Democratic Underground Latest Greatest Lobby Journals Search Options Help Login
Google

Potential Confounders in medical research.

Printer-friendly format Printer-friendly format
Printer-friendly format Email this thread to a friend
Printer-friendly format Bookmark this thread
This topic is archived.
Home » Discuss » Topic Forums » Health Donate to DU
 
Fire_Medic_Dave Donating Member (1000+ posts) Send PM | Profile | Ignore Wed Sep-03-08 12:32 PM
Original message
Potential Confounders in medical research.
Ok I know what a confounding variable is. My question is what role does a "potential" confounder play in a statistical model? It seems like an ambiguous term. Just wondered if any of you statisticians might be able to clear it up.

David
Printer Friendly | Permalink |  | Top
chicagomd Donating Member (437 posts) Send PM | Profile | Ignore Wed Sep-03-08 02:15 PM
Response to Original message
1. My understanding
is that when authors point out "potential" confounders it refers to something that may change the data set before statistical analysis but is unable to be, or was by study design not, controlled for. Essentially they are pointing out variables that need to be considered when you are looking at the statistical outcome, as they are not considered within the model used.

For example, a potential confounder in a research phone poll on who people are going to vote for in November is prejudice against African Americans. People might say on the phone that race is not a variable, but it potentially is not measurable because you cannot control for people's level of dishonesty.

It is also a nice way of pointing out the limitations of your own study and beating the critics to the punch.
Printer Friendly | Permalink |  | Top
 
Fire_Medic_Dave Donating Member (1000+ posts) Send PM | Profile | Ignore Wed Sep-03-08 04:23 PM
Response to Reply #1
2. Thanks for the explanation.
Seems to invalidate or significantly weaken conclusions made from studies where confounders are not factored into the model. Especially if the confounders are statistically significant.

David
Printer Friendly | Permalink |  | Top
 
WildClarySage Donating Member (1000+ posts) Send PM | Profile | Ignore Thu Sep-04-08 10:51 AM
Response to Original message
3. But maybe this needs to be posted a few more times
before everyone is convinced. I mean, repeated posts on the same topic *must* be more truthy.
Printer Friendly | Permalink |  | Top
 
Fire_Medic_Dave Donating Member (1000+ posts) Send PM | Profile | Ignore Thu Sep-04-08 11:55 AM
Response to Reply #3
4. Sorry I've never seen this posted, I was just trying to gain some knowledge.
Printer Friendly | Permalink |  | Top
 
WildClarySage Donating Member (1000+ posts) Send PM | Profile | Ignore Thu Sep-04-08 06:21 PM
Response to Reply #4
8. I get that. I was referring to the numerous
repeated threads started by the same person saying exactly the same thing in each thread.
Printer Friendly | Permalink |  | Top
 
Fire_Medic_Dave Donating Member (1000+ posts) Send PM | Profile | Ignore Thu Sep-04-08 08:46 PM
Response to Reply #8
9. No worries.
Didn't know where you were going with it.

David
Printer Friendly | Permalink |  | Top
 
cosmik debris Donating Member (1000+ posts) Send PM | Profile | Ignore Thu Sep-04-08 12:18 PM
Response to Original message
5. I believe that a "potential confounder"
could be used to introduce other sources of error.

If you are trying to correct for a "potential confounder" you have to assign a numerical value for the "potential" of occurrence. And you also have to have some numerical value for the impact of the "confounder".

Any variability in these two values will perturb your statistical model.

But that is based on my limited understanding of the term.
Printer Friendly | Permalink |  | Top
 
Fire_Medic_Dave Donating Member (1000+ posts) Send PM | Profile | Ignore Thu Sep-04-08 04:25 PM
Response to Reply #5
6. That would make sense.
Thanks.

David
Printer Friendly | Permalink |  | Top
 
xchrom Donating Member (1000+ posts) Send PM | Profile | Ignore Thu Sep-04-08 04:34 PM
Response to Original message
7. kick
Printer Friendly | Permalink |  | Top
 
DU AdBot (1000+ posts) Click to send private message to this author Click to view 
this author's profile Click to add 
this author to your buddy list Click to add 
this author to your Ignore list Fri Aug 29th 2014, 04:03 PM
Response to Original message
Advertisements [?]
 Top

Home » Discuss » Topic Forums » Health Donate to DU

Powered by DCForum+ Version 1.1 Copyright 1997-2002 DCScripts.com
Software has been extensively modified by the DU administrators


Important Notices: By participating on this discussion board, visitors agree to abide by the rules outlined on our Rules page. Messages posted on the Democratic Underground Discussion Forums are the opinions of the individuals who post them, and do not necessarily represent the opinions of Democratic Underground, LLC.

Home  |  Discussion Forums  |  Journals |  Store  |  Donate

About DU  |  Contact Us  |  Privacy Policy

Got a message for Democratic Underground? Click here to send us a message.

© 2001 - 2011 Democratic Underground, LLC