• E-mail
• Print
• RSS

Tip: Simple vs. Stratified Sampling

Healthcare Auditing Weekly, November 6, 2007

When determining whether to choose simple random sampling or stratified random sampling, keep the following benefits and drawbacks in mind:

Simple random sampling

Benefits:

• It's quick and easy.

Drawbacks:

• Few populations have similar items, like hospital to hospital, department to department, or year to year, so conclusions about populations based on simple random sampling are not as reliable as conclusions based on methods that better "fit" the population. For example, stratify the population among records that are likely to have a high error rate and those that are likely to have a low error rate. Doing so will give you a more reliable conclusion about the population.

Stratified random sampling Benefits:

• Results are more reliable if the population has widely disparate characteristics-such as numerous low-,middle-, and high-dollar value items, or there are several record locations for the population
• It can maximize the dollars protected. Reviewers, for example, can review a large portion of the total dollars in a population without having to use a large sample.
• It increases the chance of reviewing error-prone items if these items are included in one stratum and relatively few items are in another. For example, when you know that new residents' medical record documentation will be more error-prone than second- or third-year residents' medical records, sample more heavily in the error-prone stratum.
• It allows for the development of separate estimates for individual strata, if necessary.

Drawbacks:

• The increased precision in attribute sampling is usually not worth the additional work required in developing strata and samples for each stratum.

• E-mail to a Colleague
• Print
• RSS
• Archive