help rrlogitalso see: rrlogit postestimation -------------------------------------------------------------------------------

Title

rrlogit-- Logistic regression for randomized response data

Syntax

rrlogitdepvar[indepvars] [if] [in] [weight] [,options]

optionsDescription ------------------------------------------------------------------------- Modelpwarner(#|varname)probability of non-negated question in Warner's model; default ispwarner(1)pyes(#|varname)probability of a surrogate "yes"; default ispyes(0)pno(#|varname)probability of a surrogate "no"; default ispno(0)noconstantsuppress constant termoffset(varname)includevarnamein model with coefficient constrained to 1asisretain perfect predictor variablesconstraints(constraints)apply specified linear constraintsSE/Robust

vce(vcetype)vcetypemay beoim,robust,opg,bootstrap, orjackkniferobustsynonym forvce(robust)cluster(varname)adjust standard errors for intragroup correlationReporting

level(#)set confidence level; default islevel(95)orreport odds ratiosMax options

maximize_optionscontrol the maximization process; seldom used -------------------------------------------------------------------------depvarandindepvarsmay contain time-series operators; see tsvarlist.bootstrap,by,jackknife,rolling,statsby,stepwise, andxiare allowed; see prefix.fweights,iweights, andpweightsare allowed; see weight. See rrlogit postestimation for features available after estimation.

Description

rrlogitfits a maximum-likelihood logistic regression for randomized response data following (see, e.g., Maddala 1983:54-56).depvar=0 indicates a negative outcome (a "no" answer);depvar!=0 &depvar<. (typicallydepvar=1) indicates a positive outcome (a "yes" answer).

rrlogitis suited for the analysis of data from Warner's RRT scheme, the forced-response RRT, or the unrelated-question RRT with a known distribution for the non-sensitive question (see, e.g., Fox and Tracy 1986).

Options+-------+ ----+ Model +------------------------------------------------------------

pwarner(#|varname)specifies the probability of the non-negated question in Warner's RRT scheme.#must be in [0,1] and may not be 0.5. The default ispwarner(1). Individually varying probabilities may be specified bypwarner(varname), wherevarnameholds the individual probabilities.

pwarner()is conditional on not being directed to the surrogate "yes" or "no". That is, the overall probability of the non-negated question is (1-pyes-pno)*pwarner.

pyes(#|varname)specifies the probability of a surrogate "yes" answer.#must be in [0,1]. The default ispyes(0). Individually varying probabilities may be specified bypyes(varname), wherevarnameholds the individual probabilities.

pno(#|varname)specifies the probability of a surrogate "no" answer.#must be in [0,1]. The default ispno(0). Individually varying probabilities may be specified bypno(varname), wherevarnameholds the individual probabilities.

pyes()andpno()are unconditional (overall) probabilities. For example, in an unrelated-question RRT where the probability to be directed to the non-sensitive question is 0.4 (i.e. the probability to answer the sensitive question is 60%) and the probability to answer "yes" to the non-sensitive question is known to be, say, 0.75, the overall probability of a surrogate "yes" is 0.4*0.75 = 0.3. Likewise, the overall probability of a surrogate "no" is 0.4*(1-0.75) = 0.1.

noconstant,offset(varname); see estimation options.

asisforces retention of perfect predictor variables and their associated perfectly predicted observations and may produce instabilities in maximization; seeprobit.

constraints(constraints); see estimation options.+-----------+ ----+ SE/Robust +--------------------------------------------------------

vce(vcetype); seevce_option.

robust,cluster(varname); see estimation options.+-----------+ ----+ Reporting +--------------------------------------------------------

level(#); see estimation options.

orreports the estimated coefficients transformed to odds ratios, i.e., exp(b) rather than b. Standard errors and confidence intervals are similarly transformed. This option affects how results are displayed, not how they are estimated.ormay be specified at estimation or when replaying previously estimated results.+-------------+ ----+ Max options +------------------------------------------------------

maximize_options:difficult,technique(algorithm_spec),iterate(#), []nolog,trace,hessian,gradient,showstep,tolerance(#),ltolerance(#),gtolerance(#),nrtolerance(#),nonrtolerance,from(init_specs); see maximize. These options are seldom used.

Examples. rrlogit infidelity male age, pyes(0.5)

ReferencesFox, James Alan, and Paul E. Tracy. 1986. Randomized response: A method for sensitive surveys. London: Sage.

Maddala, G. S. 1983. Limited Dependent and Qualitative Variables in Econometrics. Cambridge: Cambridge University Press.

AuthorBen Jann, Institute of Sociology, University of Bern, jann@soz.unibe.ch

Also seeOnline: rrlogit postestimation;

constraint,logistic,logit,rrreg(if installed)