could you please provide me "Watson"s" formula that helped to...

Asked February 19, 2014, 5:13 AM EST

could you please provide me "Watson"s" formula that helped to determine sample size out of a given population.

Outside United States

1 Response

I am not sure who "Watson" is, but determining sample sizes typically involves the following steps:
1) What research and null hypotheses will be tested using this sample?
2) What is the expected difference in responses between groups? If a current therapy/intervention has a cure rate of 40% and your new therapy/intervention has a proposed cure rate of 60%, then the difference in improvement between the therapies is 20%.Another way to address this is to specify the effect size that is of scientific interest. With a large enough sample, every effect is statistically significant. Determine an effect that has the most “real world” significance.
3
) What is the standard error/standard deviation in measurement? This information is usually obtained from pilot studies or historical data that has a similar outcome. If you have the standard error, you can calculate the standard deviation (and vice versa).
4) What is the level of statistical significance? This is the probability of falsely concluding the therapies/interventions significantly differ (typically expressed as 0.05).
5
) What is the intended power calculation? This is the probability of correctly concluding therapies/interventions significantly differ (typically expressed as 0.8 or 0.9).
As a general rule of thumb, a random sample of over 384 individuals is enough to represent a population of 10,000.
Additional resources:
http://www.praccreditation.org/secure/documents/coachHO16.pdf
http://www.surveysystem.com/sscalc.htm
http://bphc.hrsa.gov/policiesregulations/performancemeasures/patientsurvey/calculating.html