What does the IRS actually look for when it selects returns for audit?

Where the claim comes from

Lists of audit triggers are popular because they promise control over something that feels unpredictable. Some items on those lists reflect real risk factors, such as large unexplained deductions or income that does not match information returns. But the lists present guesses as if they were the IRS's actual criteria, and the claim that audits are purely random is not accurate either.

Audit rates published in the IRS Data Book also shape these beliefs. They show that examination rates differ by income level and by type of return, but they describe past results, not the scoring rules. Knowing that a group is audited more often does not tell any one taxpayer why a specific return was chosen.

What the law actually says

IRM 4.1.2.6 explains that DIF "is a mathematical technique used to score income tax returns for examination potential." The manual states that the formulas were developed from National Research Program (NRP) data, that each return measured under DIF receives a score, and that, generally, the higher the score, the greater the audit potential. It also states that the DIF score assigned to a return should not be disclosed.

Congress protected those standards. IRC §6103(b)(2) provides that nothing in the law "shall be construed to require the disclosure of standards used or to be used for the selection of returns for examination, or data used or to be used for determining such standards," if the Secretary determines disclosure would seriously impair enforcement.

The IRS describes other selection methods on its audit page. Under random selection and computer screening, "sometimes returns are selected based solely on a statistical formula," comparing your return with norms for similar returns. Under related examinations, returns may be selected when they involve issues or transactions with other taxpayers, such as business partners or investors, whose returns were selected. NRP research examinations are a true random sample used to update the formulas.

Separately, the Automated Underreporter program described in IRM 4.19.3 compares the income you report with Forms W-2, 1099, and other information returns. A mismatch often produces a CP2000 notice proposing changes, which is a correspondence matter rather than a full examination.

What is true and what is not

It is true that a small share of examinations come from random research samples, and that most selection is driven by scoring, matching, and related-party connections. It is true that some patterns commonly draw attention, such as income missing from a return when a Form 1099 was filed, or deductions that are unusually large compared with income.

It is not true that anyone outside the IRS knows the DIF formulas or can tell you your score. It is also not true that avoiding a list of items keeps you safe. A return that leaves off reported income will generate a notice regardless of its DIF score.

Selection is also not the same as a finding of error. A high score means the return resembles returns that have produced changes in the past, and many examinations close with no change or a small one. The outcome depends on the documents you can provide.

• True: some returns are chosen through random research samples.
• True: information-return mismatches are detected by computer matching.
• Not true: online lists reflect the actual scoring formulas.
• Not true: legitimate deductions should be skipped to avoid scrutiny.

What to do instead

Report every item of income that appears on an information return, and reconcile your Forms 1099 and 1099-K with your books before filing. Claim the deductions you are entitled to, and keep the records that support them. Skipping legitimate deductions out of fear costs money without making the return more accurate.

If you receive a notice, respond by the deadline. Many notices are resolved by showing that the income was reported elsewhere or by explaining a mismatch with documents.

Keep records for at least three years from the date you file, and longer for property, carryovers, and situations where a longer assessment period can apply. The IRS generally examines recent returns, but it can go back further when it finds substantial errors.

If you use a payment app or marketplace, review any Form 1099-K for personal transfers or reimbursements that are not income, and keep the records that explain the difference.

How ebotCPA helps

We reconcile information returns against your books before filing, document the positions on your return, and respond to IRS notices and examinations on your behalf.

Book a $497 Case Analysis to have Ebot Mbi, CPA, EA review your facts before you act.

Primary sources

1. IRM 4.1.2.6. Discriminant Function (DIF) system.

   “is a mathematical technique used to score income tax returns for examination potential.”

   Describes DIF scoring, its basis in NRP data, and that higher scores mean greater audit potential.

2. 26 U.S.C. §6103(b)(2). Nondisclosure of return selection standards.

   “Nothing in the preceding sentence, or in any other provision of law, shall be construed to require the disclosure of standards used or to be used for the selection of returns for examination, or data used or to be used for determining such standards, if the Secretary determines that such disclosure will seriously impair assessment, collection, or enforcement under the internal revenue laws.”

   Protects return selection standards from public disclosure.

3. IRS, IRS Audits. How returns are selected for audit.

   “sometimes returns are selected based solely on a statistical formula”

   Describes random selection, computer screening, and related examinations.

4. IRM 4.19.3. IMF Automated Underreporter Program.

   Sets procedures for matching information returns to individual income tax returns.

5. 26 U.S.C. §7602(a). Examination authority.

   Authorizes the IRS to examine records and take testimony once a return is selected.

Frequently asked questions