How to separate instrument, method, operator, environment, and product variation when two calipers disagree.
Start by making the measurements comparable
A difference between two caliper readings does not by itself show that the product is defective or that one caliper is inaccurate. First confirm that both readings refer to the same feature, measurement point, orientation, unit, and surface condition. A small shift in position can matter when a part is tapered, curved, flexible, burred, coated, or otherwise nonuniform.
The contact method must also match. Jaw alignment, contact force, closing speed, cleanliness, and whether the caliper is held square to the feature can change the indication. The operators should use one defined procedure before comparing the instruments.
Separate instrument effects from technique
Inspect both calipers for zero error, damaged or dirty jaws, smooth movement, suitable range, adequate resolution, and relevant calibration information. Then use repeated measurements in a controlled sequence. It is helpful for each operator to measure the same point with both instruments and, where appropriate, repeat the sequence in a different order.
Patterns are more informative than one pair of readings. If one instrument remains consistently offset across repetitions and operators, an instrument-related cause becomes more plausible. If readings change mainly with the operator or contact method, technique may be influential. If both instruments vary similarly by position, the feature itself may not be uniform. These are diagnostic indications, not final proof.
Draw only the conclusion supported by records
Record the individual readings rather than only their averages. Keep the instrument identifiers, operators, conditions, measurement-point definition, and sequence attached to the data. Any comparison should also consider the instruments' stated uncertainty and resolution.
A single disagreement cannot identify the source of the difference. A product-specific conclusion requires a defined measurement-point diagram, information for both instruments, and repeated comparison records linked to the relevant SKU or sample. Without those records, no SEWARD product deviation or instrument failure should be asserted.
Should the reading with more decimal places be preferred?
Not automatically. Additional display digits indicate finer resolution, but the method, instrument condition, uncertainty, and repeatability determine whether those digits are meaningful.