Examples of using Sampling ratio in English and their translations into Russian
{-}
-
Official
-
Colloquial
Sampling ratio= 10.
The weights here are generally the reciprocal of sampling ratios, Nk/nk.
Sampling ratio 10.
The median CV per cent increases twofold when the sampling ratio is 10 per cent.
Sampling ratios, and thus weights, must be taken into account to compute valid significance tests for sample data.
At the seminar it was discussed which minimal sampling ratio is necessary for getting significant results.
In complex sampling designs, e.g. stratified orclustered sampling, sampling ratios vary.
The final sampling ratio is the product of successive levels of sampling ratios, and the same for their reciprocals.
It's easy to see gains of quality increasing the sampling ratio and higher quality for larger areas.
Table 5 describes the percentage distributions of the estimates p by classes of CV for each tested sampling ratio.
In stratified and clustered samples, sampling ratios may differ from one part of the sample to another, for different reasons.
If weights are discarded, and significance tests applied on the unweighted data file,the tests may give wrong results when sampling ratios vary between groups of cases.
The intention is to choose a sampling ratio of 33 per cent with the aim to preserve the richness of the census information as much as possible.
In table 1 the median CV per cent among the areas is showed for different classes of estimated p and for each of four tested sampling ratio.
When a sampling ratio of 33 per cent is adopted, percentages p between 0,5 per cent and 1 per cent can be estimated with a median CV per cent of about 16 per cent.
The main result is that hypercubes with more than 20,000 cells can be estimated with a low percentage of units in critical cells even for lower sampling ratio.
For this it is essential that the dilution ratio or the sampling ratio rd or rs be determined such that the accuracy requirements of paragraph 9.4.4.
These weights(Eq. 4) accomplish the two functions at once: they adjust sample figures to population scale, andthey also correct imbalances in sampling ratios from one stratum to the next.
For this it is essential that the dilution ratio or the sampling ratio rd or rs be determined such that the accuracy requirements of paragraph 9.4.6.2. are fulfilled. 9.4.6.1.
The existence of replacements as well as unplanned changes in sample size may involve assumptions and decisions, which are to some degree arbitrary, affecting which sample size(planned or actual)is to be used to compute sampling ratios and thus to compute weights, as will be explained below.
In the case of severe budget constraints, the sampling ratio could be reduced considering a trade-off between needed financial savings and accuracy required at different territorial domains.
In general, the gain of efficiency measured as relative difference of CVs is about 33- 38 per cent when the sampling ratio goes from 10 to 20 per cent and about 53-58 per cent when the sampling ratio increases from 10 to 33 per cent.
It is possible to observe that for 33 per cent sampling ratio in seven of the eight hypercubes considered, the percentage of individuals classified in critical cells is less than 5 per cent for all the areas, while for the largest hypercube H.B1.E1. R4 the indicator is less than 5 per cent in 15 areas and the value is between 5 and 10 per cent in the remaining 5 areas.
Correcting for this is done with proportional weights,which are specific to every subdivision of the sample for which the sampling ratio is homogeneous(e.g. each stratum), and can be designated as kπ with the general form πk=% of stratum in population/% of stratum in the sample. .
Relating to this hypercube,Table 4 shows for each tested sampling ratio(adopting the simple random sampling of households from population registers) and for three Italian NUTS2 areas with different size(Molise is one of the smaller, Marche has a medium size and Sicilia is one of the larger) the critical thresholds, the percentages of critical cells and the percentages of individuals in critical cells.
For this it is essential that the dilution ratio or the sampling ratio rdil or rs be determined such that the accuracy limits of paragraph 3.2.1. of Appendix 5 to this Annex are fulfilled.
For instance, in case of 10 per cent sampling ratio, the average CVs fall under 10 per cent whencounts go over 1,000; when sampling ratio increases up to 33 per cent the threshold decreases to 250.
In order totake the simulation study manageable, each sampling ratio was tested on 10 municipalities, given that 5 of them were common for all the tests and the others were chosen in the same class of population.
For instance, considering Sicilia area, in case of 33 per cent sampling ratio, the related critical threshold is 100, 59.4 per cent of cells have absolute frequencies lower than 100 but in these cells only 1.0 per cent of eligible people is classified.
Rs average sample ratio over the test cycle.