Using t, normal, Z-to-DPMO, and p-value reference tools

Learning objectives

• Use t, normal, Z-to-DPMO and p-value reference functions in a free spreadsheet.
• Compute DPMO from garment defect data.
• Convert between short-term and long-term Z with an explicit assumption.

Core idea

Reference functions map standardized statistics to tail areas and critical values. DPMO normalises defects against opportunities and can be translated to a Z score.

Spreadsheet lab

• Two-tailed t p-value: =T.DIST.2T(ABS(t), df).
• Two-tailed t critical: =T.INV.2T(alpha, df).
• Standard normal tail: =1 - NORM.S.DIST(z, TRUE).
• Standard normal critical: =NORM.S.INV(1 - alpha/2).
• DPMO = defects / (units × opportunities per unit) × 1,000,000.

Garment-factory example

A finishing line records 45 defects across 500 units with 8 opportunities per unit: DPMO = 45 / (500 × 8) × 1,000,000 = 11,250. Convert to Z_ST = =NORM.S.INV(1 − DPMO/1,000,000) if the opportunity model is honestly defined; state whether any 1.5-shift adjustment is applied.

Method

1. Define opportunities consistently across products.
2. Choose reference function that matches distribution and tail.
3. Report short-term vs long-term Z explicitly.

Common mistakes

• Applying a 1.5σ shift automatically.
• Comparing DPMOs from unlike opportunity models.
• Using one-tailed tables for a two-tailed question.