Correlation vs Causation: What Can Be Inferred?

  • Context: High School 
  • Thread starter Thread starter gl0Wyrm
  • Start date Start date
  • Tags Tags
    Correlation
Click For Summary
SUMMARY

This discussion centers on the distinction between correlation and causation, emphasizing that correlation alone does not imply causation. Participants highlight the necessity of conducting randomized controlled experiments to establish causal relationships. Additionally, the importance of assessing conditional probabilities within a well-ordered stochastic process is noted as a method for evaluating causality. Understanding these concepts is crucial for anyone looking to draw valid inferences from statistical data.

PREREQUISITES
  • Basic understanding of correlation and causation concepts
  • Familiarity with randomized controlled experiments
  • Knowledge of conditional probabilities
  • Introduction to stochastic processes
NEXT STEPS
  • Study the principles of randomized controlled trials (RCTs)
  • Learn about conditional probability and its applications
  • Explore stochastic processes and their relevance in statistics
  • Investigate statistical software tools for conducting experiments, such as R or Python's SciPy library
USEFUL FOR

This discussion is beneficial for statisticians, data analysts, researchers, and anyone interested in understanding the nuances of statistical inference and the methodologies required to establish causality.

gl0Wyrm
Messages
8
Reaction score
0
If variable X is positively correlated with variable Y, under what circumstances am I allowed to infer that X does not causally prevent Y with reasonable probability?

I don't know much about statistics and I'm trying to motivate myself to learn by doing. For example: https://drive.google.com/open?id=0B9wG-PC9QbVEUTBIVmgxeElCQnc (https://drive.google.com/open?id=0B9wG-PC9QbVEZ2NEU3BvYzkwM0U)
 
Physics news on Phys.org
In general, you can't infer causality solely from correlations in observational studies. You would need to perform some randomized, controlled experiments in order to establish causality.
 
Hey gl0Wyrm.

If you want to assess causality then you should assess conditional probabilities in terms of a well ordered stochastic process.
 

Similar threads

Undergrad Correlation vs causation in biology/biochemistry
  • · Replies 21 ·
Replies
21
Views
4K
High School Does a correlation coefficient represent probability?
  • · Replies 4 ·
Replies
4
Views
6K
Graduate How to find which variables w, x, y.... 'most' influence z?
  • · Replies 7 ·
Replies
7
Views
2K
Graduate Post-Selection: Pre-existing Correlations or Action At A Distance?
  • · Replies 54 ·
2
Replies
54
Views
6K
Prob/Stats The Causal Revolution and Why You Should Study It
  • · Replies 6 ·
Replies
6
Views
3K
High School The Drunkard's Walk: A Discussion of the Text
  • · Replies 17 ·
Replies
17
Views
4K
High School Some questions about "superdeterminism" and Bell's Theorem
  • · Replies 75 ·
3
Replies
75
Views
13K
Undergrad QFT made Bohmian mechanics a non-starter: missed opportunities?
  • · Replies 376 ·
13
Replies
376
Views
25K
Graduate Probability via Expectation and Callen's criterion
  • · Replies 14 ·
Replies
14
Views
3K
High School Quantum Entanglement: Hidden Information vs. Time
  • · Replies 19 ·
Replies
19
Views
4K