A sample can be measured carefully yet fail to represent the lot it is meant to describe. If only the easiest boards to reach can be selected, parts of the delivery may never enter the sample. Define the available population and selection method before inspection begins.
Statistics Canada’s sample-selection guidance distinguishes the target population, observed population, and sampling frame. The same distinction is useful in manufacturing: the boards you want to characterize, the boards available for selection, and the list or physical structure used to select them may differ.
Describe the lot precisely
Identify the delivery, product revision, production lot, and any known subdivisions such as panels, cartons, or manufacturing dates. State whether the objective concerns all received boards or only a defined subset. A mixed delivery should not become one homogeneous lot merely because it arrived in one box.
Build a selection frame that includes every eligible unit at the intended level. If selection begins with panels and then chooses positions within panels, record both stages. A board selected through a panel is not necessarily sampled in the same way as a board selected from a complete individual-unit list.
Look for coverage gaps
| Situation | Question |
|---|---|
| Only top trays are accessible | Can lower trays enter the sample? |
| Several panel positions exist | Does selection cover the relevant positions? |
| Some cartons are held elsewhere | Are they inside the claimed population? |
| Damaged units are set aside | Are exclusions recorded and justified? |
Choose and document the selection method before observing the inspection outcome. Replacing an inconvenient or suspicious selected board with a cleaner-looking one changes the sample. Record any unavoidable substitution and its reason, retaining the identity of the originally selected board.
Preserve the selection record
- Freeze the eligible population definition.
- Record the sampling frame or physical selection procedure.
- Identify selected units and their locations.
- Track exclusions, missing units, and replacements.
- Limit conclusions to the population actually represented.
This review does not prescribe a sample size or acceptance threshold. Those belong to the agreed inspection plan and its statistical purpose. A larger sample does not automatically repair systematic exclusion of one carton, panel region, or production period.
Keep the selection evidence with the inspection results. It should be possible to explain how each measured board entered the sample and which units had no opportunity to enter. That boundary is essential when using the findings to accept a lot, compare suppliers, or investigate a localized manufacturing issue.
Move from reading to a build-ready project
Use the fabrication and assembly checklists to prepare drawings, production files and test requirements.
Open the DFM checklist