Professor Franklin
Accessibility settings

Preferences are saved in this browser when storage is available.

Chapter 10 · Reading 4 of 5

Testing whether a control worked in practice

Security+ SY0-801 · Objectives 5.5

A department says every privileged role change receives independent approval. The policy requires it, the service owner believes it and an interview confirms that staff know the rule. What would let an auditor decide whether the control operated during the semester?

Begin with a claim you can test

Agree on the population, time period, required evidence and exception criteria. “All privileged access is appropriate” is broad. “Privileged grants during the fall term have an approval from an authorized person other than the requester before activation” gives the reviewer a transaction, a sequence and an independence requirement. The audit charter and scope determine which systems and people the reviewer is authorized to examine.

Establish that the population is reasonably complete before sampling it. If the reviewer samples only the ticket queue, emergency grants made directly in the application might never appear. Reconcile a list from the target system with the request records. Protect identities in the working papers and limit collection to the information the review needs.

Read the individual cases

For a selected change, identify the requester, approver, role, target and activation time. A manager approving the request the following morning may demonstrate later review while still missing a prior-approval requirement. An approval recorded before activation may also fail if the approver had no delegated authority for that system. In either case, describe what the evidence supports and ask whether an authorized exception process applied.

Sampling provides evidence within limits. Selecting only recent successful tickets is convenient but can exclude failed, canceled or exceptional work. A small targeted sample helps investigate a suspected weakness. It may offer little basis for estimating a population-wide error rate. Preserve the selection method, sample size and coverage alongside the results so readers know how far the conclusion travels.

Read a finding from the other side of the desk

The reviewer finds two cases without a ticket attachment. Before issuing a final finding, ask the owner where approvals were recorded during the sampling period. A properly controlled second system may contain contemporaneous evidence linked by transaction ID. If it does, document the link and consider whether reconciliation itself is reliable. An email obtained today saying the manager remembers the request is a different kind of evidence.

If approval evidence is still missing, report that specific condition. You may have established a failure to demonstrate the control without proving that the manager never spoke to the requester. The distinction matters to the person whose action is being reviewed and to the manager choosing a repair. Requiring an immutable approval record before activation addresses a different weakness from retraining people on who may approve.

Tell the owner what to do next

Report the requirement, observed condition, potential consequence and responsible owner. Invite factual corrections and keep evidence for disputed observations. Agree on a corrective action and a follow-up test. If the system now prevents self-approval, inspect the configuration and sample new transactions after the change. A completed project ticket shows that work occurred. The follow-up examines whether the control now operates as intended.

This approach applies just as readily to the privacy requests in the next reading. A policy can promise careful handling. A person asking about their records depends on the process that actually operated.

Access review sample

Inspect the three records, then use the evidence to answer the question.

1. Review population

Register lists 40 privileged role changes.

2. Selected cases

Reviewer sampled 8 changes from the specified period.

3. Evidence received

Six sampled cases include the required independent approval.

Apply the reading

What population could a sample of approved tickets accidentally exclude?

Compare your reasoning

Direct grants and emergency changes can exist only in the target system. Reconcile those records before selecting the sample.