To be maximally prepared for the ERIM-summerschool a set of exercises should be completed.
The files needed for the exercises can be downloaded here.
- Getting Started
During the summerschool you can choose to work with the software Mplus or (b)lavaan (which are R-packages – lavaan is the non-Bayesian equivalent of Mplus in R and blavaan the Bayesian version). Make your choice before the summerschool starts and complete Exercise 1 (which will teach you how to get the same results as in SPSS [or your own preferred software]).
If you decide to work with R:
Before you start, make sure that the latest version of R is installed on your laptop (this is really important!). I highly recommend to use Rstudio and run R from within Rstudio. Also, install JAGS on your laptop because this software is used in the background of blavaan.
If you decide to work with Mplus: make sure that the latest version (v8) on your laptop installed. Most of the exercises can be performed with the demo version.
Read the gentle introduction paper and perform exercise 2 (see the Dropbox link). In this exercise you will get a first feeling of how prior information can be updated with data and how background knowledge and/or sample size influences the results. Bring the results to class.
3. A simple model
In the dropbox folder you can find exercise 3 in which you will learn how to run a ‘simple’ Bayesian regression model either in Mplus or blavaan. Although it might seem really easy to switch from a classical estimator to Bayes estimation there is much more to it… but this will exactly be the topic of the summerschool J
4. Reading list
To be maximally prepared for the summerschool read the following papers:
Today I will introduce the technical background and we will discuss the WAMBS checklist.
In the exercise (step 10 of the checklist) we refer to this paper.
After the course you will have to complete a final exercise to get the ECTS-credits (to be handed in before July 12 via email).
More details about this exercise will be provided during the course, but the main idea will be to re-analyze a ‘real’ data set using Bayesian statistics following the WAMBS checklist.
- Just like yesterday, we want to predict the delay of PhD-students , but this time you will have access to a larger part of the Dataset (you might want to quickly read the original report with descriptive statistics and background information).
- Open the file containing the description of the variables (but DO NOT open the data yet).
- Out of the many variables, select 3 to max 6 variables that you think might be relevant for predicting the Gap.
- Briefly report why you chose these variables.
- Fill in Table 1 of the WAMBS checklist (so think about the priors, and try to come up with some background information, but just your gut feeling might also do).
- Run the Bayesian model following the WAMBS checklist; briefly report on all the steps (but keep it really short!)
- Also, run the same model but then using your own, non-Bayesian, software.
- Compare the results, which results would you use for a paper?