10.2 Sampling & Extent of Inspection

Sampling — Not 100%

A fundamental difference between initial verification and periodic inspection is the extent of testing. Initial verification tests 100% of new work — every circuit is dead-tested (continuity, insulation resistance, polarity) then live-tested (Zs, RCD). Periodic inspection, by contrast, is carried out on a sample basis: a representative percentage of the installation's accessories and circuits is inspected and tested, and the results are taken as indicative of the overall condition.

The sample percentage is agreed with the client beforehand and follows GN3 guidance. A typical starting sample might be a percentage of the accessories on each circuit type — for example, examining a proportion of socket-outlets on a ring final circuit rather than every single one. The principle is that a properly chosen sample gives a reasonable assessment of the overall condition without the cost and disruption of testing everything.

Why Sampling Is Used

Fully testing a large existing installation — often occupied and operational — is frequently impractical. The main reasons sampling is used:

• Cost and time: 100% testing of a commercial or industrial installation would be prohibitively expensive and disruptive to the occupier.
• Occupied premises: isolation of every circuit is often impossible where essential loads must remain energised (cold stores, IT equipment, medical areas).
• Representative condition: a properly chosen sample gives a reasonable assessment of overall condition — if the sample is sound, it is reasonable to infer that the rest is sound.

The trade-off is certainty: sampling gives a high, but not absolute, confidence. That is why the sample is extended when defects appear, and why the report records the sampling approach rather than pretending 100% was achieved.

Extending the Sample

If the sample reveals defects, deterioration, or non-compliance, the inspector increases the sample size (extends the inspection) to determine how widespread the problem is. For example, if the first few socket-outlets tested on a ring final circuit show high earth-fault loop impedance, the inspector tests more — possibly all — outlets on that ring to map the extent of the fault.

The increased sampling is recorded on the EICR, along with the findings and the reason for the extension. This is important: the EICR must reflect not only the original sample but any extension and why it was needed. An installation that appears sound on the initial sample may, on extension, reveal systematic deterioration — and the outcome reflects the extended findings, not just the first sample.

GN3 gives guidance on when to extend: broadly, where the sample shows a pattern of defect or deterioration rather than an isolated item, the inspector should widen the sample until the extent is established.

Testing on a Sample

The standard tests — continuity of protective conductors, insulation resistance, polarity, earth-fault loop impedance (Zs), and RCD operation — are applied on a sample of each circuit type. One hundred per cent testing is not always possible on live, occupied installations.

A further difference from initial verification is the test sequence. Periodic inspection often begins with live tests (Zs, RCD operation) to confirm the installation is safe to work on, before isolating for dead tests (continuity, insulation resistance). Initial verification always starts dead (the installation is new and not yet energised) and ends with live tests. This reversal reflects the reality of working on an existing, energised installation: the inspector must first establish that it is safe to isolate and handle.

Recording the Approach

The EICR records:

• the extent of the inspection (what was included);
• the sample percentage applied to each circuit type;
• any extension of the sample and the reason;
• the limitations (agreed and operational, as in section 10.1).

This transparency lets the reader judge the weight to place on the report. A report stating a 20% sample with no extension and no defects carries a different weight from one stating a 20% initial sample extended to 100% after defects were found.

GN3 Sampling Guidance in Practice

GN3 suggests a percentage-based sample as a starting point — for example, inspecting a proportion of each accessory type on a circuit. A typical approach on a domestic ring final circuit might be to test every socket-outlet for polarity and Zs, but to inspect the terminations (remove accessories for visual examination) on a smaller percentage, say 25%, unless the sample reveals defects. The exact percentage is the inspector's professional judgement informed by the installation's age, usage, and visible condition.

A worked example: an inspector surveys a small commercial unit with eight lighting circuits and six socket circuits. The agreed sample is 25% of accessories per circuit. On lighting circuit 3, two of the four sampled luminaires show heat-damaged terminations. The inspector extends the sample on that circuit to 100% and finds three more damaged terminations. The EICR records the original 25% sample, the extension to 100% on circuit 3, the five C2 observations, and the fact that the other circuits remained at 25% with no defects. The reader can then see exactly what was found and where the certainty lies.

The key principle: sampling is not a licence to miss defects — it is a structured way to find them efficiently, with a built-in escalation when they appear.