Optional randomized response in stratified unequal probability sampling—A simulation based numerical study with Kuk’s method

Test ◽  
2007 ◽  
Vol 16 (2) ◽  
pp. 346-354 ◽  
Author(s):  
Amitava Saha
Keyword(s):  
Numerical Study ◽  
Randomized Response ◽  
Probability Sampling ◽  
Unequal Probability Sampling ◽  
Unequal Probability ◽  
Simulation Based

Extending Takahasi‐Sakasegawa’s Indirect Response Technique to Cover Sensitive Surveys in Unequal Probability Sampling

2007 ◽  
Vol 59 (3-4) ◽  
pp. 265-276 ◽  
Author(s):  
Sanghamitra Pal
Keyword(s):  
Random Sampling ◽  
Simple Random Sampling ◽  
Subject Classification ◽  
Randomized Response ◽  
Probability Sampling ◽  
Indirect Response ◽  
Variance Estimators ◽  
Unequal Probability Sampling ◽  
Unequal Probability ◽  
Sensitive Characteristics

Abstract: Randomized Response (RR) Technique (RRT), introduced by Warner (1965), is a well‐known way to unbiasedly estimate proportions of people bearing sensitive characteristics. Takahasi and Sakasegawa (1977) narrated a novel procedure avoiding any particular RR device unlike most researchers in this field. Most RRT’s in the literature give estimators for population totals and variance estimators thereof, allowing exclusively simple random sampling with replacement (SRSWR) schemes, including as well the above too. We present formulae applicable to general varying probability sampling, even without replacement applying Takahasi‐ Sakasegawa device. AMS (2000) Subject Classification: 62D05.

Evaluation of sampling alternatives to quantify tree leaf area

2003 ◽  
Vol 33 (1) ◽  
pp. 82-95 ◽  
Author(s):  
H Temesgen
Keyword(s):  
Leaf Area ◽  
Poor Performance ◽  
Tree Crown ◽  
Two Stage ◽  
Zero Bias ◽  
Probability Sampling ◽  
First Order ◽  
Unequal Probability Sampling ◽  
Unequal Probability ◽  
Tree Leaf

High within- and among-tree crown variation have contributed to the difficulty of tree-crown sampling and single-tree leaf area (area available for photosynthesis) estimation. Using reconstructed trees, simulations were used to compare five sampling designs for bias, mean square error (MSE), and distribution of the estimates. All sampling designs showed nearly zero bias. For most sample trees, stratified random sampling resulted in the lowest MSE values, followed by ellipsoidal, two-stage systematic, simple random, and then by two-stage unequal probability sampling. The poor performance of two-stage unequal probability sampling can be ascribed to the unequal probability of inclusion of first-order branches and twigs.

Sign in / Sign up

Export Citation Format

Share Document