Summary Lecture Notes Advanced Research Methods | EUR | 2025/26

Advanced research methods
Lecture 1.1: Knowledge clip 1 – Directed Acyclic Graphs
After watching this knowledge video you will be able to:
 Understand DAG terminology
 Apply DAG rules to answer a research question

DAG – Theory
Directed Acyclic Graphs (DAGs) are graphical representions of the causal
structure underlying a research question:




Information about X
- How long to the gym
- What exercises
Also need information about diet; smoking, healthy or unhealthy, alcohol or
underlying chronic illnesses.
o Both can effect their health or gym exposure
o You need to know the causal structure of your study, so you can see
your underlying of going to the gym and health outcome
 DAGs help to visualize the causal structure underlying a research question
 You need a priori theoretical/subject knowledge about the causal structure
to draw a DAG (e.g., from previous studies, literature, common sense, …)
 Collect data on all relevant variables
 ‘Simple’ rules can be applied to determine for which variables to adjust in
regression analysis and how to interpret the results

DAG terminology
- Paths
- Causal paths and backdoor paths
- Open and closed paths, and colliders
- Blocking open paths (associotion between exposure and outcome consist
of all open paths and some need to be blocked)
- Opening blocked path
o Avoid opening backdoor paths at all costs

, o All backdoor paths need to be closed to be able to answer your
research question in an unbiased matter. Isolate the association
between exposure and outcome, then causal

Paths




 Paths are directly related to your research question and begin with x and
end with Y
 Fallow any route between x and Y
 A path is any route between exposure X and outcome Y
 Paths do not have to follow the direction of the arrows!!!!
Q: How many paths between X and Y?  4 (welke route kan je bewandelen naar
Y, beginnend bij x)
- Causal structure thay underlines the RQ is underlined by 4 paths
- Only interested in association between x and y, having data on the
association between 4 paths is important for answering the RQ in an
unbiased manner

Causal paths and backdoor paths
 A causal path follows the direction of the arrows
 A backdoor path does not
Q: Which are causal or backdoor paths?
- Two causal paths; X-V-Y and XY
- Two backdoor paths; X-W-Y and X-L-Y

Open and closed paths
 All paths are open, unless they collide somewhere on a path
 A path is closed if arrows collide in one variable on that path
Q: How many paths are open and closed?
- W is a collider; path is closed
o Path is blocked because the arrow
- Open paths go from the exposure to the outcome, if arrows collide in one
variable on a path, this path is blocked

,Blocking open paths (II)




 Open (if arrows do not collide) (causal= fallow the direction path or
backdoor= does not fallow the direction) paths transmit association
 The association between X and Y consists of
the combination of all open paths between them
 Here: all paths except X  W  Y
Examine the influence on X to Y, remove all associations and that are not
relevant and by blocking those open paths
- We block all open backdoor paths by including an variabele on that
backdoor path in the regression analysis, in this case ; we include variable
L in the analysis
 An open path is blocked when we adjust for a variable (L) along the
path
 This means that we remove the disruptive influence of L from
the association between X and Y
 How? By including variable L in the regression analysis
 Backdoor paths always need to be closed
 Causal paths need to be open/closed depending on RQ

Open blocked paths
 Including a collider (W) in the analysis means you open (you assume the
direction of the arrows is open and not closed) the blocked backdoor path
 This introduces bias in the association between X and Y
 Unbiased estimate; by closing the backdoor path again
 Removing a collider from the analysis
After watching this knowledge video you are expected to:
 Understand DAG terminology
 Be able to apply DAG rules to answer a
research question in an unbiased manner
(block backdoor paths!)
Not certain that you fully grasp the DAG terminology and rules yet?

, ARM – Lecture 1.1/1.2 - Introduction to
causal inference
After lecture 1.2 you will be able to:
 Explain the three different reasons for examining (statistical) associations
 Explain the potential outcomes approach in causal inference
 Define ‘causal effect’
 Apply the concepts of consistency, positivity, and exchangeability to make
a causal claim




What is the effect lead to a certain outcome
- Focus is on the x variable, influence on whatever outcome
- X variables are named independent variables, exposure variable, they
are all the same variables
- What would happen if the x-factor is different
o When we know we can say if there is an effect

Example of the loreal effect
- “70% less imperfections in 4 weeks of use”
Improves the quality of your skin’ implies a causal effect:
 X leads to Y
 Use of True Match Minerals powder (X) leads to a better skin (Y)
Critics?
- Small sample size
- There is no control group
o All under dermatological control (?), but no control group
o What would have happened had the women not used the
powder?
o No information on other factors that may influence the result