confidence intervals for relative risk ratios

[DrBob]

i must have gone through this about a dozen times already, but it goes in one ear and out the other. HK, are you familiar with standard deviations and inferential statistics? i understand how to get specific SD of the mean(eg. what % are 2SD below a given mean). i also dont have any trouble calculating odds ratios or relative risks for observational studies. what i dont understand is one small detail, which is...

'if the given confidence interval(CI) for a relative risk(RR) conatins "1.0", then there is no statistically significant effect of exposure.'
for example:

RR=1.57
95%CI=(1.12-2.25)
interpretation=statistically significant(increased risk)

RR=1.65
95%CI=(0.89-2.34)
interpretation=Not statistically significant(risk is the same)

RR=0.76
95%CI=(0.56-0.93)
interpretation=Statistically significant(decreased risk)

huh? i dont get what they mean by "if the CI contains a 1.0...


I, ah...I'm not much of a "word" person...

[Alice S. Fuzzybutt]

I think I just broke a blood vessel in my brain by reading that.

<IMG SRC="http://atamichimpo.50megs.com/images/fuzzyadfsig.gif">

I'm a big wheel down at the cracker factory.

[IrishAlkey]

Fuck!

That was your last one!

:)

<marquee><font color=red size=4>I'M KIDDING!</font></marquee>

[blakjeezis]

This literally made my testicles sore.

<IMG SRC =http://blakjeezis.homestead.com/files/buckwheat.gif>
"What are the scores George Dawes?"
<marquee>White people are so scared of blakjeezis</marquee>

[DarkHippie]

can't you use a graphing calculator?

<IMG SRC=http://thereisnogod.faithweb.com/images/darkhippie2.gif>
<marquee>Who is DarkHippie? "You look like an Amish child molester"-- Jim Norton. "Watch out for this one. Someday he's gonna snap and kill you all."-- Rich Vos </marquee>
<i>LABELS ARE FOR PRODUCTS, NOT PEOPLE! DON'T HUG A TREE, PLANT ONE!
</i><a href=http://www.freeopendiary.com/entrylist.asp?authorcode=A537085>Gonads and Strife: a journal</a>

[TheMojoPin]

<img src=http://www.simpsonslibrary.com/martin3.gif>

<img src=http://thereisnogod.faithweb.com/images/themojopin2.gif>
VP #2 for the Coalition of Angry Micks, and Minister of Bloody Mayhem.
"You can tell some lies about the good times you've had/But I've kissed your mother twice and now I'm working on your dad..."

[ADF]

I'm pretty confident that I don't know what the fuck you're talking about.

<center><img src=http://thereisnogod.faithweb.com/images/adf5a.gif>
Ce sachet d'emballage n'est pas un jouet.</center>

This message was edited by The Al Dukes Fan on 12-6-02 @ 7:43 PM

[HordeKing1]

I can do basic statistical formulas by hand. Standard deviation is no problem. However, I rely on computer programs for every other statistics or math related task. I'm sure I have the information somewhere (I've done these before) but I don't recall it offhand.

There are people who make a very good living running the numbers from a research study. The interpretation is still up to the researcher of course.



<img src="http://members.aol.com/rnfpantera/hking1">

[RonMwangaguhunga]

Dr. Bob will be found on Monday with various body parts of underpriveledged young men in his refrigirator.

[silera]

I hate feeling this stupid.


<center>http://groups.yahoo.com/group/silera/files/Silera/sig4.gif

<font size=3><font color=red>I can't stand myself either.</font></font></center>
<font color=white>

[TheMojoPin]

I hate feeling this stupid.

Me too. But then I go have some more drug-addled sex and beat up somebody dressed like a wizard and I feel grand. SAN DIMAS HIGH SCHOOL FOOTBALL RULES.

<img src=http://home.ix.netcom.com/~camman/_uimages/mojopin3.gif>
VP #2 for the Coalition of Angry Micks, and Minister of Bloody Mayhem.
"You can tell some lies about the good times you've had/But I've kissed your mother twice and now I'm working on your dad..."

[silera]

I'm scared that I understood everything you just said.



<center>http://groups.yahoo.com/group/silera/files/Silera/sig4.gif

<font size=3><font color=red>I can't stand myself either.</font></font></center>
<font color=white>

[TheMojoPin]

If you wanna go and take a ride wit me
wit three women in the fo' with the gold D's
Oh why do I live this way? (Hey, must be the money!)

If you wanna go and get high wit me
Smoke a L in the back of the Benz-y
Oh why must I feel this way? (Hey, must be the money!)

Yeah, I suck, I know.

<img src=http://home.ix.netcom.com/~camman/_uimages/mojopin3.gif>
VP #2 for the Coalition of Angry Micks, and Minister of Bloody Mayhem.
"You can tell some lies about the good times you've had/But I've kissed your mother twice and now I'm working on your dad..."

[TooCute]

You need to read what the
definition of a confidence
interval is. A CI is the range
of numbers for which your
true mean will fall between
95% (in this case - your
example) of the time that
you sample.

Since a RR value of 1.0
means that you have
neither a positive nor a
negative effect, if your
confidence intervals for that
RR span 1.0, you can't be
sure that you have a
significant effect at all.

e.g. let's say you calculate a
95% CI of .95 - 1.05, and
you calculate a RR of 1.04
(an increased risk), you
have no way of knowing if
you have a significant effect
or not - you might run your
experiment again and come
up with a RR of .96 (a
decreased risk).

However, same experiment,
same RR of 1.04, but you
have a 95% CI of 1.01-1.05,
you can be 95% sure that if
you run the experiment
again, you will come up with
a positive effect again.

I hope that helps.


<img src=http://thereisnogod.faithweb.com/images/toocute2.gif>
ADF makes great sigpics!

[silera]

That was just mean.

I go cry now.



<center>http://groups.yahoo.com/group/silera/files/Silera/sig4.gif

<font size=3><font color=red>I can't stand myself either.</font></font></center>
<font color=white>

[Michael Fury]

Okay, enough of the talent show already. Let's move to the Swimsuit and Evening Gown segments. ..






Your bark was loud, but your bite wasn't vicious
And them rhymes you were kickin were quite bootylicious

[The Chairman]

As Bill Clinton would say, "it depends upon what [your]definition of a confidence interval is is".

95%?

Given the RR span of 1.0, a
significant effect isn't ensured.

You need to re-run your experiment.

<img src = http://thereisnogod.faithweb.com/images/kaga.gif>
thanks ADF for the SigPic!!
I'll take that Chesterfield now...

[Dewey]

Too Cute,

It really turns me on when you talk dirty like that. Too bad I don't have enough of a Confidence Interval (CI) to ask you out.

<IMG SRC="http://www.agw-werbeartikel.de/images/easy-rider.jpg"><br>"Still searching for America."

[The Chairman]

Too bad I don't have enough of a Confidence Interval (CI) to ask you out.


he he

Is it because you lack self confidence or you lack the balls to ask her out since she's taken by the Chairman? :-)

<img src = http://thereisnogod.faithweb.com/images/kaga.gif>
thanks ADF for the SigPic!!
I'll take that Chesterfield now...

[HordeKing1]

I heard that she has a magic aura surrounding her that can change an undesired suitor into a troll.

Or perhaps it's advanced technology. I believe it was Clark who said that "technology, sufficiently advanced from our own is indistinguishable from magic."

<img src="http://members.aol.com/rnfpantera/hking1">