Concept
What Counts as Remedial
Remedial measures
What Counts as Remedial
Rule 407 bars evidence only when it describes the right kind of step. Its opening clause sets the condition: the rule speaks to “[w]hen measures are taken that would have made an earlier injury or harm less likely to occur.”1 A court applying that clause asks three questions about the measure. What kind of change is it? When did it come, measured against the injury? Who made it? The text answers the first two in its own words, “would have made … less likely” and “earlier injury or harm.” It leaves the third unanswered, because the clause is passive and names no actor.
The two justifications from the last section supply the answers. The first holds that a response to harm proves little about earlier fault. The Advisory Committee called such conduct “equally consistent with injury by mere accident or through contributory negligence,” and it described the rule as rejecting the notion that “because the world gets wiser as it gets older, therefore it was foolish before.”2 The second holds that the law should encourage safety steps, or at least not discourage them, and a party who expects its repair to be offered against it at trial has reason to delay the repair.3 Together the two reach any response that a party might postpone and a jury might misread. They explain why timing matters, since only a change that follows the injury can react to it. They bear most heavily on the actor, because the rule protects a party’s own decision. The subsections take the three questions in that order.
Types of Measures
Rule 407 defines the covered measures by what they would have done, not by what they are. A measure counts if, taken before the harm, it would have made the harm less likely; the text names no repair, device, or rule as a category.4 The Advisory Committee confirmed the breadth. Courts had applied the principle to repairs, safety devices, changes in company rules, and discharges of employees, and the note says “the language of the present rule is broad enough to encompass all of them.”5 A nailed-down board and a rewritten procedure are equally ambiguous proof of fault and equally easy to postpone, so a functional test treats them alike.
The categories below illustrate the test; they do not close it.6
- Repairs and alterations to premises, construction, or equipment.
- New safety devices, such as signs, lights, gates, and guards.
- Changes in company rules, policies, procedures, or business practice.
- Discharge of the employee thought responsible, or reorganization of a department or unit.
- Changes in labels, warnings, or instructions.
- Product design changes.
- Tightened financial or security controls, including changes made after an audit.7
Design changes are the product-case version of the same idea. A manufacturer that learns a design element has led to accidents can be expected to change the design to prevent more, even knowing the change could be used in suits over accidents already past.8 Plaintiffs want the evidence for that reason, because eliminating the very danger that hurt the plaintiff can be powerful proof that the improvement was practicable and reasonable when made.9 The rule exists to blunt that pull. In an Iowa case, a boat-trailer jack maker’s post-injury design modification stayed out because the plaintiff offered it to prove negligence or culpable conduct.10 And Jones v. Crawforth records two further worries: manufacturers may modify a design for many reasons other than to remedy a defect, and evidence of the later change may divert the jury from whether the product was defective at the relevant time to what was done afterward.11
The label on the claim does not change coverage. The policy protects a defendant who must pay the judgment whether the jury is judging the defendant’s conduct in a negligence claim or the product in a strict liability claim, and one court saw no significance in the distinction.12 The next section develops the product cases.
The function test also carries the rule well past physical safety. It covers new company rules, discharges, added warnings, and changed trade practices. It covers any measure that would make less likely the recurrence of an asserted violation of any tangible or intangible legal interest, including discrimination, employment, and antitrust claims. The rule thus extends from “safety measures” to the broader category of “remedial” or “corrective” measures.13 A bank sued over an employee’s embezzlement shows how far the idea travels. Auditors investigate and the bank tightens its controls, and the new controls are covered when offered to show the old ones were negligent, even though the concern is business security rather than safety.14
Higginbotham v. Baxter International, Inc. applies the same reasoning to a shareholder suit over overstated profits at a Brazilian operation. The court observed, “[q]uite independent of Rule 407,” that changing the accounting protocols “does not show that earlier ones were recognized as deficient.” For all the complaint revealed, the improved controls may have been a public-relations measure, or a way to forestall litigation whose cost could exceed the losses from the fraud.15 Pugh v. Tribune Co. reached a comparable result for a reform adopted after an internal investigation.
The text sets one boundary on this breadth. Commentary reads “injury” to include discrimination and employment claims. A court asked to apply Rule 407 to a city’s later promotion method in a disparate-impact suit doubted the fit, because such claims are “not naturally described as ‘an injury or harm allegedly caused by an event.’”16 The court added that even if the rule applied, the later method bore on whether an alternative method had been available earlier, which placed the evidence within the exception for another purpose.17
Post-event investigations are the least settled case. One treatise treats an auditors’ investigation after an embezzlement as covered at least when it produced changes, and perhaps even when it did not.18 Another reports that courts generally do not exclude post-accident investigative reports.19 The definition suggests a way to reconcile them. A measure is something that, taken earlier, would have lowered the risk, while an investigation only discovers what might be changed. The safer analysis separates the report from the changes it triggers, treats the changes as covered, and treats the report’s status as unsettled.
A change made for an unrelated purpose is not covered. The covered steps are those taken to keep the harm from recurring, and a change made to make a product or property look better is a different thing.20
A public agency’s recommendation does not take a measure out of the rule. In a study problem, Island Express put in place, after an accident, the changes that an NTSB report recommended. The analysis concludes that the changes are inadmissible to prove negligence, because the agency recommended them but did not require them.21 That stated reason leaves a question open. A change that an agency compels is arguably not the party’s own decision, and the party’s own decision is what the rule protects. The last subsection returns to the point.
Coverage opens the exclusion, but it does not finish the analysis. A covered measure stays barred only for the forbidden purposes, and the court may admit it for another purpose such as impeachment or, if disputed, ownership, control, or feasibility.22 In a suit over a woman’s rape at a motel, the defendants testified that they had done everything necessary for a secure motel and that chain locks and peepholes would not work, yet after the assault they installed those very devices. The plaintiffs could show that the devices were feasible and could impeach the testimony with the later installation. They could not use the evidence to prove negligence, and a limiting instruction would warn the jury of that restriction.23 A manufacturer likewise controverts feasibility when it denies the technological possibility of a safety measure it later adopts, even implicitly, as by asserting that the product that injured the plaintiff was the safest possible.24 If it explicitly admits feasibility, proof that it later adopted the measure normally should be excluded.25 The exceptions have their own sections below. The point here is that “covered” labels the evidence and does not settle its fate.
Timing After the Injury
The rule fixes its starting point in the words “earlier injury or harm.”27 The injury the lawsuit is about sets the date; the dates of manufacture and sale do not, so a change made after the sale but before the injury lies outside the exclusion.28 A product’s history runs in a fixed order: manufacture, sale, any redesign, and last the injury. Only a redesign that follows that last event is a subsequent remedial measure, because the rule presupposes a harmful event first and a measure after it.29
The line sits at the injury because the rule’s reasoning does. The Advisory Committee describes the rule as rejecting the notion that “because the world gets wiser as it gets older, therefore it was foolish before.”30 A change that predates the injury is no reaction to the plaintiff’s injury, so that inference has nothing to work on. The incentive ground has less to hold as well, since a maker that improves a design before anyone is hurt is not weighing how a suit over that injury will use the change.
The rule’s history shows why the point needed settling. Before the 1997 amendment, courts divided over measures taken after manufacture or sale but before any accident. Doe v. Johnston called the choice between the readings “not without controversy” and reported the majority view that such measures are not “remedial” under the rule.31 The 1997 amendment put the injury into the text: the rule applies to measures taken “after an injury or harm allegedly caused by an event.” In the Third Circuit the amendment also abrogated decisions that had applied the rule to changes made before the accident.32
A decision applying the amended rule shows how little the sale date matters. In Diehl v. Blaw-Knox, the manufacturer’s redesign came in 1983, sixteen years before the accident. The court held that Rule 407 did not apply to it.33 In Straley v. United States, plaintiffs offered post-manufacture warning letters and later standards to show the defendants’ knowledge of dangers and breach of a continuing duty to warn. The court called Rule 407 “clearly not applicable,” because the “event” was the accident and the warnings were mailed before it.34 Both decisions turn on one fact, the date of the accident.
Falling outside Rule 407 does not make the evidence admissible. The Diehl court added that evidence of the redesign “must still be analyzed under Rules 401 and 403.”35 A pre-injury change can carry the same persuasive force as a post-injury one,36 and Rule 403 lets a court exclude evidence whose probative value is outweighed by the danger of unfair prejudice. One court affirmed exclusion under Rule 403 of a defendant’s later changes to a checkpoint’s operation. The defendant was already in the case, other trial evidence covered how the checkpoint ran, and the changes came a year after the accident with no proof that the defendant knew of any danger beforehand, so the probative value was slight.37 Whether a pre-injury change tells the jury anything about the product’s condition at the time of sale remains a case-by-case question.
Who Took the Measure
The text of Rule 407 says that measures are taken but not by whom.39 Courts supply the answer by asking whose decision the rule protects, and the answer is the party against whom the evidence is offered. One treatise states the consequence directly: when a third person took the measure, “the policy ground for exclusion is absent,” and the evidence, if otherwise admissible, is not excluded.40
The reason follows from the incentive rationale. A rule built to encourage or discourage conduct achieves its goal only in litigation where the person it hopes to influence is a party. Another treatise makes the point with the compromise-offer rule, which is designed to encourage compromise: admitting an offer in a case to which the offeror is not a party will not discourage compromise.41 The same reasoning applies to a repair. Admitting a nonparty’s repair exposes the nonparty to nothing, so the evidence gives it no reason to hold back. The evidentiary ground fades too. The Advisory Committee describes the rule as excluding later measures “as proof of an admission of fault,”42 and a stranger’s repair is not the defendant’s conduct, so no admission by the defendant is present.
The First Circuit states the limit by the actor’s identity. In Raymond v. Raymond Corp., the court described Rule 407 as applying “only to subsequent remedial measures taken voluntarily by the defendant.”43 The word “voluntarily” ties the rule to the decision it protects. Espeaignnette v. Gene Tierney Co. later applied the same understanding to a third party’s actions on a machine.
A mandated change raises a harder question. A measure that a regulator compels leaves the party no free choice for the rule to protect, and the third-person cases are the closest analogy, since in both settings the policy ground for exclusion is absent.44 The sources here do not settle the point, and the agency-recommendation problem above sits on the other side of it.
Outside the rule does not mean admitted. Raymond adds that Rule 403 may be used to exclude evidence of later third-party actions that bear “only marginally” on whether a product was unreasonably dangerous at the time of manufacture. It cites Grenada Steel Industries, Inc. v. Alabama Oxygen Co., where evidence of nonparty repairs was properly excluded because it “lacked probative value and injected the dangers of confusion and misleading the jury.”45
Two errors tempt on an exam. The first is to exclude a nonparty’s change because it looks like a repair and followed the same accident, when the rule asks who took the measure. The second is to treat a measure outside Rule 407 as automatically admissible, when relevance and Rule 403 still decide whether the jury hears it.