Why sampling matters

When the IRS audits a return or business, it rarely examines every transaction. Instead, examiners commonly review a sample and use that sample to estimate the rest. That process reduces audit time and cost for both the IRS and taxpayers but can also lead to projected adjustments that materially increase a proposed deficiency if the sample results show errors.

The IRS explains its audit procedures and taxpayer rights in its Audit Process guidance and Publication 556 (Examination of Returns, Appeal Rights, and Claims for Refund) — see IRS Audit Process and IRS Publication 556 for details (IRS.gov).

Types of sampling the IRS (plain language)

  • Probabilistic (statistical) sampling: Every item in the population has a known chance of selection. Common approaches include simple random sampling, stratified sampling and probability‑proportional‑to‑size (PPS or dollar‑unit) sampling. Statistical samples let examiners estimate an overall error rate and attach a measure of reliability (confidence) to that estimate.

  • Pros: Reduces bias, supports defensible projections, allows calculation of confidence limits.

  • Cons: Requires clear population definition, enough records to support sampling, and can still produce large extrapolated adjustments.

  • Non‑probabilistic (judgmental or targeted) sampling: The examiner selects items based on risk indicators or professional judgment — for example, unusually large deductions, a vendor with many related‑party transactions, or items flagged by prior returns.

  • Pros: Focuses on likely problem areas and can be quicker to execute.

  • Cons: Harder to defend statistically if contested and can feel more subjective to taxpayers.

How examiners use sampling in practice

  • Correspondence audits: These are usually limited, item‑specific reviews done by mail. The IRS may ask for records for a targeted set of items (a form of judgmental selection) and use those to verify reported amounts. If you receive a letter, follow the requests precisely and respond by the deadline. For a checklist and example response letter, see our guide on Preparing for a Correspondence Audit: Document Checklist and Sample Response Letter.

  • Office and field audits: For complex returns or businesses with thousands of transactions, examiners often use statistical sampling to draw conclusions and extrapolate results to the entire population. A field audit may select a sample of invoices, expense claims or payroll records for verification. If errors are found in the sample, the examiner typically projects those errors to the full population to compute proposed additional tax and penalties.

  • Specialized audits: Certain types of audits (e.g., employment tax, excise tax, or large partnerships) use established sampling protocols tailored to the tax type and the available data.

What “projected” adjustments mean and how they’re calculated

If an examiner finds errors in the sample, those findings are often extrapolated to estimate the total error for the population. A simple example:

  1. Population = 10,000 transactions; sample = 200 transactions.
  2. Errors exist in 20 of the 200 sample items (10% error rate in the sample).
  3. Examiner applies the 10% error rate to the full population to estimate 1,000 problematic transactions; dollar amounts for the projected adjustment are calculated from this estimate.

In statistical sampling, the examiner should consider confidence intervals and tolerable error; in practice, however, taxpayers often see a point projection without the full statistical report. You can ask the examiner for the sampling workpapers and the extrapolation method — these are essential to evaluate whether the projection is correct and whether the sample was representative.

How sampling choices affect your outcome

  • A probabilistic sample with solid documentation tends to limit the IRS’s ability to claim bias and makes their projection easier to test. A well‑designed statistical sample gives both sides a better basis for negotiation.
  • Judgmental selection can generate intense scrutiny of specific items; if those items are vital to your tax outcome (e.g., large charitable gifts or vehicle deductions), you may face larger adjustments.
  • If records exist but the taxpayer cannot produce them for sampled items, the examiner may treat those items as unsupported and increase the projected error rate.

Practical steps if your return is sampled

  1. Read the audit notice carefully and meet deadlines. Missing deadlines limits your options.
  2. Request the sampling documentation and workpapers. Ask how the sample was drawn, what population was used, and the extrapolation formula. Examiners often must provide workpapers under procedural rules (see IRS guidance in Publication 556).
  3. Recreate the population when possible. If the IRS defined the population too broadly, offer a narrower, defensible population and propose stratification to reduce projection error.
  4. Produce contemporaneous substantiation for sampled items (receipts, mileage logs, invoices, bank records). A single well‑organized binder or electronic folder speeds review and improves credibility.
  5. If sampled results produce a large projected adjustment, propose alternative sampling or a 100% review of high‑dollar items. You can often negotiate to examine only the higher‑value items rather than accepting a straight projection.
  6. Consult a CPA, tax attorney or enrolled agent experienced with IRS statistical sampling. Complex sampling disputes are technical and benefit from professional representation.

For a practical walkthrough on assembling records used in sampled audits, see our article on Preparing a Trial Balance for an IRS Audit: What Auditors Look For.

Negotiation, appeals and taxpayer rights

If you disagree with the examiner’s sampling method or projection, you have options:

  • Ask the examiner to explain and provide the sampling plan and calculations.
  • Request supervisory review or consider asking for an Independent Office of Appeals conference if you cannot reach an agreement with the examiner.
  • Use the administrative appeals process described in IRS Publication 556 to preserve rights and deadlines.

Appeals can challenge the extrapolation method, the population definition, or the examiner’s treatment of unsupported items. A settlement via Appeals often reduces projected adjustments without costly litigation.

Recordkeeping and prevention checklist

  • Keep receipts, vendor invoices, bank statements and contemporaneous logs (e.g., mileage) for at least three years — longer for certain items (see IRS recordkeeping guidance).
  • Reconcile bank accounts monthly and maintain a simple digital filing system (PDFs named by date and type make production faster).
  • Limit rounding and approximations on returns; use reasonable estimates and be able to explain how you calculated them.
  • For businesses, maintain a trial balance and summaries that tie to supporting documents — this makes sampling and reconciliation easier during an audit.

Common misconceptions

  • “Sampling means the IRS is guessing.” While sampling produces estimates, modern statistical methods provide defensible projections. The key is whether the sample and extrapolation were done correctly and transparently.
  • “If I was sampled, I did something wrong.” Sampling is often a practical response to volume, not proof of wrongdoing. Still, high‑risk items attract judgmental selection and merit careful review.

Quick FAQ

  • What triggers sampling? Large data volumes, prior discrepancies, and high‑risk items often lead examiners to use sampling.
  • Can you force a full review instead of a projection? You can ask, but the IRS is unlikely to do a full population review for very large data sets. Negotiation, stratification, or concentrating on high‑dollar items are more realistic paths.
  • Will sampling increase penalties? If the projected errors show negligence or fraud, penalties can apply. Good documentation reduces penalty exposure.

Sources and further reading

Internal guides on FinHelp.io:

Professional disclaimer

This article is educational and reflects general practices as of 2025; it does not replace personalized tax advice. For specific audit representation or legal counsel, consult a licensed tax professional or attorney.