Search

ABSTRACT Recreational fishing surveys have an important role in providing data to inform fisheries management. The selection of a contact method is an important and often challenging consideration that influences the potential for non‐sampling errors that can result in unrepresentative data and biased estimates. Telephones are used for many off‐site recreational fishing surveys, but widespread use of digital technology suggests that online surveys could be a viable alternative. We compared phone surveys with alternative online surveys using probability sampling from a licence database. Response rates were lower for online surveys (13%–26%) than phone surveys (97%–98%). The proportion of respondents who reported marine fishing in the previous 12 months did not differ between survey methods (98% for both survey modes), but the proportion of avid fishers was higher for online surveys (41%–54%) than phone surveys (37%–40%). These differences reflect higher respondent burden for online surveys where data are self‐administered and respondents are more likely to opt‐out. Higher non‐response and over‐representation of avid fishers from online surveys may compromise data reliability. We recommend comparative studies to quantify and correct for biases when evaluating alternative survey modes for new or ongoing surveys. [ABSTRACT FROM AUTHOR]

In order to improve upon the efficiency of an estimate in double sampling for estimating population mean of character under study using an auxiliary variable, a part of survey resources are used to collect the information on auxiliary variable. Some authors have suggested exponential-type estimators and some others advocated for log-type estimators. But combination of such is required for specific situation. This paper presents a class of logarithmic-cum-exponential ratio estimators in double sampling setup. The expressions for the mean squared error and bias of the proposed class of estimators are derived for two different cases(sub-sample and independent sample). Sometimes the persons involved in the sample survey have to undergo for risk on life. For example, data collection in naxalites area, working in intense forest, interview during spread of epidemic or data collection in politically disturbed region. Such risk may affect the accuracy, efficiency of estimation. A linear Risk function is used for the proposed class of estimators. Two cases of double sampling are compared in terms of relative efficiency in view to risk aspect. It is found that the proposed class of estimators has a lower mean squared error than the simple mean estimator, usual ratio, usual exponential, usual log estimators in the double sampling setup. In addition, these theoretical results are supported by a numerical example. Risk function based simulated study is performed for the support of findings of the content. Optimal sample sizes under risk are derived and compared under two cases. [ABSTRACT FROM AUTHOR]

Traditional methods for collecting robust samples of recreational effort and catch data from offsite mail or telephone surveys are becoming increasingly difficult to obtain. Smartphone applications (apps) that allow recreational fishers to report their effort and catch are a potentially viable data collection tool, but may be biased. We evaluated demographics, effort, and catch of freshwater anglers in south‐west Western Australia by comparing independent data from concurrent phone‐recall and app‐diary surveys for a 12‐month period in 2017–2018. Survey modes differed, with higher proportions of app respondents angling in freshwater and being avid anglers. For both survey modes, unweighted mean effort (days fished) was higher in dams than rivers, but was higher from the phone‐recall survey for dams and rivers. The unweighted distribution of kept and released catches (number of fish per angler) of rainbow trout (Oncorhynchus mykiss) and redfin (Perca fluviatilis) were similar between survey modes, with higher retention of redfin and higher release of rainbow trout. Kept and released rates of freshwater cobbler (Tandanus bostocki) and brown trout (Salmo trutta) were less common in both survey modes. Lower response rates and higher survey participation by avid anglers (≥20 days per year) from the app‐diary survey may bias survey data. Such bias needs to be addressed when using digital data collection methods, along with improving recruitment and retention of app participants. [ABSTRACT FROM AUTHOR]

Classical estimators fail to be efficient in practical scenarios when data is riddled with extreme values known as outliers. Robust estimation strategies are insensitive to outliers and may be used in such cases. The current work is focused on developing a novel robust estimation strategy using Huber M-estimation. A new chain-product type estimator for population mean has been suggested utilizing data on two auxiliary variables. A numerical comparison has been carried out between the proposed robust estimator and the corresponding classical estimator using real and simulated data containing outliers. Recommendations have been made for its practical use based on the encouraging results. [ABSTRACT FROM AUTHOR]

In this article, two non-randomized scrambling models have been presented for estimation of mean of a quantitative sensitive attribute. Information on two unrelated non-sensitive qualitative attributes have been utilized for the purpose. An extensive simulation study has been conducted to compare the performance of the proposed models with respect to a contemporary model in terms of both percentage relative efficiency and privacy protection for various values of the parameters. The encouraging results suggest that the proposed models may be useful for survey statisticians in their field work. [ABSTRACT FROM AUTHOR]

Our work was motivated by the question whether, and to what extent, well‐established risk factors mediate the racial disparity observed for colorectal cancer (CRC) incidence in the United States. Mediation analysis examines the relationships between an exposure, a mediator and an outcome. All available methods require access to a single complete data set with these three variables. However, because population‐based studies usually include few non‐White participants, these approaches have limited utility in answering our motivating question. Recently, we developed novel methods to integrate several data sets with incomplete information for mediation analysis. These methods have two limitations: (i) they only consider a single mediator and (ii) they require a data set containing individual‐level data on the mediator and exposure (and possibly confounders) obtained by independent and identically distributed sampling from the target population. Here, we propose a new method for mediation analysis with several different data sets that accommodates complex survey and registry data, and allows for multiple mediators. The proposed approach yields unbiased causal effects estimates and confidence intervals with nominal coverage in simulations. We apply our method to data from U.S. cancer registries, a U.S.‐population‐representative survey and summary level odds‐ratio estimates, to rigorously evaluate what proportion of the difference in CRC risk between non‐Hispanic Whites and Blacks is mediated by three potentially modifiable risk factors (CRC screening history, body mass index, and regular aspirin use). [ABSTRACT FROM AUTHOR]

This article presents a new method that combines item response theory techniques with small-area estimation approaches to handle missing data. We propose an unbiased estimator for the average skill parameter of three-parameter logistic models using plausible values as the imputation method for missing data. We conduct a thorough simulation study to compare our estimator with the Horvitz-Thompson estimator in complex sampling. Synthetic data experiments demonstrate that our proposal has lower standard errors than its competitor. Additionally, we apply our method to the results in mathematics of the 2015 Program for International Student Assessment and compare our findings with previous studies. These findings indicate that our method is a competitive alternative for generating accurate official statistics. [ABSTRACT FROM AUTHOR]

Objective: The aim of this paper is to present bootstrap algorithms for measuring the accuracy of estimation and prediction in design-based and model-based approaches in survey sampling and small area estimation. Three proposals of prediction-mean squared error estimators are also examined. Research Design & Methods: Various bootstrap procedures are shown and used to estimate the design- and prediction-mean squared errors based on real data. Computations are supported by two R packages. Findings: Three prediction-mean squared error estimators are proposed. Implications / Recommendations: The bootstrap algorithms used in the design-based approach give similar results for the considered data for the variance estimates of the considered estimator, implying that the speed of the algorithms may be important for practitioners in cases of similar properties. The proposed estimators of the prediction mean squared error produce higher estimates than other estimators in the model-based approach, indicating a positive bias that can be interpreted as a pessimistic accuracy estimate. Contribution: All the presented bootstrap algorithms are easily applicable using two R packages available at R CRAN and GitHub. Three double bootstrap prediction-MSE estimators are proposed and analysed in the real-data application.

Background:The Together for Health-Virginia (T4H-VA) Research Program aimed to advance cancer prevention, education, and outreach in Virginia. Creating a representative and inclusive cohort is critical to the program’s mission and quality of outcomes. The T4H-VA Research Program utilized a multi-modal sampling approach to improve population health assessment. The current study describes the technology-based, non-probability platform developed for this purpose and compares differences between the probability-based (mail-based) and non-probability-based (e-cohort) methods with respect to participant demographics, health characteristics, and health information and technology use.Methods:T4H-VA is a research registry focusing on 54 counties within the Massey Comprehensive Cancer Center (MCCC) catchment area in Richmond, VA. Adult residents proficient in English were eligible. For the probability-based sampling, surveys were mailed to residents within the catchment area. For the non-probability sampling, an online study platform was developed and surveys were completed through the web/mobile app.Result:Both cohorts fell short of recruitment goals. The study yielded 1158 participants (M = 57, SD = 16 years; 55.0% female; 72.1% White); 899 (77.6%) were sampled through the probability, mail-based approach. Participants who identified as ‘other' race were significantly less likely to be sampled by the non-probability method. Significant differences emerged, including health protective (greater moderate and high physical activity) and risk factors (greater alcohol consumption and personal history of cancer) in the non-probability, e-cohort relative to the probability sample. E-Cohort participants were significantly more likely to report using electronic health records.Discussion:Overall difficulties in recruiting were caused, at least in part, by the onset of the COVID-19 pandemic and related factors. The e-cohort, which used exclusively digital recruitment strategies, fell significantly short of recruitment goals. This suggests in-person and mail-based strategies remain important for recruitment. Moreover, instead of favoring a singular approach, a combined approach to survey sampling may capitalize on the strengths of each sampling mode to increase diversity in sociodemographic and health risk characteristics.Abbreviations:ABS: address-based sample; BRFSS: Behavioral Risk Factor Surveillance System; BMI: body mass index; CI: confidence interval; COVID-19: coronavirus disease 2019; e-cohort: electronic cohort; EHR: electronic health record; eConsent: electronic consent; HINTS: Health Information National Trends Survey; HIT: health information technology; HOP: Healthy Oregon Project; MCC: Massey Cancer Center; NCI: National Cancer Institute; NIH: National Institute of Health; OR: odds ratio; REDCap: Research electronic data capture; T4H-VA: Together for Health-Virginia; VCU: Virginia Commonwealth University.

Estimates of recreational fishing effort and catch are needed to inform fisheries management, but such estimates can be challenging to obtain. We combined a zig‐zag line transect survey of buoys, a fisheries enforcement survey and a drone survey to estimate recreational effort in the open‐access trap fishery for signal crayfish (Pacifastacus leniusculus) in the sixth largest lake in Europe, Lake Vättern, Sweden. Using supporting variables such as size of nearby ports and bathymetry, our model‐assisted estimate of total effort was 6400 (±4300 CI) buoys during the three‐weekend season. Using the frequency of buoys to traps as a constant, the total number of crayfish traps was estimated to be 25,500 (±17,000 CI). The sampling design and methods presented are useful for ecologists and resource managers to design future multi‐source surveys of recreational trap fishing effort. [ABSTRACT FROM AUTHOR]

Objective: The aim of this paper is to present bootstrap algorithms for measuring the accuracy of estimation and prediction in design-based and model-based approaches in survey sampling and small area estimation. Three proposals of prediction-mean squared error estimators are also examined. Research Design & Methods: Various bootstrap procedures are shown and used to estimate the design- and prediction-mean squared errors based on real data. Computations are supported by two R packages. Findings: Three prediction-mean squared error estimators are proposed. Implications / Recommendations: The bootstrap algorithms used in the design-based approach give similar results for the considered data for the variance estimates of the considered estimator, implying that the speed of the algorithms may be important for practitioners in cases of similar properties. The proposed estimators of the prediction mean squared error produce higher estimates than other estimators in the model-based approach, indicating a positive bias that can be interpreted as a pessimistic accuracy estimate. Contribution: All the presented bootstrap algorithms are easily applicable using two R packages available at R CRAN and GitHub. Three double bootstrap prediction-MSE estimators are proposed and analysed in the real-data application. Article type: original article. [ABSTRACT FROM AUTHOR]

Survey data are often randomly drawn from an underlying population of inferential interest under a multistage, complex sampling design. A sampling weight proportional to the number of individuals in the population that each sampled individual represents is released. The sampling design is informative with respect to a response variable of interest if the variable correlates with the sampling weights. The distribution for the variables of interest differs in the sample and in the population, requiring correction to the sample distribution to approximate the population. We focus on model-based Bayesian inference for repeated (continuous) measures associated with each sampled individual. We devise a model for the joint estimation of response variable(s) of interest and sampling weights to account for the informative sampling design in a formulation that captures the association of the measures taken on the same individual incorporating individual-specific random-effects. We show that our approach yields correct population inference on the observed sample of units and compare its performance with competing method via simulation. Methods are compared using bias, mean square error, coverage, and length of credible intervals. We demonstrate our approach using a National Health and Nutrition Examination Survey dietary dataset modeling daily protein consumption. [ABSTRACT FROM AUTHOR]

The endogeneity problem arises when the auxiliary variables correlate to the error terms. In such cases, appropriate instrumental variables ensure efficient estimation. Calibration has recognized itself as an important methodological tool at a large scale to estimate the population total in survey sampling. Which does not offer efficient estimation in the presence of endogeneity. When endogeneity is present in the auxiliary variables, the calibration using endogenous auxiliary variables may produce biasedness and increase variance due to inappropriate model assumptions. In this article, we propose instrumental-variable calibrated estimators by using the classical instrumental-variables approach for the case of exact identification that are more efficient than conventional calibration estimators when some auxiliary variables are endogenous. The necessary properties of the proposed estimators are presented. Our study is backed by both the simulation study and a real data example to check the performance of the proposed estimators.

Increasing demand in surveys focused on quantitative characteristics (population mean among others) of controversial issues like corruption, tax evasion, drugs consumption of sensitive variables like spending on drugs or illegal sources of income which lead to lively research, are of randomized response techniques for quantitative variables. Therefore, we propose complex multicriteria evaluation methodology for randomized response techniques for population mean. Based on extensive review in the literature, following ranges of criteria were proposed: statistical properties of estimator, implementation and parameter choice, respondent burden and credibility and confidentiality protection of respondents’ data. Finally, we evaluate in this setting standard techniques using scramble variables and recently proposed techniques of dichotomous question.

Surveys involving sensitive characteristics entail applying innovative tools such as Randomized Response Techniques (RRTs) to elicit responses from the sampled units. This manuscript proposes one such new two-stage RRT, involving two simple and common randomization devices, namely, a coin and a die. Empirical study has been conducted to show its efficacy over contemporary estimator in terms of both Percentage Relative Efficiency and Privacy Protection. Application of the proposed RRT to election polling has been discussed efficiently. The appropriateness of such application has been illustrated via detailed simulation study. The results have been tabulated and recommendations made for the use of the proposed estimation strategy. [ABSTRACT FROM AUTHOR]

Summary: The Iowa Seat‐Belt Use Survey is an annual survey designed to provide estimates of seat‐belt use rates for the state of Iowa in the United States. A desire for county level (substate) estimates motivates the need for small area estimation. Developing a small area model for the seat‐belt survey data is challenging for two mean reasons. First, the data consist of multivariate counts. Second, the same sampling units are observed for five different time points. An appropriate model should reflect multivariate dependencies and the longitudinal data structure. We address these challenges though a unit‐level Bayesian hierarchical model. The observed counts have Poisson distributions. Latent random effects capture multivariate associations and correlations among the observations for the same sampling unit observed at different time points. We employ the posterior predictive distribution for model comparisons. Using the selected model, we construct small area predictors of two measures of seat‐belt use at the county level for 5 years. [ABSTRACT FROM AUTHOR]

Nonprobability (convenience) samples are increasingly sought to reduce the estimation variance for one or more population variables of interest that are estimated using a randomized survey (reference) sample by increasing the effective sample size. Estimation of a population quantity derived from a convenience sample will typically result in bias since the distribution of variables of interest in the convenience sample is different from the population distribution. A recent set of approaches estimates inclusion probabilities for convenience sample units by specifying reference sample-weighted pseudo likelihoods. This paper introduces a novel approach that derives the propensity score for the observed sample as a function of inclusion probabilities for the reference and convenience samples as our main result. Our approach allows specification of a likelihood directly for the observed sample as opposed to the approximate or pseudo likelihood. We construct a Bayesian hierarchical formulation that simultaneously estimates sample propensity scores and the convenience sample inclusion probabilities. We use a Monte Carlo simulation study to compare our likelihood based results with the pseudo likelihood based approaches considered in the literature. [ABSTRACT FROM AUTHOR]

In this article, a modified principal component calibration estimator is proposed by using second moments of principal components for estimating the population total in survey sampling using simple random sampling. A simulation scheme and real-life example are used to evaluate the performance of the proposed estimator. The new estimator is more efficient and reduces bias as compared to the conventional principal component calibration estimator. [ABSTRACT FROM AUTHOR]

Increasing demand in surveys focused on quantitative characteristics (population mean among others) of controversial issues like corruption, tax evasion, drugs consumption of sensitive variables like spending on drugs or illegal sources of income which lead to lively research, are of randomized response techniques for quantitative variables. Therefore, we propose complex multicriteria evaluation methodology for randomized response techniques for population mean. Based on extensive review in the literature, following ranges of criteria were proposed: statistical properties of estimator, implementation and parameter choice, respondent burden and credibility and confidentiality protection of respondents' data. Finally, we evaluate in this setting standard techniques using scramble variables and recently proposed techniques of dichotomous question. [ABSTRACT FROM AUTHOR]

Survey sampling has a wide range of applications in biomedical, meteorological, stock exchange, marketing, and agricultural research based on data collected through sample surveys or experimentation. The collected set of information may have a fuzzy nature, be indeterminate, and be summarized by a fuzzy number rather than a crisp value. The neutrosophic statistics, a generalization of fuzzy statistics and classical statistics, deals with the data that have some degree of indeterminacy, imprecision, and fuzziness. In this article, we introduce a fuzzy decision-making approach for deciding a sample size under a fuzzy measurement cost modeled by an intuitionistic fuzzy cost function. Our research introduces neutrosophic ratio-type estimators for estimating the population mean of the neutrosophic study variable Y N ∈ [ Y L , Y U ] utilizing all the indeterminate values of the neutrosophic auxiliary variable X N ∈ [ X L , X U ] rather than only the extreme values X L and X U . Three simulation studies are carried out to explain the proposed methods of parameter estimation, sample size determination, and efficiency comparison. The results reveal that the proposed neutrosophic class of estimators produces more accurate and precise estimates of the neutrosophic population mean than the existing neutrosophic estimators in simple random sampling, which is the ultimate goal of inferential statistics. [ABSTRACT FROM AUTHOR]

Unit-level modeling strategies offer many advantages relative to the area-level models that are most often used in the context of small area estimation. For example, unit-level models aggregate naturally, allowing for estimates at any desired resolution, and also offer greater precision in many cases. We compare a variety of the methods available in the literature related to unit-level modeling for small area estimation. Specifically, to provide insight into the differences between methods, we conduct a simulation study that compares several of the general approaches. In addition, the methods used for simulation are further illustrated through an application to the American Community Survey. [ABSTRACT FROM AUTHOR]

Model-based small area estimation is frequently used in conjunction with survey data to establish estimates for under-sampled or unsampled geographies. These models can be specified at either the area-level, or the unit-level, but unit-level models often offer potential advantages such as more precise estimates and easy spatial aggregation. Nevertheless, relative to area-level models, literature on unit-level models is less prevalent. In modeling small areas at the unit level, challenges often arise as a consequence of the informative sampling mechanism used to collect the survey data. This article provides a comprehensive methodological review for unit-level models under informative sampling, with an emphasis on Bayesian approaches. [ABSTRACT FROM AUTHOR]

Background: Population-based seroprevalence studies are crucial to understand community transmission of COVID-19 and guide responses to the pandemic. Seroprevalence is typically measured from diagnostic tests with imperfect sensitivity and specificity. Failing to account for measurement error can lead to biased estimates of seroprevalence. Methods to adjust seroprevalence estimates for the sensitivity and specificity of the diagnostic test have largely focused on estimation in the context of convenience sampling. Many existing methods are inappropriate when data are collected using a complex sample design. Methods: We present methods for seroprevalence point estimation and confidence interval construction that account for imperfect test performance for use with complex sample data. We apply these methods to data from the Chatham County COVID-19 Cohort (C4), a longitudinal seroprevalence study conducted in central North Carolina. Using simulations, we evaluate bias and confidence interval coverage for the proposed estimator compared with a standard estimator under a stratified, three-stage cluster sample design. Results: We obtained estimates of seroprevalence and corresponding confidence intervals for the C4 study. SARS-CoV-2 seroprevalence increased rapidly from 10.4% in January to 95.6% in July 2021 in Chatham County, North Carolina. In simulation, the proposed estimator demonstrates desirable confidence interval coverage and minimal bias under a wide range of scenarios. Conclusion: We propose a straightforward method for producing valid estimates and confidence intervals when data are based on a complex sample design. The method can be applied to estimate the prevalence of other infections when estimates of test sensitivity and specificity are available. [ABSTRACT FROM AUTHOR]

Antoch, Mola and Vozár (2022) proposed recently new randomized response technique for population mean or total of a quantitative variable. The aim of the paper is to solve the open problem to derive unbiased variance estimator of these procedures. In their proposal, unlike other randomized response techniques for population mean or total the randomized response is not a linear function of a sensitive variable. However, standard techniques to derive variance estimators in this setting are based on this assumption. That is why an interviewer needs also to know values pseudorandom numbers (i.e., results of individual randomization experiments). Respondents might perceive this relaxation of privacy protection negative. The performance of the approximate two-sided confidence intervals of distributions with different shape including their coverage is assessed by a simulation study for simple random sampling without replacement.

In cluster-specific studies, ordinary logistic regression and conditional logistic regression for binary outcomes provide maximum likelihood estimator (MLE) and conditional maximum likelihood estimator (CMLE), respectively. In this paper, we show that CMLE is approaching to MLE asymptotically when each individual data point is replicated infinitely many times. Our theoretical derivation is based on the observation that a term appearing in the conditional average log-likelihood function is the coefficient of a polynomial, and hence can be transformed to a complex integral by Cauchy's differentiation formula. The asymptotic analysis of the complex integral can then be performed using the classical method of steepest descent. Our result implies that CMLE can be biased if individual weights are multiplied with a constant, and that we should be cautious when assigning weights to cluster-specific studies. [ABSTRACT FROM AUTHOR]

There is a lack of data visualization tools, such as UpSet plot, suitable for survey data collected via complex sampling. Ignoring the design specifications when plotting the data could result in ambiguous findings. We discuss a two-step approach that can be used to construct the UpSet plots of weighted frequencies and relative weighted frequencies (percentages). In the first step we compute weighted frequencies corresponding to all possible intersection sets using the SAS® Survey Package software. In the second step, we import the SAS output table in R and use it to construct the UpSet plots in R. We illustrate the application of the two-step method in the study of perceived reasons for e-cigarette use among current e-cigarette users. In the study we analyzed reports of adults (n = 3,614) who responded to the 2014–15 Tobacco Use Supplement to the Current Population Survey. The UpSet plots were constructed to identify the most commonly reported sets of reasons for e-cigarette use. We expect that the discussed method can be adopted for analysis of data from other complex surveys and thus, offer a valuable tool for constructing UpSet plots of intersection sets based on survey data with weights. [ABSTRACT FROM AUTHOR]

Inverse probability weighting (IPW) is a general tool in survey sampling and causal inference, used in both Horvitz–Thompson estimators, which normalize by the sample size, and Hájek/self-normalized estimators, which normalize by the sum of the inverse probability weights. In this work, we study a family of IPW estimators, first proposed by Trotter and Tukey in the context of Monte Carlo problems, that are normalized by an affine combination of the sample size and a sum of inverse weights. We show how selecting an estimator from this family in a data-dependent way to minimize asymptotic variance leads to an iterative procedure that converges to an estimator with connections to regression control methods. We refer to such estimators as adaptively normalized estimators. For mean estimation in survey sampling, the adaptively normalized estimator has asymptotic variance that is never worse than the Horvitz–Thompson and Hájek estimators. Going further, we show that adaptive normalization can be used to propose improvements of the augmented IPW (AIPW) estimator, average treatment effect (ATE) estimators, and policy learning objectives. Appealingly, these proposals preserve both the asymptotic efficiency of AIPW and the regret bounds for policy learning with IPW objectives, and deliver consistent finite sample improvements in simulations for all three of mean estimation, ATE estimation, and policy learning.

Rubin’s variance estimator of the multiple imputation estimator for a domain mean is not asymptotically unbiased. Kim et al. derived the closed-form bias for Rubin’s variance estimator. In addition, they proposed an asymptotically unbiased variance estimator for the multiple imputation estimator when the imputed values can be written as a linear function of the observed values. However, this needs the assumption that the covariance of the imputed values in the same imputed dataset is twice that in the different imputed datasets. In this study, we proposed a bootstrap variance estimator that does not need this assumption. Both theoretical argument and simulation studies show that it was unbiased and asymptotically valid. The new method was applied to the Hox pupil popularity data for illustration.

There are established methods for estimating disease prevalence with associated confidence intervals for complex surveys with perfect assays, or simple random sample surveys with imperfect assays. We develop and study methods for the complicated case of complex surveys with imperfect assays. The new methods use the melding method to combine gamma intervals for directly standardized rates and established adjustments for imperfect assays by estimating sensitivity and specificity. One of the new methods appears to have at least nominal coverage in all simulated scenarios. We compare our new methods to established methods in special cases (complex surveys with perfect assays or simple surveys with imperfect assays). In some simulations, our methods appear to guarantee coverage, while competing methods have much lower than nominal coverage, especially when overall prevalence is very low. In other settings, our methods are shown to have higher than nominal coverage. We apply our method to a seroprevalence survey of SARS‐CoV‐2 in undiagnosed adults in the United States between May and July 2020. [ABSTRACT FROM AUTHOR]

Modeling nonlinear data is a common practice in data science and Machine Learning (ML). It is aberrant for the outcome of a natural process to vary linearly with the values of input variable(s). A robust and easy methodology is needed for accurately and quickly fitting a sampled dataset with a set of covariates, assuming that the sampled data can be a complicated nonlinear function. A novel approach to the estimation of finite population parameter r, which is a linear combination of the population values, is considered in this article under superpopulation setting with known Basis Functions Regression (BFR) models. The problems of subsets selection with a single predictor using an automatic matrix approach and ill-conditioned regression models are discussed. Prediction error variance of the proposed estimator is estimated based on widely used feature selection criteria in ML. Finally, the Expected Squared Prediction Error (ESPE) of the proposed estimator and the expectation of estimated error variance under bootstrapping as well as simulation study with different regularizers are obtained to observe the long-run behavior of the proposed estimator. [ABSTRACT FROM AUTHOR]

Summary: Several exact inference procedures for logistic regression require the simulation of a 0‐1 dependent vector according to its conditional distribution, given the sufficient statistics for some nuisance parameters. This is viewed, in this work, as a sampling problem involving a population of n units, unequal selection probabilities and balancing constraints. The basis for this reformulation of exact inference is a proposition deriving the limit, as n goes to infinity, of the conditional distribution of the dependent vector given the logistic regression sufficient statistics. It is proposed to sample from this distribution using the cube sampling algorithm. The interest of this approach to exact inference is illustrated by tackling new problems. First it allows to carry out exact inference with continuous covariates. It is also useful for the investigation of a partial correlation between several 0‐1 vectors. This is illustrated in an example dealing with presence‐absence data in ecology. [ABSTRACT FROM AUTHOR]

Survey calibration is a widely used method to estimate the population mean or total score of a target variable, particularly in medical research. In this procedure, auxiliary information related to the variable of interest is used to recalibrate the estimation weights. However, when the auxiliary information includes qualitative variables, traditional calibration techniques may be not feasible or the optimisation procedure may fail. In this article, we propose the use of linear calibration in conjunction with a multidimensional scaling-based set of continuous, uncorrelated auxiliary variables along with a suitable metric in a distance-based regression framework. The calibration weights are estimated using a projection of the auxiliary information on a low-dimensional Euclidean space. The approach becomes one of the linear calibration with quantitative variables avoiding the usual computational problems in the presence of qualitative auxiliary information. The new variables preserve the underlying assumption in linear calibration of a linear relationship between the auxiliary and target variables, and therefore the optimal properties of the linear calibration method remain true. The behaviour of this approach is examined using a Monte Carlo procedure and its value is illustrated by analysing real data sets and by comparing its performance with that of traditional calibration procedures. [ABSTRACT FROM AUTHOR]

It is a common interest in medicine to determine whether a hospital meets a benchmark created from an aggregate reference population, after accounting for differences in distributions of multiple covariates. Due to the difficulties of collecting individual-level data, however, it is often the case that only marginal distributions of the covariates are available, making covariate-adjusted comparison challenging. We propose and evaluate a novel approach for conducting indirect standardization when only marginal covariate distributions of the studied hospital are known, but complete information is available for the reference hospitals. We do this with the aid of two existing methods: iterative proportional fit, which estimates the cells of a contingency table when only marginal sums are known, and synthetic control methods, which create a counterfactual control group using a weighted combination of potential control groups. The proper application of these existing methods for indirect standardization would require accounting for the statistical uncertainties induced by a situation where no individual-level data is collected from the studied population. We address this need with a novel method which uses a random Dirichlet parametrization of the synthetic control weights to estimate uncertainty intervals for the standard incidence ratio. We demonstrate our novel methods by estimating hospital-level standardized incidence ratios for comparing the adjusted probability of computed tomography examinations with high radiations doses, relative to a reference standard and we evalauate out methods in a simulation study.

A new randomized response technique for estimating the population total, or the population mean of a quantitative variable is proposed. It provides a high degree of protection to the respondents because they never report their data. Therefore, it may be favorably perceived by them and increase their willingness to cooperate. Instead of revealing the true value of the characteristic under investigation, the respondent only states whether the value is greater (or smaller) than a number which is selected by him/her at random and is unknown to the interviewer. For each respondent, this number, a sort of individual threshold, is generated as a pseudorandom number. Furthermore, two modifications of the proposed technique are presented. The first modification assumes that the interviewer also knows the generated random number. The second modification deals with the issue that, for certain variables, such as income, it may be embarrassing for the respondents to report either high or low values. Thus, depending on the value of the fixed threshold (unknown to the respondent), the respondent is asked different questions to avoid being embarrassed. The suggested approach is applied in detail to the simple random sampling without replacement, but it can be, after a straightforward modification, applied to many sampling schemes, including cluster sampling, two-stage sampling, or stratified sampling. The results of the simulations illustrate the behavior of the proposed technique.

Remote sensing (RS) has enhanced forest inventory with model-based inference, that is, a family of statistical procedures rigorously estimates the parameter of a variable of interest (VOI) for a spatial population, e.g., the mean or total of forest carbon for a study area. Upscaling in earth observation, alias to this estimation, aggregates VOI from a finer spatial resolution to a coarser one with reduced uncertainty, serving decision making for natural resource management at larger scales. However, conventional model-based estimation (CMB) confronts a major challenge: it only supports RS wall-to-wall data, meaning that remotely sensed data must be available in panorama and non-wall-to-wall but quality data such as lidar or even cloud-masked satellite imagery are not supported due to incomplete coverage, impeding precise upscaling with cutting-edge instruments or for large scale applications. Consequently, this study aims to develop and demonstrate the use and usefulness of RS non-wall-to-wall data for upscaling with Hierarchical model-based estimation (HMB) which incorporates a two-stage model for bridging RS non– and wall-to-wall data; and for optimizing cost-efficiency, to evaluate the effects of non-wall-to-wall sample size on upscaling precision. Three main conclusions are relevant: (1) the HMB is a variant of the CMB estimator through trading in the uncertainty of the second-stage model to enable estimation using RS non-wall-to-wall data; (2) a quality first-stage model is key to exerting the advantage of HMB relative to the CMB estimator; (3) the variance of the HMB estimator is dominated by the first-stage model variance component, indicating that increasing the sample size in the first-stage is effective for increasing the overall precision. Overall, the HMB estimator balances tradeoffs between cost, efficiency and flexibility when devising a model-based upscaling in earth observation.

Web surveys have replaced Face‐to‐Face and computer assisted telephone interviewing (CATI) as the main mode of data collection in most countries. This trend was reinforced as a consequence of COVID‐19 pandemic‐related restrictions. However, this mode still faces significant limitations in obtaining probability‐based samples of the general population. For this reason, most web surveys rely on nonprobability survey designs. Whereas probability‐based designs continue to be the gold standard in survey sampling, nonprobability web surveys may still prove useful in some situations. For instance, when small subpopulations are the group under study and probability sampling is unlikely to meet sample size requirements, complementing a small probability sample with a larger nonprobability one may improve the efficiency of the estimates. Nonprobability samples may also be designed as a mean for compensating for known biases in probability‐based web survey samples by purposely targeting respondent profiles that tend to be underrepresented in these surveys. This is the case in the Survey on the impact of the COVID‐19 pandemic in Spain (ESPACOV) that motivates this paper. In this paper, we propose a methodology for combining probability and nonprobability web‐based survey samples with the help of machine‐learning techniques. We then assess the efficiency of the resulting estimates by comparing them with other strategies that have been used before. Our simulation study and the application of the proposed estimation method to the second wave of the ESPACOV Survey allow us to conclude that this is the best option for reducing the biases observed in our data. [ABSTRACT FROM AUTHOR]

Many surveys are performed using non-probability methods such as web surveys, social networks surveys, or opt-in panels. The estimates made from these data sources are usually biased and must be adjusted to make them representative of the target population. Techniques to mitigate this selection bias in non-probability samples often involve calibration, propensity score adjustment, or statistical matching. In this article, we consider the problem of estimating the finite population distribution function in the context of non-probability surveys and show how some methodologies formulated for linear parameters can be adapted to this functional parameter, both theoretically and empirically, thus enhancing the accuracy and efficiency of the estimates made. [ABSTRACT FROM AUTHOR]

Although K-fold cross-validation (CV) is widely used for model evaluation and selection, there has been limited understanding of how to perform CV for non-iid data, including those from sampling designs with unequal selection probabilities. We introduce CV methodology that is appropriate for design-based inference from complex survey sampling designs. For such data, we claim that we will tend to make better inferences when we choose the folds and compute the test errors in ways that account for the survey design features such as stratification and clustering. Our mathematical arguments are supported with simulations, and our methods are illustrated on real survey data. [ABSTRACT FROM AUTHOR]

Rubin's variance estimator of the multiple imputation estimator for a domain mean is not asymptotically unbiased. Kim et al. derived the closed-form bias for Rubin's variance estimator. In addition, they proposed an asymptotically unbiased variance estimator for the multiple imputation estimator when the imputed values can be written as a linear function of the observed values. However, this needs the assumption that the covariance of the imputed values in the same imputed dataset is twice that in the different imputed datasets. In this study, we proposed a bootstrap variance estimator that does not need this assumption. Both theoretical argument and simulation studies show that it was unbiased and asymptotically valid. The new method was applied to the Hox pupil popularity data for illustration. [ABSTRACT FROM AUTHOR]

Survey sampling has a wide range of applications in biomedical, meteorological, stock exchange, marketing, and agricultural research based on data collected through sample surveys or experimentation. The collected set of information may have a fuzzy nature, be indeterminate, and be summarized by a fuzzy number rather than a crisp value. The neutrosophic statistics, a generalization of fuzzy statistics and classical statistics, deals with the data that have some degree of indeterminacy, imprecision, and fuzziness. In this article, we introduce a fuzzy decision-making approach for deciding a sample size under a fuzzy measurement cost modeled by an intuitionistic fuzzy cost function. Our research introduces neutrosophic ratio-type estimators for estimating the population mean of the neutrosophic study variable YN∈[YL,YU] utilizing all the indeterminate values of the neutrosophic auxiliary variable XN∈[XL,XU] rather than only the extreme values XL and XU. Three simulation studies are carried out to explain the proposed methods of parameter estimation, sample size determination, and efficiency comparison. The results reveal that the proposed neutrosophic class of estimators produces more accurate and precise estimates of the neutrosophic population mean than the existing neutrosophic estimators in simple random sampling, which is the ultimate goal of inferential statistics.

The calibration method has been widely used to incorporate auxiliary information in the estimation of various parameters. Specifically, some authors adapted this method to estimate the distribution function, although their proposal is computationally simple, its efficiency depends on the selection of an auxiliary vector of points. This work deals with the problem of selecting the calibration auxiliary vector that minimizes the asymptotic variance of the calibration estimator of distribution function. The optimal dimension of the optimal auxiliary vector is reduced considerably with respect to previous studies so that with a smaller set of points, the minimum of the asymptotic variance can be reached, which in turn allows to improve the efficiency of the estimates. [ABSTRACT FROM AUTHOR]

In large-area forest inventories, a trade-off between the amount of data to be sampled and the corresponding collection costs is necessary. It is not always possible to have a very large data sample when dealing with sampling-based inventories. It is therefore important to optimize the sampling design with the limited resources. Whereas this sort of inventories are subject to these constraints, the availability of remote sensing (RS) data correlated with the forest inventory variables is usually much higher. For this reason, the RS and sampled field measurement data are often used in combination for improving the forest inventory estimation. In this study, we propose a model-based data sampling method founded on Bayesian optimization and machine learning algorithms which utilizes RS data to guide forest inventory sample selection. We evaluate our method in empirical experiments using real-world volume of growing stock data from the Aland region in Finland. The proposed method is compared against two baseline methods: simple random sampling and the local pivotal method. When a suitable model link is selected, the empirical experiments show on best case on average up to 22% and 79% improvement in population mean and variance estimation respectively over baselines. However, the results also illustrate the importance of model selection which has a clear effect on the results. The novelty of the study is in the application of Bayesian optimization in national forest inventory survey sampling. [ABSTRACT FROM AUTHOR]