Efficient balanced sampling: The cube method

Deville, Jean-Claude; Tillé, Yves

doi:10.1093/biomet/91.4.893

Cited by 204 publications

(198 citation statements)

References 20 publications

Supporting

Mentioning

193

Contrasting

Unclassified

Order By: Relevance

“…When a rounding problem is encountered, it is not possible to find a vector s of zeros and ones that exactly satisfies the equation Deville and Tillé (2004) proposed the cube method, which allows for the selection of balanced samples. The cube method is an algorithm composed of two phases: the flight phase and the landing phase.…”

Section: Balanced Samplingmentioning

confidence: 99%

“…At the end of the landing phase, every unit has been selected or rejected. Deville and Tillé (2004) have proposed two ways of running the landing phase: by linear programming or by suppression of variables. The landing phase by linear programming consists of solving a linear programming problem through the simplex algorithm.…”

Section: Balanced Samplingmentioning

confidence: 99%

“…These auxiliary variables are called the balancing variables. Deville et al (1988) described a method to obtain balanced samples and, later, the cube method (Deville and Tillé, 2004) was proposed for the same purpose. Some methods have been proposed for the computing of optimal inclusion probabilities for balanced sampling as for instance those given in Tillé and Favre (2005), Nedyalkova and Tillé (2008), and Chauvet et al (2011a).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Fast balanced sampling for highly stratified population

Hasler

Tillé

2014

Computational Statistics & Data Analysis

View full text Add to dashboard Cite

a b s t r a c tBalanced sampling is a very efficient sampling design when the variable of interest is correlated to the auxiliary variables on which the sample is balanced. A procedure to select balanced samples in a stratified population has previously been proposed. Unfortunately, this procedure becomes very slow as the number of strata increases and it even fails to select samples for some large numbers of strata. A new algorithm to select balanced samples in a stratified population is proposed. This new procedure is much faster than the existing one when the number of strata is large. Furthermore, this new procedure makes it possible to select samples for some large numbers of strata, which was impossible with the existing method. Balanced sampling can then be applied on a highly stratified population when only a few units are selected in each stratum. Finally, this algorithm turns out to be valuable for many applications as, for instance, for the handling of nonresponse.

show abstract

Section: Balanced Samplingmentioning

confidence: 99%

Section: Balanced Samplingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fast balanced sampling for highly stratified population

Hasler

Tillé

2014

Computational Statistics & Data Analysis

View full text Add to dashboard Cite

show abstract

“…This method is divided into a ight phase and a landing phase (see Deville and TillÃ e, 2002). Each phase can be applied with several options.…”

Section: Measure Of Accuracy Of the Approximations Of Variancementioning

confidence: 99%

Variance approximation under balanced sampling

Deville

Tillé

2005

Journal of Statistical Planning and Inference

View full text Add to dashboard Cite

A balanced sam pling design has the interesting property that Horvitz-Thom pson estimators of totals for a set of balancing variables are equal to the totals we want to estim ate, therefore the variance of Horvitz-Thom pson estim ators of variables of interest are reduced in function of their correlations with the balancing variables. Since it is hard to derive an analytic expression for the joint inclusion probabilities, we derive a general approxim ation of variance based on a residual technique. This approximation is useful even in the particular case of unequal probability sampling with ÿxed sample size. Finally, a set of numerical studies with an original methodology allows to validate this approximation.

show abstract

“…A family of algorithms is available (Deville and Tillé, 2005a) to select a balanced random sample. It is thus possible to select at time t a sample, so that the identities…”

Section: Balanced Samplingmentioning

confidence: 99%