Survey Sampling

Some social surveys address sensitive topics for which respondents do not report reliable responses. Randomized response techniques (RRTs) are employed to increase privacy levels and provide honest answers. However, estimates obtained from this method tend to exhibit increased variances. Repeating randomized responses for each individual increases the sample size, and the mean of observations for each individual reduces the variance of the parameter's estimator, bringing them closer to reality. In this study, considering continuous additive repeated randomized responses (RRRs), we apply the averaged RR of each individual using the linear regression model for sensitive variable mean. Data on the income of family heads were collected from students, and each respondent was asked to randomize their responses five times. The maximum likelihood estimators of parameters are obtained by two methods. In the first method, the response variable is the first reported observation, and in the second method, we considered the averaged RR for each individual. The results emphasize that the estimators from the second method are closer to reality and have lower variance.

Near-cutoff sampling for multiple-attribute establishment surveys can be very useful. When the same data items are in an occasional census as those in more frequent samples, testing and accurate production of Official Statistics is possible, with proper stratification, when needed. -
This poster was for the 2022 Joint Statistical Meetings (JSM) in Washington DC.

Various uses of cutoff sampling are shown. (Near-cutoff sampling with ratio model prediction can work very well for Official Statistics.)  -  Poster for a speed session for the International Conference on Establishment Statistics ICES VII, Glasgow 2024.

"Knaub, J.R., Jr. (2017), 'Comparison of Model-Based to Design-Based Ratio Estimators,' JSM Proceedings, Survey Research Methods Section, American Statistical Association, was on this topic and focused on fundamentals, whereas here we focus on the analysis of structure and performance. In studying this topic, lesser-known facts appear to contribute greatly to the structure and performance of various methods. Heteroscedasticity is a key feature. --- This is for a conference, the JSM 2026, to be held in August 2026, in Boston."

Abstract:
Ratio models, used to assist in estimation and for prediction, are a basis for more advanced methods of inference from samples to populations, providing valuable insight, with their own substantial practical applications. Performances are compared for model-assisted design-based ratio estimators and prediction-based ratio estimators, which all consider heteroscedasticity. Randomized selection as a basis for inference is compared to prediction, where a model for a common mean for each member of a population, for a given data item, is analogous to simple random sampling. When a size measure, variable predicted-y, is introduced, regression is possible. For ratio models, the size measure for each item would be bx, or just x, as relative size is what is needed here. Then, a multiple-attribute sample may be selected without the same size measure for each item during estimation/prediction. Further, there are some lesser-known facts regarding ratio estimation and/or ratio-model prediction: one, due to Ken Brewer, might impact new work; another involves accuracy in repeated surveys; others include adjustments to variance estimators, sampling, and other issues impacting structure and performance. -

This is a poster scheduled for the Joint Statistical Meetings (JSM) 2026, in August, in Boston. - 
June 4, 2026 - JRK

Les formations végétales en Algérie sont façonnées par une combinaison de facteurs climatiques, édaphiques et anthropozoïques, qui influencent continuellement leur physionomie. Cette flore se développe sous un climat typiquement méditerranéen, marqué par des précipitations concentrées durant la saison froide et une sécheresse estivale prolongée s'étendant sur plusieurs mois (jusqu'à 7 à 8 mois), période durant laquelle les risques d'incendies sont élevés. Le substrat, quant à lui, se caractérise par une forte hétérogénéité, dominée par des sols marneux ou calcaires. Par ailleurs, l'action anthropozoïque incessante -due au pâturage, aux incendies volontaires ou involontaires, ainsi qu'aux coupes illicitesperturbe et dégrade durablement ces milieux naturels.  Phytogéographie : Étude de la répartition des végétaux et de leurs causes. Considérée comme l'ancêtre de la phytosociologie.  Phytoécologie : Approche réductionniste initiée par Louis Emberger. Elle étudie la végétation à travers le prisme du taxon (l'espèce). On établit le profil écologique de chaque espèce par rapport à des variables mesurées (pH, humidité...). Les espèces aux profils similaires sont regroupées en "groupes écologiques", et les communautés végétales sont vues comme l'assemblage de ces groupes. C'est une approche très utilisée en typologie des stations forestières.  Phytosociologie : Approche holistique qui étudie directement la communauté végétale (l'association) comme une entité en soi, avant de la caractériser écologiquement. Avant d'entamer toute phase d'échantillonnage et de relevé de végétation, il est impératif de définir clairement le cadre de l'étude en s'interrogeant sur les points suivants : quel est l'objet précis de la recherche ? Dans quel contexte s'inscrit-elle ? Quels sont les objectifs visés et les hypothèses sous-jacentes ? Quels moyens sont disponibles ? À quel public destination les résultats ? Et enfin, quelles valorisations en sont attendues ? Pour illustrer l'importance de ces questionnements, voici des exemples de contextes aux objectifs distincts : • Objectif recherche : la démarche vise à répondre à une question scientifique, par exemple, étudier les impacts des changements globaux sur les communautés végétales. • Objectif gestion : il s'agit ici de réaliser une typologie ou un catalogue d'habitats, comme la rédaction d'un document d'objectifs pour la gestion d'un site. Exemple: Dresser un catalogue exhaustif des différents types d'habitats forestiers rencontrés sur le site d'étude.

In applying the jackknife procedure to stratified samples, jackknifed variances are often estimated based on overall cross-strata pseudo means. The stratum weights employed in estimating overall cross-strata pseudo means may deviate from the stratum weights based on the sample frame. The focus of this study is on investigating the effects of the departure of stratum weights on overall cross-strata jackknifed variance estimation. Canonical variance analysis is employed to partition the total variation into several components, the formula based variance, the variance due to the departure of stratum weights, and the interaction part. The partition of variation leads to a better understanding of the structure of the variance estimation for stratified samples; it enhances the efficiency of the jackknife procedure and provides valid variance estimation for stratified combined samples. The empirical results are based on a real large-scale assessment; nevertheless, the findings can be extended to offer analyses of stratified samples in which jackknifing procedure is employed.

Jiahe Qian This paper discusses minimum mean square error (MSE) estimation to a population in the context of stratified simple random sampling. We have derived adjusted stratum weights that minimize the MSE of the weighted sample mean. Moreover, we have proposed a generalized model for the minimum MSE weights of a mean estimator when the bias of the estimated mean is constrained in size. The estimators based on the minimum MSE weights are consistent with the design. This paper has shown that the mean estimator with the minimum MSE weights from a stratified sample is equivalent in expectation to a Stein-type estimator. This connection reveals the relationship among estimates based on the minimum MSE weights, the minimum variance weights, and the unbiased case weights. The application of Stein estimator is justified as an optimal compromise between a designunbiased approach and a model-dependent approach to survey data. An inspection of the properties of the weights shows that bias classifies them into equivalent groups which form lines on the plane defined by the weights. To observe the effects of bias upon the mean estimator, this paper analyzes the squared relative bias, i.e. ratio of the bias squared to variance. The empirical estimates of the minimum MSE weights and their related iii .

El Análisis de Redes es un enfoque
de estudio de reciente creación y popularización, al grado que
cada día son más las disciplinas científicas sociales que lo aplican a su materia de estudio. Su perspectiva especial consiste en el uso de modelos y aplicaciones que consideran en forma relacional
a los individuos o instituciones.

The problem mentioned here is to obtain unbiased estimate of the proportion of people having sensitive characteristics like accumulated savings, intentional tax evasion, drunken driving etc. in a society. Warner presented a technique known as Randomized Response Technique for reducing evasive answer bias or non response bias. Several researchers extended different Randomized Response models to complex surveys. Amitava extended Kim and Warde model to complex surveys. Here we present an alternative form of Kim and warde's mixed Randomized Response Model in complex surveys. Our proposed model performs very well when P 2 is close to 0.5 and the results also good but not up to a great extent when P 2 is far away from 0.5.

The Franciscana dolphin ( Pontoporia blainvillei), a small cetacean endemic to southwestern Atlantic coastal waters, is the most endangered marine mammal species in the south Atlantic. In the Espírito Santo State, in southeastern Brazil, the Franciscana dolphin distribution overlaps regions suppressed and threatened by an intensive industrial and port development in the coastal zone. Moreover, most of Franciscana’s home range was recently impacted by the collapse of a dam that released millions of iron mining waste into the Atlantic Ocean. Considering the restricted coastal habitat of the species, these impacts can be considered an immediate threat, increasing the probability of local extinction of the species. This research calls attention to the necessity for improvement in the regional Environmental Impact Assessments, Conservation Action Plans, and to the development of an effective zoning for ports and conservation areas in the region. These actions could increase conservation ...

A wide variety of local weight and land measurement units in Bangladesh. These are considered a major problem in registering land records and in pricing and marketing agricultural commodities. This study reports the findings of a countrywide sample survey to highlight the characteristics of the units and their implications.

The central limit theorem is central to modern statistics, particularly for sampling theory. Despite what you may have heard, it hasn’t been proven mathematically in relation to samples, unless they approach infinity in size, which samples never do. You may also have heard that it is the reason so many phenomena follow a normal distribution curve. This isn’t the case, and it can’t be. In spite of that, the central limit theorem is vitally important, and it really does work. This material explains why.

In this paper, we study, within a modeling framework, the joint treatment of nonignorable dropout and informative sampling for longitudinal survey data, by specifying the probability distribution of the observed measurements when the sampling design is informative. The sample distribution of the observed measurements model is extracted from the population distribution model, assumed to be multivariate normal. The sample distribution is derived first by identifying and estimating the conditional expectations of first order sample inclusion probabilities, (assuming complete response at the first time period), given the study variable, based on a variety of models, such as linear, exponential, logit and probit. Next, we consider a logistic model for the informative dropout process. The proposed method combines two methodologies used in the analysis of sample surveys: for the treatment of informative sampling and informative dropout. One incorporates the dependence of the first order inclusion probabilities at the initial time period on the study variable, see , while the other incorporates the dependence of the probability of nonresponse on unobserved or missing observations, see . An empirical example based on British Labour Force Survey data illustrates the methods proposed.

Objectives: Cycle 5 of the National Survey of Family Growth (NSFG) was conducted by the National Center for Health Statistics (NCHS) in 1995. The NSFG collects data on pregnancy, childbearing, and women's health from a national sample of women 15-44 years of age. This report describes how the sample was designed, shows response rates for various subgroups of women, describes how the sampling weights were computed to make national estimates possible, shows how missing data were imputed for a limited set of key variables, and describes the proper ways to estimate sampling errors from the NSFG. The report includes both nontechnical summaries for readers who need only general information and more technical detail for readers who need an in-depth understanding of these topics. Methods: The 1995 NSFG was based on a national probability sample of women 15-44 years of age in the United States and was drawn from 14,000 households interviewed in the 1993 National Health Interview Survey (NHIS). Of the 13,795 women eligible for the NSFG, 10,847 (79 percent) gave complete interviews. Results: This report recommends using weighted data for analysis and a software package that will estimate sampling errors from complex samples (for example, SUDAAN or comparable software). The rate of missing data in the 1995 NSFG was very low. However, missing data were imputed for 315 key variables, called "recodes." Of the 315 recodes defined for Cycle 5, 271 variables had missing data on less than 1 percent of the cases; only 44 had 1 percent or more with missing data. These missing values were imputed for all of these 315 variables. The imputation procedures are described in this report.

Boxplots showing distribution of all antecedent dry periods and the antecedent dry periods for runoff events during which runoff samples were collected at U.S. Geological Survey monitoring stations on State Route 2A in Boston (422108071052501), Interstate 90 near Weston (422025071154501), and State Route 20 near Quinsigamond Village (421247071470201) in eastern Massachusetts, 2014-16 .

The normalized odd Collatz map is analyzed as a discrete dynamical system in the ring of 2-adic integers. It is proven that the existence of periodic macro-cycles is equivalent to the solvability of a Diophantine system governed by an ensemble of projective 2×2 matrices. The finiteness of a macro-cycle is established to be consistent with the formation of invariant attractor hole windows. The topological isolation of these windows induces an overdetermined system of independent cyclic traversals, which strictly necessitates bitwise palindromic symmetry of the local 2-adic configurations (hole kernels). It is shown that for window widths L ≥ 9, the fundamental incommensurability of left-sided multiplicative expansion and right-sided additive truncation deterministically breaks this symmetry. Within the kinematic corridor of the orbit, this dynamical symmetry breaking algebraically forbids the existence of real periodic macro-cycles of length N > 7.

Social surveys generally assume that a sample of units (students, individuals, employees, . . . ) is observed by two-stage selection from a finite population, which is grouped into clusters (schools, household, companies, . . . ). This design involves sampling from two different populations: the population of schools or primary stage units and the population of students or secondstage units. Calibration estimators for student statistics can be defined by using combined information based on school totals and student totals. Auxiliary information from the units at the two stages can be calibrated by integrated weighting, as proposed by Lemaı ˆtre and Dufour or Estevao and Sa ¨rndal. Two calibration estimators for the population total based on unit weights are defined. The first estimator satisfies a calibration equation at the unit level, and the second one, at the cluster level. The proposed estimator shrinks the unit estimator toward the cluster. A simulation study based on two real populations is carried out to study the empirical performance of this shrinkage estimator. The populations studied were obtained from the Programme for International Student Assessment database and from the Spanish Household Budget Survey.

The calibration method has been widely discussed in the recent literature on survey sampling, and calibration estimators are routinely computed by many survey organizations. The calibration technique was introduced in [12] to estimate linear parameters as mean or total. Recently, some authors have applied the calibration technique to estimate the finite distribution function and the quantiles. The computationally simpler method in is built by means of constraints that require the use of a fixed value t 0 . The precision of the resulting calibration estimator changes with the selected point t 0 . In the present paper, we study the problem of determining the optimal value t 0 that gives the best estimation under simple random sampling without replacement. A limited simulation study shows that the improvement of this optimal calibrated estimator over possible alternatives can be substantial.

The convenience of online surveys has quickly increased their popularity for data collection. However, this method is often non-probabilistic as they usually rely on selfselection procedures and internet coverage. These problems produce biased samples. In order to mitigate this bias, some methods like Statistical Matching and Propensity Score Adjustment (PSA) have been proposed. Both of them use a probabilistic reference sample with some covariates in common with the convenience sample. Statistical Matching trains a machine learning model with the convenience sample which is then used to predict the target variable for the reference sample. These predicted values can be used to estimate population values. In PSA, both samples are used to train a model which estimates the propensity to participate in the convenience sample. Weights for the convenience sample are then calculated with those propensities. In this study, we propose methods to combine both techniques. The performance of each proposed method is tested by drawing nonprobability and probability samples from real datasets and using them to estimate population parameters.

This study introduces a general framework on inference for a general parameter using nonprobability survey data when a probability sample with auxiliary variables, common to both samples, is available. The proposed framework covers parameters from inequality measures and distribution function estimates but the scope of the paper is broader. We develop a rigorous framework for general parameter estimation by solving survey weighted estimating equations which involve propensity score estimation for units in the non-probability sample. This development includes the expression of the variance estimator, as well as some alternatives which are discussed under the proposed framework. We carried a simulation study using data from a real-world survey, on which the application of the estimation methods showed the effectiveness of the proposed design-based inference on several general parameters.

ipfweight performs a stepwise adjustment (known as iterative proportional fitting or raking) of survey sampling weights to achieve known population margins. The iterative process is repeated until the difference between the sample margins and the known population ...

The controversy surrounding the Mandatory Country-of-Origin Labeling (COOL) has attracted research attentions. A number of studies have reported consumers are willing to pay more for beef labeled with U.S. origin versus beef from unknown or other origins. Despite that, relatively little is known about what motivates consumers’ preference for origin-labeled food products (Lusk et al 2006). Using Individual-Level Parameters following a mixed logit model, we found that U.S. consumers were willing to pay significantly less for imported steak from Australia and Canada compare to U.S. steak. Further, we found that the negative willingness to pay is associated strongly with consumers’ perception of food safety on the exporting country

Applications of several adjustment methods and the Without Replacement Bootstrap (BWO) are presented, using data from the 1997 Annual Business Survey, conducted by Portugal's National Statistics Institute. The application of these methods is motivated by issues related with frame problems and total nonresponse. The BWO algorithm is used for variance estimation of the considered estimators: the adjustment cell estimator, the post-stratified estimator and the post-stratified estimator with adjustment cells. The results agree with the theoretical evidences referred in the literature.

This article describes the prevalence of multi-morbidity and its association with socioeconomic and demographic factors using National Sample Survey 2017-18 data, on 42756 older adults aged 60+. The prevalence of multimorbidity is assessed as the count of self-reported morbidity among the older adults. Bivariate and logistic regression is estimated to examine the covariates of multi-morbidity. Overall, our study found that there was 277 older adults per 1000 population who reported their morbidity, co-morbidity and multi-morbidity. The prevalence of multi-morbidity was higher at a higher level of education, income and in urban areas. While a higher prevalence in multi-morbidity is observed among the other backward community compared to the forward caste. The multivariate analysis suggests that compared to individuals aged 70 years and above were more likely to report multi-morbidity as compared to individuals aged 60-64 years [OR:2.17; CI: 1.602-2.939]. In the economic groups, individuals belonging top 20% of the income groups has higher odds of reporting multi-morbidity as compared to those individuals belonging to the bottom 20% of the population [OR; 10.63, CL; 5.223-21.657]. Further, hypertension along with diabetes was the most common reported multi-morbidity. Policy initiatives aiming to reduce the burden of these multi-morbidities must focus on reorienting programme through increasing public investment in health. These investments must increase by providing preventive care facilities for the early prevention of diseases and improving the provision of geriatric curative services both in rural and urban areas as needed by the community in India.

• Meaning and Importance of Population and Sample
• Meaning and Principles of Sampling
• Probability Sampling designs – Simple Random, Stratified Random, Clustered Sampling
• Non-probability Sampling Designs - Purposive, Convenient and Quota Sampling

Ranked set sampling (RSS) is an important survey technique aimed at efficient estimation of population characteristics. Various RSS methods can be used to collect an RSS sample, but the reproducibility of estimates using these methods remains unexplored. Reproducibility refers to obtaining similar estimates when the survey is repeated under identical conditions. This study compares the reproducibility of population mean estimates using four basic RSS methods: classical RSS (RSS), Median RSS (MRSS), Extreme RSS (ERSS), and Paired RSS (PRSS). We assess the reproducibility of these methods using Nonparametric Predictive Inference (NPI) bootstrapping. Simulations are conducted for varying sample sizes with both perfect and imperfect rankings, and results are compared for weak and strong associations between the study and concomitant variables. Additionally, we apply these methods to agricultural data from Punjab, India. Our findings indicate that MRSS provides the best reproducibility for population mean estimates, while ERSS performs the worst in this regard.

This paper investigates the impact of minimum wages on wages and employment in Greece between 2009 and 2017. Our main contribution is the examination of the effects of minimum wages under a dramatically changing context, as during this period Greece has experienced the deepest recession in its recent history, extensive labour market reforms, and several changes in the minimum wage, including a large decrease. Employing a unique administrative panel matched employer-employee data set and a range of estimators, such as difference-in-differences, fixed effects, and Instrumental Variables, we find that minimum wages have a positive and significant effect on individual and firm-level wages with significant positive wage spill-overs extending, sometimes, above the median wage, but no systematic employment effects.

The paper considers a Finnish human survey that takes advantage of explicit stratification. Stratification is not ordinary, since there are two types of stratification. In one, the strata are municipality based that are much used earlier. But in the second, these take advantage of the grid data base in which the grid size is 250 metres times 250 metres. Using the grid data base the two explicit strata are constructed, one consisting of such grids where the median tax income of the adults is low (called ‘poor’ grids), whereas the median income is high in the second explicit stratum of such grids (‘rich’ grids). In order to correctly handle these both stratified data, it is required first to merge them together, next to create the single sampling weights, and finally, due to 60 per cent non-response to adjust for these weights. This helps in reducing the bias of the estimates. The adjustment is strengthened with good auxiliary data both from the grid data base and other administrative...

This memorandum responds to a request from the House Subcommittee on Census and Population for information concering the planning for the 1990 Census. It identifies eleven major issues concerning the implementation of the 1990 Census: (1) determination of who shall be counted; (2) provision of data for apportionment and redistricting; (3) questionnaire content; (4) use of the Postal Service; (5) use of a two-stage census; (6) use of sampling; (7) use of administrative lists; (8) automation-data processing, computer-assisted telephone interviewing, geographical mapping and procurement; (9) data processing; (10) local review; and (11) use of adjustment procedures. Provided for each issue covered is a brief description of the possible options for the 1990 Census, the dates when the Census Bureau recommends decisions should be made, and an analy'is of alternative legislative proposals. The information on the proposti scaedule for decisions about the procedures to be used for the 1990 Census was obtained from the Census Bureau. The specified dates are the Bureau's estimates as to when decisions should be made if the Census is to proceed efficiently. (PN)

Local species richness and between-site similarity in species composition of parasitoid wasps (Hymenoptera: Ichneumonidae; Pimplinae and Rhyssinae) were correlated with those of four plant groups (pteridophytes, Melastomataceae, Burseraceae and Arecaceae) in a western Amazonian lowland rain forest mosaic. The mosaic structure of the forest was related to variation in soils within the non-inundated terrain. Significant matrix correlation between patterns in parasitoid wasp species composition and plant species composition was found. Most of the overall correlation was due to idiobiont parasitoids of weakly concealed hosts, which attack host larvae and pupae in exposed situations, with two of the four ecologically defined parasitoid groups showing no correlation at all. A positive correlation between the number of plant species and the number of Pimplinae and Rhyssinae species at a site was found when the latter was corrected for collecting effort. Consequently, the degree of floristi...

This paper presents the first record of Naucrates ductor ( Linnaeus 1758) from Syrian waters. One specimen (300 mm TL, 294.29 g TW) was caught by purse-seine nets at about 60 m depth from Lattakia coast, on 25 September 2020. This record represents the first of this cosmopolitan species that entered the Syrian waters (Eastern Mediterranean).

Respondent driven sampling (RDS) is a relatively new network sampling technique typically employed for hard-to-reach populations. Like snowball sampling, initial respondents or "seeds" recruit additional respondents from their network of friends. Under certain assumptions, the method promises to produce a sample independent from the biases that may have been introduced by the non-random choice of "seeds." We conducted a survey on health communication in Guam's general population using the RDS method, the first survey that has utilized this methodology in Guam. It was conducted in hopes of identifying a cost-efficient non-probability sampling strategy that could generate reasonable population estimates for both minority and general populations. RDS data was collected in Guam in 2013 (n=511) and population estimates were compared with 2012 BRFSS data (n=2031) and the 2010 census data. The estimates were calculated using the unweighted RDS sample and the weighte...

This Paper aims to test the causal relationship between the Volatility Index (VIX) and global gold prices during the period from 4 January 2021 to 28 August 2025, using Granger causality testing within a vector autoregressive (VAR) model. The results indicate a unidirectional predictive causality from the VIX to gold prices, supporting gold's role as a hedging instrument during times of market uncertainty. The research recommends that investors and policymakers use the VIX as an early warning bell for potential market volatility, enabling them to adjust their investment strategies and manage risks effectively.

Using a set of random telephone and Internet (web− based) survey samples for a national advisory referendum, we implement Beta models to handle proportional budget information, and allow for consistency in modeling assumptions and the calculation of ...

0 1 1 ( ) ( ) ( ) 0 e e e     . in both the cases I and II. The other expected values ignoring finite population correction (fpc) terms in case I and case II are given by, Case I: Case II: 2 1 1 ( ) x e C n   , 2 0 1 ( ) x e e CC n   ,

This paper proposes two ratio and product-type estimators using transformation based on known minimum and maximum values of auxiliary variable. The biases and mean squared errors of the suggested estimators are obtained under large sample approximation. Conditions are obtained under which the suggested estimators are superior to the conventional unbiased estimator, usual ratio and product estimators of population mean. The superiority of the proposed estimators are also established through some natural population data sets.

This paper is an attempt to develop an estimator for finite population mean. Motivated by , a ratio in ratio type exponential strategy is developed for estimation of population mean in double sampling for stratification. To compare with relevant considered estimators, expressions for bias and mean squared error of the developed estimator have been derived. The developed estimator has been compared with usual unbiased estimator, , ratio estimator and ratio type exponential estimator given by Tailor et al (2014) theoretically as well as empirically.

This paper addresses the problem of estimating the finite population mean of the study variable y in the presence of auxiliary attributes. An improved class of estimators for population mean has been defined along with properties under large sample approximation. An empirical study was carried out to observe the performance of the proposed and the existing estimators. It turns out that the suggested estimators are substantially more precise than the recent estimators envisaged by Koyuncu (2012) and Haq and Shabbir (2014, 2018).

Use of auxiliary information has been in practice to improve the efficiency of the estimators of parameters. Ratio, product and regression methods are good examples of use of auxiliary information. Ratio, product and regression type estimators essentially require the knowledge of population mean of auxiliary variates. But many times, the information on population mean of the auxiliary variate is not available. In this type of situations, double sampling is used. Ajagaonkar (1975) and Sisodia and Dwivedi (1982) discussed problem of estimation using single auxiliary variate whereas Khan and Tripathi (1967), Rao (1975) and Singh and Namjoshi (1988) considered the use of multi auxiliary variates in double sampling. Singh (1967) used information on two auxiliary variates and envisaged a ratio-cum-product estimator of finite population mean of the study variate assuming that the population mean of the auxiliary variates are known. Upadhyaya and Singh (1999) proposed some ratio type estima...

This paper is an attempt to develop an estimator for finite population mean. Motivated by Kiregyera (1984), a ratio in ratio type exponential strategy is developed for estimation of population mean in double sampling for stratification. To compare with relevant considered estimators, expressions for bias and mean squared error of the developed estimator have been derived. The developed estimator has been compared with usual unbiased estimator, Ige and Tripathi (1987), ratio estimator and ratio type exponential estimator given by Tailor et al (2014) theoretically as well as empirically.

A ratio estimator is proposed for the ratio of two population means using auxiliary information in stratified random sampling. Bias and mean squared error expressions are obtained under large sample approximation, and the proposed estimator is compared both theoretically and empirically with the conventional estimator of ratio for two population means in stratified random sampling.

proposed a general family of estimators for population mean using known value of some population parameters in simple random sampling. The objective of this paper is to propose a family of combined-type estimators in stratified random sampling adapting the family of estimators proposed by under non-response. The properties of proposed family have been discussed. We have also obtained the expressions for optimum sample sizes of the strata in respect to cost of the survey. Results are also supported by numerical analysis. * st y under non-response.