Estimation of the Coefficient of Variation with Minimum Risk: A Sequential Method for Minimizing Sampling Error and...

Area of Science:

• Psychometrics
• Statistical Methods

Background:

• The coefficient of variation is a valuable effect size measure in psychology.
• Current fixed sample size methods cannot simultaneously minimize sampling error and study costs.
• A need exists for efficient estimation procedures that balance precision and resource allocation.

Purpose of the Study:

• To develop a general theory for sequential estimation of the population coefficient of variation.
• To create a procedure that considers both sampling error and study cost without distributional assumptions.
• To provide a method that optimizes resource allocation for accurate estimation.

Main Methods:

• A sequential sampling procedure for estimating the coefficient of variation was developed.
• The method involves planning a pilot sample size and sequentially collecting data until a stopping rule is met.
• A risk function guides the sequential data collection to balance error and cost.

Main Results:

• The proposed sequential procedure is the first of its kind for coefficient of variation estimation.
• The method effectively balances sampling error and study costs, preventing unnecessary expenses.
• Monte Carlo simulations demonstrated the properties of the final sample size distribution across various conditions and distributions.

Conclusions:

• The developed sequential estimation method offers a more efficient approach to estimating the coefficient of variation in psychological research.
• This procedure allows researchers to achieve desired precision while controlling study costs.
• Freely available R functions in the MBESS package facilitate the implementation of these methods.