China
Africa
Explore chapters and articles related to this topic
Sampling Theory
Published in Marcello Pagano, Kimberlee Gauvreau, Heather Mattie, Principles of Biostatistics, 2022
Marcello Pagano, Kimberlee Gauvreau, Heather Mattie
What all these surveys have in common is that they utilize probability samples; whether a particular person is included in the sample is decided by a random device. Probability samples allow us to not only make inference about an important characteristic of a population, but also to calculate a measure of uncertainty associated with that inference. The simple random sample described below is one example of such a sampling scheme, but there are other ways to obtain a probability sample. Each has its own advantages and limitations, and we discuss some of these in this chapter.
Sampling
Published in Louis Cohen, Lawrence Manion, Keith Morrison, Research Methods in Education, 2017
Louis Cohen, Lawrence Manion, Keith Morrison
A probability sample, because it draws randomly from the wider population, is useful if the researcher wishes to be able to make generalizations, because it seeks representativeness of the wider population. (It also permits many statistical tests to be conducted with quantitative data.) This is a form of sampling used in randomized controlled trials. Randomization has two stages – random selection from a population and random allocation to groups (e.g. a control and an experimental group) – and these are key requirements for many experiments and statistics. Randomization, as one of its founding figures, Ronald Fisher (1966), remarked, is designed to overcome myriad within-group and between-group differences. It ensures that the average result, taking into account range and spread, within one group is similar to the average within another group (Torgerson and Torgerson, 2008, p. 29); as the authors remark, ‘[t]he presence of all variables that could affect outcome … in all groups will cancel out their effect in the analysis’ (p. 29), and if, by chance, other variables are not the same in both groups, then this is unlikely to affect the outcome. Indeed Fisher commented that randomization, intended to overcome individual differences, is sufficient ‘to guarantee the validity of the test of significance’ in an experiment (1966, p. 21). Randomization has the potential to address external validity, i.e. generalizability, and internal validity, i.e. to avoid selection bias (p. 29).
Validating a measurement of psychological, physical and sexual abuse against women in gynecological care within the Chilean health system
Published in Health Care for Women International, 2022
Manuel Cárdenas Castro, Stella Salinero Rates
The instrument that we have created must continue accumulating evidence that validates its use in investigative contexts. Similarly, future studies must overcome certain limitations that are present in this article. The first is the use of a non-probability sample, which makes it difficult to generalize the results unto the general population. On the other hand, the fit of the alternative models made with the item response technique should be examined in order to detect overlapping and possible continuum portions that are not covered by the items. The same way, we hope to begin studies that allow this scale to be tested in other groups that also require access to adequate and timely gynecological care, such as transgender people, who most obviously suffer from the rejection of a medical system that is based on a binary model of gender (with this rejection being most evident in our particular context, gynecological and obstetric medical specialties).
Sexual Health Among Lesbian, Gay, Bisexual, and Heterosexual Older Adults: An Exploratory Analysis
Published in Clinical Gerontologist, 2021
Mark Brennan-Ing, Jennifer E. Kaufman, Britta Larson, Kristi E. Gamarel, Liz Seidel, Stephen E. Karpiak
Comparing data between a community sample and a national probability sample has limitations. We cannot rule out the possibility that differences in participant recruitment strategies, data collection methods, other aspects of research protocols, and the representativeness of these samples may explain some of the differences we observed related to sexual health and sexual orientation. However, our ability to use identical questions for comparison is a notable strength. Furthermore, given the lack of nationally representative samples of older LGB adults, and the dearth of information on their sexual health, our findings help to advance the scientific literature on this topic. An additional limitation was the inability to examine bisexuals separately from lesbians and gay men, but this breakdown was not available in the NSHAP-LGB data, so we were unable to make those comparisons.
Associations between body shape, body adiposity and other indices: a case study of hypertension in Chinese children and adolescents
Published in Annals of Human Biology, 2019
Data collected from the 2011 wave of the China Health and Nutrition Survey (CHNS) were used in this study. The CHNS is an on-going open-cohort study, across a large part of China, which started in the year 1989 and is followed up every 3 or 4 years. Multistage cluster random sampling was applied to draw a stratified probability sample from each selected province in the CHNS (Popkin et al. 2010; Zhang et al. 2014). Within each province, two cities and four counties were selected based on their income levels reported by the State Statistical Bureau in 1988 (Du et al. 2014). In each city or county, four communities were selected and within each community 20 households were selected, from which all household members were interviewed (Du et al. 2014). In 2011, new participants with different geographical situations and health indicators were selected from nine provinces and the three largest municipal cities in China. By 2011, the CHNS had included 47% of China’s population (Zhang et al. 2014). Detailed information on the study design and characteristics of the CHNS has been published elsewhere (Liu 2008; Zhang et al. 2014). In the present study we included 863 children (aged 7–12 years) and 524 adolescents (aged 13–17 years) with valid information on their blood pressure, anthropometric measurements (weight, height, WC and hip circumference), age and gender (Figure 1).