More Is Not Always Better: An Experimental Individual-Level Validation of the Randomized Response Technique and the Crosswise Model

Social desirability and the fear of sanctions can deter survey respondents from responding truthfully to sensitive questions. Self-reports on norm breaking behavior such as shoplifting, non-voting, or tax evasion may therefore be subject to considerable misreporting. To mitigate such misreporting, various indirect techniques for asking sensitive questions, such as the randomized response technique (RRT), have been proposed in the literature. In our study, we evaluate the viability of several variants of the RRT, including the recently proposed crosswise-model RRT, by comparing respondents’ self-reports on cheating in dice games to actual cheating behavior, thereby distinguishing between false negatives (underreporting) and false positives (overreporting). The study has been implemented as an online survey on Amazon Mechanical Turk (N = 6,505). Our results indicate that the forced-response RRT and the unrelated-question RRT, as implemented in our survey, fail to reduce the level of misreporting compared to conventional direct questioning. For the crosswise-model RRT, we do observe a reduction of false negatives (that is, an increase in the proportion of cheaters who admit having cheated). At the same time, however, there is an increase in false positives (that is, an increase in non-cheaters who falsely admit having cheated). Overall, our findings suggest that none of the implemented sensitive questions techniques substantially outperforms direct questioning. Furthermore, our study demonstrates the importance of distinguishing false negatives and false positives when evaluating the validity of sensitive question techniques.

Sensitive Questions in Online Surveys: Experimental Results for the Randomized Response Technique (RRT) and the Unmatched Count Technique (UCT)

Gaining valid answers to so-called sensitive questions is an age-old problem in survey research. Various techniques have been developed to guarantee anonymity and minimize the respondent's feelings of jeopardy. Two such techniques are the randomized response technique (RRT) and the unmatched count technique (UCT). In this study we evaluate the effectiveness of different implementations of the RRT (using a forced-response design) in a computer-assisted setting and also compare the use of the RRT to that of the UCT. The techniques are evaluated according to various quality criteria, such as the prevalence estimates they provide, the ease of their use, and respondent trust in the techniques. Our results indicate that the RRTs are problematic with respect to several domains, such as the limited trust they inspire and non-response, and that the RRT estimates are unreliable due to a strong false "no" bias, especially for the more sensitive questions. The UCT, however, performed well compared to the RRTs on all the evaluated measures. The UCT estimates also had more face validity than the RRT estimates. We conclude that the UCT is a promising alternative to RRT in self-administered surveys and that future research should be directed towards evaluating and improving the technique.

RRREG: Stata module to estimate linear probability model for randomized response data

rrreg fits a linear probability model for randomized response data

RRLOGIT: Stata module to estimate logistic regression for randomized response data

rrlogit fits a maximum-likelihood logistic regression for randomized response data.

Sensitive Questions in Online Surveys: Experimental Results for the Randomized Response Technique (RRT) and the Unmatched Count Technique (UCT)

Gaining valid answers to so-called sensitive questions is an age-old problem in survey research. Various techniques have been developed to guarantee anonymity and minimize the respondent's feelings of jeopardy. Two such techniques are the randomized response technique (RRT) and the unmatched count technique (UCT). In this study we evaluate the effectiveness of different implementations of the RRT (using a forced-response design) in a computer-assisted setting and also compare the use of the RRT to that of the UCT. The techniques are evaluated according to various quality criteria, such as the prevalence estimates they provide, the ease of their use, and respondent trust in the techniques. Our results indicate that the RRTs are problematic with respect to several domains, such as the limited trust they inspire and non-response, and that the RRT estimates are unreliable due to a strong false "no" bias, especially for the more sensitive questions. The UCT, however, performed well compared to the RRTs on all the evaluated measures. The UCT estimates also had more face validity than the RRT estimates. We conclude that the UCT is a promising alternative to RRT in self-administered surveys and that future research should be directed towards evaluating and improving the technique.

Likelihood-based estimation of the regression model with scrambled responses

A significant problem in the collection of responses to potentially sensitive questions, such as relating to illegal, immoral or embarrassing activities, is non-sampling error due to refusal to respond or false responses. Eichhorn & Hayre (1983) suggested the use of scrambled responses to reduce this form of bias. This paper considers a linear regression model in which the dependent variable is unobserved but for which the sum or product with a scrambling random variable of known distribution, is known. The performance of two likelihood-based estimators is investigated, namely of a Bayesian estimator achieved through a Markov chain Monte Carlo (MCMC) sampling scheme, and a classical maximum-likelihood estimator. These two estimators and an estimator suggested by Singh, Joarder & King (1996) are compared. Monte Carlo results show that the Bayesian estimator outperforms the classical estimators in almost all cases, and the relative performance of the Bayesian estimator improves as the responses become more scrambled.

Parameter estimation in the presence of auxiliary information

Dissertação para obtenção do Grau de Doutora em Estatística e Gestão de Risco, Especialidade em Estatística

Sensitive Questions in Online Surveys: An Experimental Comparison of the RRT and the Crosswise Model

Self-administered online surveys provide a higher level of privacy protection to respondents than surveys administered by an interviewer. Yet, studies show that asking sensitive questions is problematic also in self-administered mode. Because respondents might not be willing to reveal the truth and provide answers that are subject to social desirability bias, the validity of prevalence estimates of sensitive behaviors gained via online surveys can be challenged. A wellknown method to combat these problems is the Randomized Response Technique (RRT). However, convincing evidence that the RRT provides more valid estimates than direct questioning in online mode is still lacking. Moreover, an alternative approach called the Crosswise Model (CM) has recently been suggested to overcome some of the deficiencies of the RRT. We therefore conducted an experimental study in which different implementations of the RRT and the CM have been tested and compared to direct questioning. Our study is a large-scale online survey on sensitive behaviors by students such as cheating in exams and paper plagiarism. The results of the study reveal poor per-formance of the RRT, while the CM yielded significantly higher estimates of sensitive behaviors than direct questioning. We conclude that the CM is a promising approach for asking sensitive questions in self-administered surveys.

MTurk Survey on "Mood and Personality". Documentation

Social desirability and the fear of negative consequences often deter a considerable share of survey respondents from responding truthfully to sensitive questions. Thus, resulting prevalence estimates are biased. Indirect techniques for surveying sensitive questions such as the Randomized Response Technique are intended to mitigate misreporting by providing complete concealment of individual answers. However, it is far from clear whether these indirect techniques actually produce more valid measurements than standard direct questioning. In order to evaluate the validity of different sensitive question techniques we carried out an online validation experiment at Amazon Mechanical Turk in which respondents' self-reports of norm-breaking behavior (cheating in dice games) were validated against observed behavior. This document describes the design of the validation experiment and provides details on the questionnaire, the different sensitive question technique implementations, the field work, and the resulting dataset. The appendix contains a codebook of the data and facsimiles of the questionnaire pages and other survey materials.

Base rates of counterproductive work behaviors

This study provides a comprehensive assessment of base rates for counterproductive work behaviors (CWB) and examines their relationship with personality and demographic variables. The Randomized-Response Technique (RRT) was employed in order to reduce the effects of social desirability. Base rates were calculated for 66 behaviors for a student and nationwide sample. Results revealed 15 significant behaviors for the student sample and 7 for the nationwide sample. In addition, low neuroticism was found to relate to higher reporting of counterproductive behavior for both groups. Also, low conscientiousness was related to higher reports of CWB in the student sample. Finally, CWB was found to differ based on ethnicity for the student sample such that Caucasians reported higher rates of CWB than Hispanics.

Hemodynamic response in one session of strength exercise with and without electrostimulation in heart failure patients: A randomized controlled trial

Background: Studies have investigated the influence of neuromuscular electrostimulation on the exercise/muscle capacity of patients with heart failure (HF), but the hemodynamic overload has never been investigated. The aim of our study was to evaluate the heart rate (HR), systolic and diastolic blood pressures in one session of strength exercises with and without neuromuscular electrostimulation (quadriceps) in HF patients and in healthy subjects. Methods: Ten (50% male) HF patients and healthy subjects performed three sets of eight repetitions with and without neuromuscular electrostimulation randomly, with one week between sessions. Throughout, electromyography was performed to guarantee the electrostimulation was effective. The hemodynamic variables were measured at rest, again immediately after the end of each set of exercises, and during the recovery period. Results: Systolic and diastolic blood pressures did not change during each set of exercises among either the HF patients or the controls. Without electrostimulation: among the controls, the HR corresponding to the first (85 +/- 13 bpm, p = 0.002), second (84 +/- 10 bpm, p < 0.001), third (89 +/- 17, p < 0.001) sets and recuperation (83 +/- 16 bpm, p = 0.012) were different compared to the resting HR (77 bpm). Moreover, the recuperation was different to the third set (0.018). Among HF patients, the HR corresponding to the first (84 +/- 9 bpm, p = 0.041) and third (84 +/- 10 bpm, p = 0.036) sets were different compared to the resting HR (80 +/- 7 bpm), but this increase of 4 bpm is clinically irrelevant to HF. With electrostimulation: among the controls, the HR corresponding to the third set (84 +/- 9 bpm) was different compared to the resting HR (80 +/- 7 bmp, p = 0.016). Among HF patients, there were no statistical differences between the sets. The procedure was well tolerated and no subjects reported muscle pain after 24 hours. Conclusions: One session of strength exercises with and without neuromuscular electrostimulation does not promote a hemodynamic overload in HF patients. (Cardiol J 2011; 18,1: 39-46)

Safety and Effectiveness of two treatment regimes with tranexamic acid to minimize inflammatory response in elective cardiopulmonary bypass patients: a randomized double-blind, dose-dependent, phase IV clinical trial

Background. In cardiopulmonary bypass (CPB) patients, fibrinolysis may enhance postoperative inflammatory response. We aimed to determine whether an additional postoperative dose of antifibrinolytic tranexamic acid (TA) reduced CPB-mediated inflammatory response (IR). Methods. We performed a randomized, double-blind, dose-dependent, parallel-groups study of elective CPB patients receiving TA. Patients were randomly assigned to either the single-dose group (40 mg/Kg TA before CPB and placebo after CPB) or the double-dose group (40 mg/Kg TA before and after CPB). Results. 160 patients were included, 80 in each group. The incident rate of IR was significantly lower in the double-dose-group TA2 (7.5% vs. 18.8% in the single-dose group TA1; P = 0.030). After adjusting for hypertension, total protamine dose and temperature after CPB, TA2 showed a lower risk of IR compared with TA1 [OR: 0.29 (95% CI: 0.10-0.83), (P = 0.013)]. Relative risk for IR was 2.5 for TA1 (95% CI: 1.02 to 6.12). The double-dose group had significantly lower chest tube bleeding at 24 hours [671 (95% CI 549-793 vs. 826 (95% CI 704-949) mL; P = 0.01 corrected-P significant] and lower D-dimer levels at 24 hours [489 (95% CI 437-540) vs. 621(95% CI: 563-679) ng/mL; P = 0.01 corrected-P significant]. TA2 required lower levels of norepinephrine at 24 h [0.06 (95% CI: 0.03-0.09) vs. 0.20(95 CI: 0.05-0.35) after adjusting for dobutamine [F = 6.6; P = 0.014 corrected-P significant]. We found a significant direct relationship between IL-6 and temperature (rho = 0.26; P < 0.01), D-dimer (rho = 0.24; P < 0.01), norepinephrine (rho = 0.33; P < 0.01), troponin I (rho = 0.37; P < 0.01), Creatine-Kinase (rho = 0.37; P < 0.01), Creatine Kinase-MB (rho = 0.33; P < 0.01) and lactic acid (rho = 0.46; P < 0.01) at ICU arrival. Two patients (1.3%) had seizure, 3 patients (1.9%) had stroke, 14 (8.8%) had acute kidney failure, 7 (4.4%) needed dialysis, 3 (1.9%) suffered myocardial infarction and 9 (5.6%) patients died. We found no significant differences between groups regarding these events. Conclusions. Prolonged inhibition of fibrinolysis, using an additional postoperative dose of tranexamic acid reduces inflammatory response and postoperative bleeding (but not transfusion requirements) in CPB patients. A question which remains unanswered is whether the dose used was ideal in terms of safety, but not in terms of effectiveness.

Can a short period of micronutrient supplementation in older institutionalized people improve response to influenza vaccine? A randomized, controlled trial

OBJECTIVES: To test the hypothesis that a micronutrient supplement can improve seroconversion after influenza immunization in older institutionalized people. DESIGN: Randomized, double-blind, placebo-controlled study. SETTING: Nursing and residential homes in Liverpool, United Kingdom. PARTICIPANTS: One hundred sixty-four residents aged 60 and older from 31 homes were initially randomized; of these, 119 (72.6%) completed the study. INTERVENTION: Participants were randomized to receive a micronutrient supplement providing the reference nutrient intake for all vitamins and trace elements or identical placebo. Tablets were taken over an 8-week period during September and October 2000; influenza vaccine was administered 4 weeks after their commencement. MEASUREMENTS: The hemagglutination-inhibiting antibody response as defined by a fourfold or greater titer rise over 4 weeks and assessed separately for each of the three antigens contained in the 2000/2001 influenza vaccine (A/New Caledonia/20/99 (H1N1), A/Moscow/10/99 (H3N2), B/Beijing/184/93 (B)). RESULTS: Despite a significant increase in serum concentrations of vitamins A, C, D-3, E, folate, and selenium in the supplemented group, there was no significant difference between groups (supplemented vs placebo, respectively) in the proportion of participants seroconverting to H1N1 (41% vs 49%, P=.374), H3N2 (49% vs 58%, P=.343), or B (41% vs 40%, P=.944). CONCLUSION: A micronutrient supplement providing the reference nutrient intake administered over 8 weeks had no beneficial effect on antibody response to influenza vaccine in older people living in long-term care.

Apolipoprotein e genotype has a modest impact on the postprandial plasma response to meals of varying fat composition in healthy men in a randomized controlled trial

BACKGROUND:Apolioprotein E (APOE) genotype is reported to influence a person's fasting lipid profile and potentially the response to dietary fat manipulation. The impact of APOE genotype on the responsiveness to meals of varying fat composition is unknown. OBJECTIVE:We examined the effect of meals containing 50 g of fat rich in saturated fatty acids (SFAs), unsaturated fatty acids (UNSATs), or SFAs with fish oil (SFA-FO) on postprandial lipemia. METHOD:A randomized, controlled, test meal study was performed in men recruited according to the APOE genotype (n = 10 APOE3/3, n = 11 APOE3/E4). RESULTS:For the serum apoE response (meal × genotype interaction P = 0.038), concentrations were on average 8% lower after the UNSAT than the SFA-FO meal in APOE4 carriers (P = 0.015) only. In the genotype groups combined, there was a delay in the time to reach maximum triacylglycerol (TG) concentration (mean ± SEM: 313 ± 25 vs. 266 ± 27 min) and higher maximum nonesterified fatty acid (0.73 ± 0.05 vs. 0.60 ± 0.03 mmol/L) and glucose (7.92 ± 0.22 vs. 7.25 ± 0.22 mmol/L) concentrations after the SFA than the UNSAT meal, respectively (P ≤ 0.05). In the Svedberg flotation rate 60-400 TG-rich lipoprotein fraction, meal × genotype interactions were observed for incremental area under the curve (IAUC) for the TG (P = 0.038) and apoE (P = 0.016) responses with a 58% lower apoE IAUC after the UNSAT than the SFA meal (P = 0.017) in the E4 carriers. CONCLUSIONS:Our data indicate that APOE genotype had a modest impact on the postprandial response to meals of varying fat composition in normolipidemic men. The physiologic importance of greater apoE concentrations after the SFA-rich meals in APOE4 carriers may reflect an impact on TG-rich lipoprotein clearance from the circulation. This trial was registered at clinicaltrials.gov as NCT01522482.