![[logo] US EPA](../gif/logo_epaseal.gif){kind=logo}
Determining Sample Sizes
Explicit description of objectives and criteria are required establishing sample sizes. In addition, information on known or suspected implementation issues (frame errors, access, etc.) can be used to increase the likelihood of successful implementation. Also see, frequently asked questions on determining sample size and why 50 samples.
Determination of sample size requires knowledge and consideration of:
- monitoring objectives
- desired levels of certainty in results
- monitoring program duration as well as funding and logistics.
Establishing sample size(s) requires detailed information on:
- all the estimates that will be produced from the survey
- the precision desired for each estimate
- knowledge of the variability expected
- Desired amount of detectable change
A reality faced in most studies is:
- The number of objectives creates a need for many more sample sites than budget and operational constraints allow
- Determination of the number of sample sites that can be sampled based on budget, logistics, etc leads to
- The requirement to assign priorities to each objective
- It is usual to have some sub-objectives dropped due to sample size limitations.
Sample size calculations for the estimation of a proportion, e.g., proportion of stream length that meets a designated use depend on:
- Proportion of interest (often approximately 50%)
- Precision required (width of confidence limits, often target at +/- 10%)
- Confidence required (usually 90% or 95%).
Oregon State Example - Precision Goals and Results
Base Samples
Sample Size 50
- 50% of units meet the criteria (i.e., 50% of stream length is impaired), precision will be +/-12% with 90% confidence.
- 20% of the units meet the criteria then the precision will be +/-9% with 90% confidence.
Sample Size 25
- 50% of units meet the criteria, precision will be +/-17% with 90% confidence.
- 20% of units meet the criteria, precision will be +/-13% with 90% confidence.
Sample Size 250 (potential result with 5 year rotating panels)
- 50% of units meet the criteria, precision will be +/-5% with 90% confidence
- 20% of units meet the criteria, precision will be +/-4% with 90% confidence
- 50% of units meet the criteria, precision will be +/-6% with 95% confidence
- 20% of units meet the criteria, precision will be +/-5% with 95% confidence
Over Samples
When known or suspected circumstances are likely to prevent sampling at some base sample sites, e.g., frame errors, denied access, hazardous site conditions, etc. prior additional sample sites can be identified. These Over-Sample sites are sampled whenever a Base-Sample site cannot be sampled. This strategy greatly increases the likelihood of attaining the desired sample sizes and fulfillment of the objectives. Often an additional 10-30% sites are selected as Over-Sample sites. See Implementation Issues for examples from existing programs. Go to Top