's what I obsess over: that what you receive is ready to defend before your committee or reviewers without errors slipping through. If something out of scope appears, we discuss it before I touch it.
Illustrative example: a two-factor CFA with standardized loadings, factor covariance and fit indices (CFI, RMSEA, SRMR). This is the kind of figure and table I deliver, ready for the Results section of your thesis or article.
Free material on the techniques that show up most in psychology theses and articles. Start here.
Questionnaire validation, SEM, mediation, invariance, reviewer response… Fill the form and I'll respond within 24h with a tentative proposal.
Biostatistics and statistical consulting for health research · All services
It depends on the number of parameters and item quality, but N=180 is usually enough for a one- or two-dimensional CFA and for not-too-complex structural models. If the model is large for your N, I propose alternatives such as SAM (Structural After Measurement), item parcels or robust estimators. We assess it in the consultation.
Yes, it's one of the most common requests. I run the configural → metric → scalar sequence (and partial if needed), interpret the CFI/RMSEA changes and draft the Results paragraph and the reviewer response. Typical turnaround: 48-72h from receiving data and letter.
The full repertoire: classical tests (t, ANOVA, ANCOVA, repeated measures), regression, mixed models and the latent side (SEM, IRT, networks). What matters is choosing the analysis that fits your question, not the most sophisticated one.
That's the priority. I don't deliver tables without explanation: every analysis comes with interpretation adapted to your level, plain language and examples of how to write up results. A Zoom session before your defense is included if you want it.