Sampling is one of the most complex issues when dealing with statistics. A
sample is a portion of the population that is selected for analysis. The worst
mistake researchers make is to inappropriately collect data. Data may be
collected through random sampling, convenience sampling, systematic sampling,
stratified sampling, and cluster sampling. The decision to use random sampling
or some non-probability sampling method depends on the research design chosen.
No matter how well you have planned your collection process, there is always
likely to be some error in the results. A sampling error is the difference
between a sample result and the true population results. Sampling error arises
from sample fluctuations.
A non-sampling error occurs when the sample data are incorrectly collected,
recorded, or analyzed. This can be result of intentionally or unintentionally
selecting a biased sample, using defective measurement instruments, or entering
data incorrectly into database management programs.
After choosing a sampling method, researchers turn to the next important
question: how many subjects should be used? You can approach this question with
statistical significance, confidence intervals, or ‘on the fly.’
Of course, if you want your results to withstand rigorous scientific inquiry,
it is best to use statistical significance and confidence intervals to
determine your sample size. Even if your population size is infinite, our
consultants are experts at ‘power statistics’ and can employ an equation known
as the Sample Size Determination for a Proportion to determine any size of a
sample given a desired confidence level (i.e. 385 subjects would be necessary
to represent an unknown population size to a 95% confidence interval).
Many researchers possess limited resources such as time and money. For this
reason, knowing the right size of a sample can be a huge cost-effective and
timesaving technique to improve your research work. Please contact a
StatsBusters representative for further information.
© 2006 StatsBusters, LLC. StatsBusters is an American and British cooperative.
The United States and United Kingdom member firms are separate and independent
legal entities and each describes itself as such. All rights reserved.
StatsBusters Online Privacy Statement and Disclaimer
