9.4 Invalid Results and Retest Decisions

Treat Invalid Results as Quality Events

A field test can be performed carefully and still produce an invalid result. Distinguish two situations the exam tests hard:

• Invalid test — the procedure or sample was compromised (blown 5-minute window, leaking air meter, dropped slump cone, segregated sample). The number does not represent the concrete and must be discarded and the test repeated when possible.
• Out-of-spec result — a valid test shows the concrete is outside the specification limit (slump too high, air too low). The number is real and must be reported and escalated, not erased.

Common causes of invalid results include: a non-representative sample (drawn from the first/last 10 percent, or not remixed); equipment faults (uncalibrated air meter, damaged slump cone, leaking pressure gauge); procedural error (wrong rod count, wrong layer count, improper strike-off); and timing failures where the test began outside the C172 window.

The ASTM C94 Single-Retest Rule

When a fresh-property test fails, the technician does not condemn the load on one reading. ASTM C94 permits a single retest: if the result of a slump or air-content test falls outside the specified tolerance, a second test is run on a new portion of the same sample. The concrete is considered to fail only if the second test also fails — and many specifications evaluate the failure based on the second test, not an average that masks a bad batch. The key exam point is that one failing fresh test is not automatic rejection; a confirming retest is required.

If the retest also fails, the concrete is reported as nonconforming and the decision passes up the QA/QC chain.

Water Additions and Never "Fixing" the Number

A tempting but wrong response to low slump is unauthorized water. ASTM C94 limits jobsite water: water may be added only up to the maximum design water-cementitious ratio, the addition must be mixed at least 30 revolutions at mixing speed, all water additions must be completed within the discharge time limit, and the fresh tests must be repeated after the addition. Every gallon must appear on the ticket. Adding water beyond the design w/cm to chase a slump number is a quality failure that weakens the concrete.

The cardinal rule is that a questionable reading is handled through the applicable ASTM method and the project communication chain, never by editing the number. If the air meter leaks, the result is invalid and the meter is repaired or replaced before retesting — the technician does not nudge the gauge to a plausible value. Documenting why a result was invalid (and what was done about it) is itself an act of quality control, because it protects everyone who relies on the record.

A Decision Tree for a Questionable Result

When a number looks wrong, the technician works through a disciplined sequence rather than reacting emotionally:

1. Was the sample representative? Drawn from the middle of the batch, remixed, not from the first/last 10 percent? If not, the result is invalid — obtain a new portion or sample.
2. Was the equipment sound and calibrated? A leaking air meter, dented cone, or out-of-tolerance scale invalidates the result. Repair or replace, then retest.
3. Was the procedure correct? Right number of layers and rod strokes, proper strike-off, correct lift time on the slump cone, test begun inside the C172 window? A procedural slip means repeat.
4. If sample, equipment, and procedure were all sound, the result is VALID — and if it is outside spec, it is an out-of-spec finding to report, applying the C94 single-retest rule before any nonconformance is declared.

Out-of-Spec Versus Invalid: A Critical Distinction

Mixing up these two leads to bad outcomes. " Treating an invalid result as out-of-spec unfairly condemns good concrete and triggers needless investigation. The exam tests whether the candidate can tell them apart from the cause: a blown time window, contamination, or broken equipment makes a test invalid; a clean test that exceeds the slump or air limit is out-of-spec but real. In both cases the data and the response are documented, but the path differs — discard and repeat for invalid, report and escalate (after one retest) for out-of-spec.

Never resolve either by quietly editing the number; that converts a recoverable quality event into falsified records.

The single-retest provision exists precisely so a technician does not have to choose between two bad options — accepting a marginal first reading or fabricating a better one. When slump or air fails, the disciplined move is simply to draw a fresh portion and run the test again, then let the two valid results speak. This protects the technician as much as the project: a documented retest sequence is defensible, whereas an unexplained adjusted number is not, and it keeps a single anomalous reading from either condemning sound concrete or excusing a genuinely defective batch.