LACTNET Archives

Lactation Information and Discussion

LACTNET@COMMUNITY.LSOFT.COM

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Susan Burger <[log in to unmask]>
Reply To:
Lactation Information and Discussion <[log in to unmask]>
Date:
Sat, 20 Nov 2010 09:09:05 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (21 lines)
I realize that double negatives are really confusing.  And yet, I could not bring myself to write it the other way.  As I said, I don't know what was in the PROBIT study, just that the results have been misused to say breastfeeding is not as beneficial as we thought.

When I worded my hypothetical situation -- I wrote it the way that I think studies should proceed.  NOT breastfeeding is the risk factor.  Any study on breastfeeding should have their null hypothesis to be to test that NOT breastfeeding (which is mostly FORMULA feeding or premature introduction of complementary foods) does NOT have higher risks.  When you create a null hypothesis, you are stating the opposite of what you believe will happen and then statistically determining whether or not a result that is counter to the null hypothesis is due to chance.

In the future I think I will simply state formula feeding and/or premature introduction of complementary foods so as to avoid the double negatives.

Most of what I wrote was hypothetical to illustrate how one should think about studies that don't show strong results when you move from clinical to intervention to large-scale intervention trials.  An very important point as you move to larger scale intervention trials is that you unit of consideration is often no longer the individual. If you randomly assign the treatment to practitioners - say one practitioner suggests frenotomy, another practitioner suggests frenotomy only, another suggests CST and frenotomy, and a fourth suggests muddling through with no intervention you have created what is known as a nested design.  The sample size is not merely based on the number of people seen by those practitioners but the number of practitioners in each treatment group.  If you only have one practitioner in each group -- your sample size for treatment is really only one.  Ditto for clinics or hospitals or larger units such as states that implement policies versus states that do not implement policies.

Best, Susan

             ***********************************************

Archives: http://community.lsoft.com/archives/LACTNET.html
To reach list owners: [log in to unmask]
Mail all list management commands to: [log in to unmask]
COMMANDS:
1. To temporarily stop your subscription write in the body of an email: set lactnet nomail
2. To start it again: set lactnet mail
3. To unsubscribe: unsubscribe lactnet
4. To get a comprehensive list of rules and directions: get lactnet welcome

ATOM RSS1 RSS2