Minimum Sample Size (MSS) for Continuous Data

Introduction:  MSS for Continuous Data

The Minimum Sample Size (MSS) is the smallest number of observations required to make a reliable statistical inference about a population. In Lean Six Sigma projects, determining the correct sample size is essential for designing valid experiments, surveys, and measurements that lead to accurate conclusions. The formula discussed here applies to continuous data, such as time, length, weight, or cost.

Background

In practice, too small a sample risks unreliable results, while too large a sample may waste time and resources. Statistical theory provides formulas to calculate the minimum number of data points needed, based on factors such as confidence level, standard deviation, and desired precision. This calculation is particularly valuable in the Measure and Analyse phases of the DMAIC cycle when working with continuous variables.

Key Elements / Features

• Definition: MSS is the minimum number of data points required for a chosen confidence level and margin of error when analysing continuous data.
• Parameters:
  • σ = population standard deviation (or estimated from pilot data).
  • d = desired precision (margin of error).
  • Zα/2 = critical z-value linked to the chosen confidence level (e.g. 1.96 for 95%).

Formula (Continuous Data):

Simplified Version (95%):

Example
Suppose a Lean Six Sigma project wants to measure process time with:

• Estimated standard deviation 𝜎 = 5 minutes
• Desired precision 𝑑 = 2 minutes
• Confidence level = 95% (Zα/2 ≈ 1.96)

Calculation:

So, a minimum of 25 measurements is required.

Relevance / Impact

• Prevents underpowered studies and weak conclusions.
• Saves resources by avoiding unnecessary data collection.
• Increases credibility of Lean Six Sigma projects through statistically valid evidence.
• Supports better decision-making in process improvement.

See also

• Confidence Interval
• Hypothesis Testing
• Central Limit Theorem
• Minimum Sample Size (for Discrete Data)