Turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- Discussions
- :
- Re: statistical method which could be used that explains if A is a cause for the...

Topic Options

- Start Article
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

statistical method which could be used that explains if A is a cause for the B

Nov 5, 2018 11:08 PM
(1450 views)

Hi

What statisttical method would help once we are trying to see if parameter A is the cause of B?

as an example is the toxicity of a particular element the cause for a disorder?

thanks

8 REPLIES 8

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: statistical method which could be used that explains if A is a cause for the B

Generally a *controlled experiment* (with specific attributes) is required to establish causation or causal relationship.

You can explore the correlation or association between a continuous or categorical response, respectively, and a factor using analysis of variance (categorical factor) or linear or non-linear regression (continuous factor) or logistic regression and other methods.

See Help > Books > Fitting Linear Models.

Learn it once, use it forever!

Highlighted
##

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: statistical method which could be used that explains if A is a cause for the B

As Mark states, there is no substitute for a designed experiment (and some domain knowledge).

Correlation does not necessarily imply causation. Case in point:

http://tylervigen.com/spurious-correlations

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: statistical method which could be used that explains if A is a cause for the B

The naivity of your question suggests that you might benefit from some training in explanatory modeling. JMP and JMP Training offers such courses.

Learn it once, use it forever!

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: statistical method which could be used that explains if A is a cause for the B

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: statistical method which could be used that explains if A is a cause for the B

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: statistical method which could be used that explains if A is a cause for the B

There are many models to choose from. You have not shared the data or the results of the regression analysis so far, so it is difficult to guide you further.

There are empirical models and theoretical models. There are models for testing hypotheses and models for interpolation. Of course, there are many methods for estimating such models. I am sure that the JMP Community can help you if you tell us more about the comparison.

Learn it once, use it forever!

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: statistical method which could be used that explains if A is a cause for the B

enclosed is the regression.

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: statistical method which could be used that explains if A is a cause for the B

I think the method you are looking for is called "science', and in science we develop theories that help us anticipate the expected causal relationships. Of course, there can be competing theories, or competing causal relationships, and so we conduct experiments to see what the real-world tells us. And here statistics can help us design the experiments so that they are more efficient in the use of resources, and to ensure that if there is a real relationship we are likely to detect it. As part of our analysis of the experiment data we will typically use regression techniques to build statistical models and these will give us more confidence in the selection of competing theories about causation.

-Dave