[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

From |
"Matt Spittal" <Matt.Spittal@cancervic.org.au> |

To |
<statalist@hsphsun2.harvard.edu> |

Subject |
RE: st: crude binary analysis |

Date |
Mon, 25 Aug 2008 10:03:28 +1000 |

Sahid, See -help xi- for information on controlling the reference category (specifically, the section 'Summary of controlling the omitted dummy'). -- Matt matt.spittal@cancervic.org.au -----Original Message----- From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu]On Behalf Of Sahid Hocine Sent: Monday, 25 August 2008 5:13 AM To: statalist@hsphsun2.harvard.edu Subject: Re: st: crude binary analysis Steven, Thanks I meant when i use the -tabodds- for binary crude analysis it gives me odds ( not OR my mistake) which i used then to calculate ratios. thus the OR=1 for the reference chosen. When use the xi can i specify the reference for the or analysis ( which will have OR=1)? On Sun, Aug 24, 2008 at 7:54 PM, Steven Samuels <sjhsamuels@earthlink.net> wrote: > Sahid- > > You asked for a version of a survey program which would match the results in > -tabodd-. -tabodd- has no line with a "reference group" and an entry of 1. > Nor does it give odds ratios (OR's) as I stated in my email. So, exactly > what question did you mean to ask? > > -Steve > > > > On Aug 24, 2008, at 2:46 PM, Sahid Hocine wrote: > >> Thanks, >> In addition when i do perform the following i have a problem since >> the or for the first line should be 1 as reference? Sahid >> >> . xi,noomit:svy:logit sd107 i.v130,nocons or >> >> (running logit on estimation sample) >> >> >> >> Survey: Logistic regression >> >> >> >> Number of strata = 35 Number of obs = >> 5600 >> >> Number of PSUs = 4823 Population size = >> 5638.5227 >> >> Design df = >> 4788 >> >> F( 2, 4787) = >> 510.77 >> >> Prob > F = >> 0.0000 >> >> >> >> >> ------------------------------------------------------------------------------ >> >> | Linearized >> >> sd107 | Odds Ratio Std. Err. t P>|t| [95% Conf. >> Interval] >> >> >> -------------+---------------------------------------------------------------- >> >> _Iv130_1 | .016556 .0021707 -31.28 0.000 .0128034 >> .0214084 >> >> _Iv130_2 | .0087124 .0062621 -6.60 0.000 .002129 >> .0356531 >> >> >> ------------------------------------------------------------------------------ >> >> On Sun, Aug 24, 2008 at 7:43 PM, Sahid Hocine <cahid2@googlemail.com> >> wrote: >>> >>> Thanks, >>> In addition when i do perform the following i have a problem since the >>> or for the first line should be 1 as reference? Sahid >>> >>> . xi,noomit:svy:logit sd107 i.v130,nocons or >>> >>> (running logit on estimation sample) >>> >>> >>> >>> Survey: Logistic regression >>> >>> >>> >>> Number of strata = 35 Number of obs = >>> 5600 >>> >>> Number of PSUs = 4823 Population size = >>> 5638.5227 >>> >>> Design df = >>> 4788 >>> >>> F( 2, 4787) = >>> 510.77 >>> >>> Prob > F = >>> 0.0000 >>> >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> >>> | Linearized >>> >>> sd107 | Odds Ratio Std. Err. t P>|t| [95% Conf. >>> Interval] >>> >>> >>> -------------+---------------------------------------------------------------- >>> >>> _Iv130_1 | .016556 .0021707 -31.28 0.000 .0128034 >>> .0214084 >>> >>> _Iv130_2 | .0087124 .0062621 -6.60 0.000 .002129 >>> .0356531 >>> >>> >>> ------------------------------------------------------------------------------ >>> >>> On Sun, Aug 24, 2008 at 4:00 PM, Steven Samuels >>> <sjhsamuels@earthlink.net> wrote: >>>> >>>> Sahid, -taboods- gives crude odds, not odds ratios. >>>> Use -svy: logit- with -xi, noomit-; drop the constant, and add option >>>> "or". >>>> >>>> Note: when referring to commands on Statalist, use hyphens around their >>>> names: -tabodds- >>>> >>>> -Steve >>>> >>>> /**************************************************/ >>>> sysuse auto, clear >>>> gen rep= rep78<=3 >>>> tab rep >>>> svyset _n >>>> >>>> tabodds rep foreign >>>> /* Now svy: logit */ >>>> xi, noomit : svy: logit rep i.foreign, nocons or >>>> /*************************************************/ >>>> >>>> On Aug 24, 2008, at 10:29 AM, Sahid Hocine wrote: >>>> >>>>> Dear Stalister, >>>>> >>>>> I have a question regarding Stata 10 being a newbee on the field. >>>>> >>>>> I am analyzing a DHS data set *(survey) and for the bivariate crude >>>>> analysis I am used to tabodds however it is not supported by the >>>>> svy , so I guess I would have to use the logistic command for binary >>>>> analysis but the output is not what I want i.e a crude odd ratio by >>>>> category. Can I proceed with the svy or the stratification is the only >>>>> solution? >>>>> >>>>> Waiting for your help >>>>> >>>>> Sahid >>>>> >>>>> >>>>> >>>>> >>>>> . svy:logistic sd107 v130 >>>>> (running logistic on estimation sample) >>>>> >>>>> Survey: Logistic regression >>>>> >>>>> Number of strata = 35 Number of obs = >>>>> 5600 >>>>> Number of PSUs = 4823 Population size = >>>>> 5638.5227 >>>>> Design df = >>>>> 4788 >>>>> F( 1, 4788) = >>>>> 0.77 >>>>> Prob > F = >>>>> 0.3795 >>>>> >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> | Linearized >>>>> sd107 | Odds Ratio Std. Err. t P>|t| [95% Conf. >>>>> Interval] >>>>> >>>>> -------------+---------------------------------------------------------------- >>>>> v130 | .5262408 .3844291 -0.88 0.380 .1256632 >>>>> 2.203743 >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> >>>>> . svy:tabodds sd107 v130 >>>>> tabodds is not supported by svy with vce(linearized); see help svy >>>>> estimation for a list of Stata estimation commands >>>>> that are supported by svy >>>>> r(322); >>>>> >>>>> . :tabodds sd107 v130 >>>>> unrecognized command: : invalid command name >>>>> r(199); >>>>> >>>>> . tabodds sd107 v130 >>>>> >>>>> >>>>> -------------------------------------------------------------------------- >>>>> v130 | cases controls odds [95% Conf. >>>>> Interval] >>>>> >>>>> ------------+------------------------------------------------------------- >>>>> muslim | 81 5218 0.01552 0.01246 0.01933 >>>>> christian | 3 298 0.01007 0.00323 >>>>> 0.03139 >>>>> >>>>> -------------------------------------------------------------------------- >>>>> Test of homogeneity (equal odds): chi2(1) = 0.55 >>>>> Pr>chi2 = 0.4602 >>>>> >>>>> Score test for trend of odds: chi2(1) = 0.55 >>>>> Pr>chi2 = 0.4602 >>>>> * >>>>> * For searches and help try: >>>>> * http://www.stata.com/help.cgi?search >>>>> * http://www.stata.com/support/statalist/faq >>>>> * http://www.ats.ucla.edu/stat/stata/ >>>> >>>> * >>>> * For searches and help try: >>>> * http://www.stata.com/help.cgi?search >>>> * http://www.stata.com/support/statalist/faq >>>> * http://www.ats.ucla.edu/stat/stata/ >>> >> * >> * For searches and help try: >> * http://www.stata.com/help.cgi?search >> * http://www.stata.com/support/statalist/faq >> * http://www.ats.ucla.edu/stat/stata/ > > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

**References**:**Re: st: crude binary analysis***From:*"Sahid Hocine" <cahid2@googlemail.com>

- Prev by Date:
**RE: st: probit vs. reg postestimation** - Next by Date:
**RE: st: how to get p-value robust to heteroskedasticity and ajusted for the clustering of observations in firms** - Previous by thread:
**Re: st: crude binary analysis** - Next by thread:
**st: converting dates from Excel .csv file** - Index(es):

© Copyright 1996–2014 StataCorp LP | Terms of use | Privacy | Contact us | What's new | Site index |