ONLINE COURSE – Path analysis, structural equations and causal inference for biologists (PSCB03) This course will be delivered live
20 October 2025 - 30 October 2025
£480.00
Event Date
Monday, October 20th, 2025
Course Format
This is a ‘LIVE COURSE’ – the instructors will be delivering lectures and coaching attendees through the accompanying computer practical’s via video link, a good internet connection is essential.
COURSE PROGRAM
TIME ZONE – Eastern Daylight Time – however all sessions will be recorded and made available allowing attendees from different time zones to follow.
Please email oliverhooker@prstatistics.com for full details or to discuss how we can accommodate you).
About This Course
This course, based primarily on my 2016 book, teaches you how to use path analysis and structural equations modelling to test causal hypotheses using observational data that is typical of research in ecology and evolution. It is taught in half-day sessions so that you can practice individually after each half-day session. You will learn how to conduct these tests, why (and
when) they are justified, and how to interpret the results. The first few lectures will primarily present the theory but practical sessions will become more prominent later in the course. The
practical work will be based on R and RStudio. Students will receive R script, datasets, and a list of R packages to install. It is highly recommended that each student have a copy of my 2016 book for the course, but not essential.
By the end of the course, participants should be able to:
Understand the logical relationships between d-separation, data, and causal hypotheses.
Know when to use piecewise SEM, when to use covariance- based SEM, and the advantages/disadvantages and assumptions of each
Be able to construct, test, and interpret measurement models involving latent variables
Be able to construct and identify equivalent models
Be able to incorporate nested or mixed models, multigroup models, and non-normal distributions into SEM
Participants are encouraged to bring their own data, as there will be opportunities throughout the course to plan, analyze, and receive feedback on structural equation models.
Intended Audiences
Scientists generally, and ecologists specifically, who want to test hypotheses concerning cause-and-effect relationships involving several variables, especially involving observational data.
Venue
Delivered remotely
Course Details
Availability – TBC
Duration – 7 x 1/2 days
Contact hours – Approx. 25 hours
ECT’s – Equal to 2 ECT’s
Language – English
Teaching Format
This course involves a mixture of theory and practical work. Data and analytical approaches will be presented in a lecture format to explain key concepts. Statistical analyses will then be presented using R. All R script that the instructor uses during these sessions will be shared with participants, and R script will be presented and explained.
Assumed quantative knowledge
Experience in using R and RStudio for statistical analysis.
A basic understanding of statistical inference and regression methods.
A familiarity of more advanced regression models (mixed models, generalized linear models) is an asset but is not essential.
Assumed computer background
Equipment and software requirements
A computer with the most recent version of R and RStudio is required. R and RStudio are both available as free and open source software for PCs, Macs, and Linux computers. A full list of required packages will be made available to participants prior to the course.
https://cran.r-project.org/
Download RStudio
UNSURE ABOUT SUITABLILITY THEN PLEASE ASK oliverhooker@prstatistics.com
PLEASE READ – CANCELLATION POLICY
Cancellations are accepted up to 28 days before the course start date subject to a 25% cancellation fee. Cancellations later than this may be considered, contact oliverhooker@prstatistics.com. Failure to attend will result in the full cost of the course being charged. In the unfortunate event that a course is cancelled due to unforeseen circumstances a full refund of the course fees will be credited
If you are unsure about course suitability, please get in touch by email to find out more oliverhooker@prstatistics.com
COURSE PROGRAMME
Monday 20th
Day 1 08:30 – 12:00
Causal inference using experiments vs. observations (1h)
Randomised experiments are the gold standard
Limitations on randomised experiments
The logic of controlled experiments
Limitations of controlled experiments
Physical control vs. observational control DAGs, d-separation and data (2h)
Translating from the language of causality to the language of statistics
Directed acyclic graphs (DAGS) and d-separation
D-separation and statistical conditioning
The difference between experimental control and statistical conditioning
The logic of causal inference using d-separation
Tuesday 21st
Day 2 08:30 – 12:00
Path analysis using piecewise structural equation modelling (1h30)
D-separation basis sets of a DAG
The steps in conducting a piecewise SEM
Rejecting or provisionally accepting your path model
Path coefficients as measures of direct causal effect
Decomposing causal effects
Practical work (2h)
Wednesday 22nd
Day 3 8:30 – 12:00
Path analysis using piecewiseSEM (2h30)
The piecewiseSEM library in R
Practical work (1h)
Thursday 23rd
Day 4 08:30 – 12:00
Equivalent models and AIC statistics (2h)
Statistical power in SEM
Provisionally accepting a causal hypothesis
What is a “d-separation equivalent” DAG
Rules for identifying equivalent models
AIC statistic to compare between non-equivalent models
How to interpret AIC statistics
Practical work (1h30)
Friday 24th
Day 5 08:30 – 12:00
Covariance-based path analysis (2h)
Translating the DAG into “structural equations”
The model-predicted covariance matrix
An intuitive explanation of maximum likelihood estimation
Estimating the free parameters via ML
The concept of “degrees of freedom”
The ML chi-squared statistic of model fit
Rejecting (or not) your SE model
Covariance-based path analysis using lavaan (1h30)
Monday 27th
Day 6 08:30 – 11:30
Latent variables and measurement models (3h)
Removing latent variables from a DAG
DAGs and MAGs
DAG.to.MAG() function
When you can’t remove a latent: measurement models
Measurement models and ML estimation
Fixing the scale of a latent variable
Measurement models and minimum degrees of freedom
Measurement models in lavaan
Empirical example: measuring soil fertility
Tuesday 28th
Day 7 08:30 – 12:00
Practical using measurement models (1h)
The full structural equation model (2h30)
Model identification: structural and empirical
Composite variables and composite latents
Consequences and solutions for small sample sizes
Consequences and solutions for non-normal data
Measures of approximate fit
Missing data
Reporting results in publications
Wednesday 29th
Day 8 08:30 – 12:00
Multigroup models (2h)
What is causal heterogeneity?
The concept of nested models
How to fit multigroup models in lavaan
Practical: putting everything together (1h30)
Thursday 30th
Day 9 9:00 – 12:00
Practical and group presentations of results
Bill Shipley
Bill Shipley is an experienced researcher and
teacher in plant ecology and statistical ecology. He has published four scientific monographs and over 170 peer-reviewed papers.