Harmless Error vs. Reversible Error: When Mistakes Matter in Court

Last updated 1 hour ago. Our resources are updated regularly but please keep in mind that links, programs, policies, and contact information do change.

A mistake at trial does not automatically mean a new trial. A trial judge can admit evidence that should have been excluded, yet an appeals court can leave the conviction standing. Federal appeals courts must disregard any error that did not affect “substantial rights,” which in most federal criminal cases means affecting the outcome.

An error that did not affect those rights is called harmless, and the judgment stands. One that did can be reversible error, a reason to change the judgment. For a defendant, a family member or anyone following an appeal, what happens next depends on whether anyone objected at trial, what kind of error it was and which court is reviewing it.

A mistake does not settle the appeal

Federal courts of appeals review whether proceedings were fair and the law was applied correctly; they do not retry cases, hear new evidence or take witness testimony. An appeal over a trial error therefore asks two questions: whether a rule was broken, and whether the break justifies changing the judgment.

In federal criminal cases, Rule 52(a) requires courts to disregard an error, defect or irregularity that does not affect substantial rights. Federal civil Rule 61 likewise requires disregarding errors that do not affect a party’s substantial rights, and says that an error is not grounds to disturb a judgment unless justice requires otherwise. The civil rule covers more than a mistaken evidence ruling: its language reaches errors by the court or a party.

Calling an error harmless does not make the original ruling correct; it means the error does not justify changing the result under the applicable test. Reversible error, likewise, is not a fixed list of forbidden acts; it is a mistake that meets the test for changing the judgment.

The record determines whether the error mattered

Federal Rule of Evidence 103 lets a party claim error in a ruling that admits or excludes evidence only if the error affects a substantial right of that party and the party raised the issue at trial in the way the rule requires. In its 1978 decision Burks v. United States, a case about when a defendant can be tried again, the Supreme Court named incorrect evidence rulings, incorrect instructions and prosecutorial misconduct as examples of trial error. Falling into one of those categories does not establish that a mistake changed the outcome.

In its 1946 decision Kotteakos v. United States, the Supreme Court confronted a prosecution that charged one conspiracy, although the evidence showed eight or more separate conspiracies by groups whose only connection was that each used the same broker to handle fraudulent loan applications. The jury was told that the acts and statements of one conspirator counted against all of them, and evidence of the broker’s dealings with other defendants was admitted against the petitioner, the defendant who brought the case to the Supreme Court. The problem was the risk that proof about one group would be used to convict someone in a different group.

The Supreme Court reversed because the petitioner’s rights were substantially prejudiced. Kotteakos says the inquiry cannot be merely whether enough evidence remains to support the result. The question is whether the error itself had substantial influence; if it did, or if the reviewing court is left in grave doubt, the conviction cannot stand.

Objecting changes the path to reversal

Federal criminal Rule 51 permits a party to preserve an error claim by telling the trial court, when a ruling is made or sought, what action the party wants or what objection and grounds it has. If there was no opportunity to object, Rule 51 says the absence of an objection does not later prejudice the party. Objecting gives the trial court a chance to fix the problem while the trial is still under way.

For admitted evidence, Rule 103 ordinarily requires a timely objection or motion to strike with a specific ground, unless that ground is apparent from context. For excluded evidence, the party ordinarily must explain its substance through an offer of proof, unless the substance is apparent from context. An offer of proof puts on the record what the missing evidence would have shown, so the court can assess what its exclusion means.

Under federal criminal Rule 30, an objection to jury instructions or a failure to give a requested instruction must specify the objection and grounds before the jury retires to deliberate. Without that objection, Rule 30 bars appellate review except as permitted by Rule 52(b), the rule on plain error.

Rule 52(b) permits consideration of a plain error affecting substantial rights even when it was not brought to the trial court’s attention. Plain error is a route for reviewing a mistake no one objected to, rather than another name for every reversible error. Under United States v. Olano, a 1993 Supreme Court decision, the error must be clear or obvious and must affect substantial rights, ordinarily by affecting the outcome. Even then, correction depends on whether the error seriously affects the fairness, integrity or public reputation of judicial proceedings.

After a timely objection, the government ordinarily bears the burden of showing the error was not prejudicial, meaning it did not affect the outcome. On plain-error review, it is the defendant who bears the burden of showing prejudice.

Olano distinguishes forfeiture, failing to assert a right on time, from waiver, intentionally relinquishing or abandoning a known right. An intentional waiver can extinguish the claimed error, making it different from an objection merely overlooked.

In Olano itself, alternate jurors were present during the jury’s deliberations, a conceded violation of a federal rule. Because the defendants had not shown prejudice, the Supreme Court held that the appeals court had no authority to correct the error. The Supreme Court would not presume that the alternates’ mere presence had chilled the deliberations. A conceded rule violation was still not enough to win relief on plain-error review.

Constitutional errors and the narrow structural exception

In Chapman v. California, decided in 1967, the Supreme Court rejected the rule that every federal constitutional error must automatically require reversal. For constitutional errors subject to harmless-error review, Chapman requires the beneficiary of the error to prove beyond a reasonable doubt that the error did not contribute to the verdict. When the violation hurt the defendant, the side that benefited is the prosecution, so the prosecution must make the case that the error was harmless. The test is about the error’s contribution to the verdict, not a new determination of whether the accused is guilty.

In Chapman, the prosecutor repeatedly used the defendants’ failure to testify to suggest guilt, and the trial judge told jurors they could draw adverse inferences from that silence. Applying its constitutional harmless-error test, the Supreme Court concluded that those errors were not harmless.

Structural errors are defects affecting the framework of a trial, rather than simply errors in its process. The Supreme Court’s recognized examples include complete denial of counsel, a biased trial judge, racial discrimination in grand-jury selection and denial of self-representation or a public trial. When a structural error is objected to at trial and raised on direct review, the ordinary appeal of a conviction, the defendant is entitled to reversal without an inquiry into harm. A constitutional label alone does not place an error in that category.

Arizona v. Fulminante, decided in 1991, treats admission of an involuntary confession as subject to harmless-error analysis. In Neder v. United States, decided in 1999, the Supreme Court said that leaving an element of the crime out of the jury instructions can be harmless when a reviewing court concludes beyond a reasonable doubt that the omitted element was uncontested and supported by overwhelming evidence, so the verdict would have been the same.

By contrast, Sullivan v. Louisiana, decided in 1993, explains that when a judge gives jurors an unconstitutional definition of reasonable doubt, there is no jury verdict of guilty beyond a reasonable doubt for harmless-error review to examine. An instruction can therefore be wrong in different ways, with different consequences for review.

In Weaver v. Massachusetts, decided in 2017, a courtroom was closed to the public during jury selection, and the defense lawyer neither objected nor raised the issue on direct review. When the closure was raised later through an ineffective-assistance-of-counsel claim, the Supreme Court held that the defendant must show prejudice to secure a new trial. Ineffective assistance concerns a lawyer’s failure to provide the constitutionally required representation, here the failure to object. Weaver shows that how an error reaches the court can matter even when the error is structural. Its limited holding does not establish a prejudice rule for every kind of structural defect.

Missing proof is different from a flawed trial

A claim that the evidence was legally insufficient asks whether the prosecution proved the charge at all, rather than whether a trial mistake was harmless. Jackson v. Virginia, decided in 1979, asks whether any rational factfinder, viewing the evidence in the light most favorable to the prosecution, could find every essential element of the crime beyond a reasonable doubt. It is not enough for the reviewing judge simply to prefer a different account of the facts. The everyday phrase “there was no evidence” can mislead: the legal question is not whether there was some evidence, but whether the proof could persuade a rational factfinder of every required element.

In Burks v. United States, decided in 1978, the Supreme Court held that reversal for legally insufficient evidence bars a second trial under the Double Jeopardy Clause and requires entry of a judgment of acquittal. The prosecution does not get another opportunity to supply the proof it failed to produce at the first trial. Reversal for trial error is different: it means the defendant was convicted through a defective judicial process, and a retrial is allowed.

Why harmless-error rules developed

There is no single case that created every part of harmless-error doctrine. Congress adopted the federal harmless-error policy in 1919, after criticism of appellate review dominated by technical errors. Kotteakos later explained how to examine the effect of an error, and Chapman addressed federal constitutional errors against a background of existing state and federal harmless-error rules. Disregarding immaterial errors lets review focus on whether a judgment should change, rather than treating every technical failure as a reason to repeat a trial.

State appeals and federal habeas use different rules

When the Supreme Court decided Chapman in 1967, it noted that all 50 states had harmless-error statutes or rules. In Georgia criminal appeals, the nonconstitutional harmless-error test asks whether it is highly probable that the error did not contribute to the verdict. That formulation is one reason a federal overview cannot be treated as a complete statement of a state’s appellate rules. Georgia also has a statutory rule requiring a new trial when a trial judge comments on the accused’s guilt, even without an objection. A specific state command can therefore change the answer that ordinary harmless-error or plain-error analysis would suggest.

Federal habeas law lets a person held under a state court judgment challenge that custody as a violation of federal law, with restrictions that include first exhausting the remedies available in the state’s courts. That later challenge is separate from the direct appeal of the conviction. Section 2244(d) generally gives a person held under a state-court judgment one year to file a federal habeas application, starting from the latest applicable statutory trigger. Those triggers are finality after direct review or expiration of the time to seek it; removal of a state-created impediment violating the U.S. Constitution or federal law that prevented filing; initial Supreme Court recognition of the asserted constitutional right, if newly recognized and made retroactive on collateral review; or when due diligence could have uncovered the claim’s factual basis. Time while a properly filed state postconviction or other collateral-review application concerning the relevant judgment or claim is pending does not count toward that period. In Brecht v. Abrahamson, decided in 1993, the prosecution had used the defendant’s silence after Miranda warnings to challenge his credibility, a due process violation. The Supreme Court held that Chapman’s harmless-beyond-a-reasonable-doubt test does not apply when a federal habeas court decides whether that error entitles a prisoner to relief. The reviewing court asks whether the trial error had a substantial harmful effect or influence on the jury’s verdict.

A federal statute, 28 U.S.C. § 2254(d), sets a further limit for a claim that a state court decided on its merits. Federal habeas relief is allowed only if the state decision contradicted or unreasonably applied clearly established federal law as determined by the U.S. Supreme Court, or rested on an unreasonable determination of the facts in light of the evidence presented in the state proceeding. Brown v. Davenport, decided in 2022, requires applying both that statutory test and Brecht before granting relief in that setting.

To understand a particular error claim, identify the court and case type, the rule allegedly broken, the objection in the record and the stage of review. Then distinguish the requested remedy: correcting a flawed trial is different from recognizing that the government never proved its case.