Page 1 of 2
Sample Size Calculation when the Sampling Method is
Simple Random Sampling (SRS)
The following summary addresses two-tailed tests.
Sample Size Calculation for SRS when Estimating Proportions
• When investigating “proportions”, the question being asked must have a "yes/no" or
other two-choice answer, leading to a proportion of the population.
• Examples:
o We may wish to estimate the proportion of public secondary educators who will
retire within the next 5 years.
o We may wish to estimate the proportion of students who receive a passing score
on a statewide assessment.
o We may wish to estimate the proportion of elementary teachers who are male.
o We may wish to estimate the proportion of high school graduates who have
completed an online course as part of their high school program of study.
o We may wish to estimate the proportion of graduates from a doctoral program
who obtain employment as university faculty.
o We may wish to estimate the proportion of public school teachers who will
change their teaching style to accommodate new educational initiatives, such as
the Common Core State Standards.
• ss =
(
)
()
http://www.surveysystem.com/sample-size-formula.htm
Where z = z value for the desired confidence interval:
• 90% / 1.645
• 95% / 1.960
• 99% / 2.575
• 99.9% / 3.29
p = percentage picking a choice, expressed as decimal
• Use 0.5 for sample size calculation when proportion in population is unknown.
• If previous research has shown a valid “p”, then that value may be used in place of 0.5.
c = confidence interval, expressed as decimal = B = bound
• Determined by researcher (e.g., 0.03 = ±.03)
• B = bound
o The bound is the distance from the mean out to either the upper limit or lower
limit of the confidence interval. These limits are referred to as the upper or lower
bounds. The bound is the figure that follows the ± sign.
Sample Size Correction for Finite Population (first calculate “ss”, then use correction).
• ss
fp
= ss/(1+ (ss-1)/pop)
Hypothetical Example: Researchers from a particular state wish to determine the proportion of
campers using national forest campgrounds whose camping stay is 5 days or longer.
Approximately 99,000 campers use the campgrounds each camping season. Previous studies
have shown 43% of camp stays are 5 days or longer. Determine the sample size that researchers
must obtain given a desired bound (B) of .02 with a 95% probability.
Answer = 2,299.3 then always ROUND UP for sample size = 2,300.