skip to content

Department of Psychology

Access to the course (Students with Raven)

This prospectus is revised (January 2021) in the light of the continuing pandemic and need for online teaching. The overall objective of this course is to give graduates starting research careers a good intuitive understanding of principles in statistics and data-analysis strategy, which will serve them in four ways, enabling them: (1) to critically assess the literature on and around their dissertation topic so as to be able to judge the strength (a broad concept, to be defined) of the evidence presented  independent of possibly misleading article conclusions using facile categorical existence statements; (2) to have sufficient quantitative confidence and sufficient basic competence in ‘data-language’, to efficiently and creatively use sources and consultancy with those more statistically experienced, to then design and execute well-linked data acquisition and data analyses which reconcile power in its wider sense (to be defined) with appropriateness, hence in-principle reproducibility); (3) to be able convincingly to explain why analyses presented take the form they do; and (4) to be able to include quantitative methods in their generic employment-oriented skills. Whilst admitting certain legitimate uses of the p-value, the course aims to make a complete break with the past of psychology in which its over-use was trivialising, fundamentally unscientific and harmful. It does not cover methods claimed as ‘AI’ as such, but refers in passing to the dangers of non-replicability and non-transparency in many applications of AI as seen from a statistical perspective.


Below you will find the rough outline of each session of this course.

  1. 1-hour recorded online lectures will be released at least a week before the discussion session.
  2. 1-hour discussion session on the dates listed below between 4-5pm
Date Topic.

Thursday 18 February 2021 at 4pm                          

Measurement levels, power and the general linear model (GLM – multivariable regression) including logistic regression: why we should do things in a particular way eg using the most powerful of the appropriate tests. Limits to extreme replicationism.

Access to the Course (Students with Raven)

Thursday 25 February 2021 at 4pm

The many (related) senses of errors and what to do about them: distributions and their importance outside the (attenuated) requirement of normality for residual error distributions

Access to the Course (Students with Raven)

Thursday 4 March 2021 at 4pm   

The range of multivariate techniques based on GLM and when they can be useful or necessary: principal component & factor analysis, canonical correlation, classification (clustering and LDA); MLM; SEM#

Access to the Course (Students with Raven)

Thursday 11 March 2021 at 4pm   

Modern developments in non-parametric techniques: when to use bootstrapping

Access to the Course (Students with Raven)

Thursday 18 March 2021 at 4pm 

Statistical strategy as part of project planning, applied between and within analyses. Role of preliminary analyses and making reporting efficient yet adequate for honesty to be judged. Multi-versing to minimise the forking path problem. What the ‘Limitations and future research’ section should say.

Access to the Course (Students with Raven)

#Brief definition only. It is hoped to hold again a physically isolated SEM workshop in the department in the summer term.