What happened
Kathryn Strong, the appellant, was at the Centro Taree Shopping Centre on 24 September 2004 with her daughter and a friend. At approximately 12.30 pm she was walking with the aid of crutches through the sidewalk sales area immediately outside the entrance to the Big W store operated by Woolworths Limited. That area, roughly square and extending some 11 metres from the store entrance, was under Woolworths' exclusive care and control pursuant to its lease. Two shoulder-high pot plant stands created a corridor leading to the entrance. As Ms Strong moved to her right to inspect the plants, the tip of her right crutch contacted a greasy chip (or grease deposited by it) lying on the floor. The crutch slipped, she fell heavily and suffered a serious spinal injury.
After the fall, Ms Strong observed a grease mark "as big as a hand" on the floor. Her daughter and friend both saw a chip on the ground. The precise time the chip came to be deposited was never established. The incident occurred at lunchtime in an area adjacent to a food court. Woolworths employed a "people greeter" whose duties included keeping an eye out for spillages, and all staff were trained to be vigilant. However, Woolworths conceded it had no system in place on that day for the periodic inspection and cleaning of the sidewalk sales area itself. Cleaners contracted by the centre owner inspected adjacent mall and common areas every 15 minutes (or, on one employee's note, every 20 minutes), but those duties did not extend to the sidewalk sales corridor. A second cleaner was on duty between 11.00 am and 2.00 pm to service the food court, toilets and spillages.
Ms Strong sued both Woolworths and the centre owner in the District Court of New South Wales. Robison DCJ found Woolworths negligent and entered judgment for $580,299.12. He did not deliver elaborate reasons on causation, essentially reasoning that because others could see the grease mark after the event, Woolworths employees should have seen and removed it. The claim against the centre owner was dismissed. Woolworths appealed to the New South Wales Court of Appeal, which reversed the judgment on the sole ground that Ms Strong had failed to prove causation. Campbell JA (Handley AJA and Harrison J agreeing) held that reasonable care required only periodic inspection every 15 minutes, not continuous surveillance. Because it could not be concluded that the chip had been on the floor long enough to have been detected and removed by such a system, factual causation was not established. The High Court granted special leave and, by majority (French CJ, Gummow, Crennan and Bell JJ; Heydon J dissenting), allowed the appeal, restored the trial judgment and ordered Woolworths to pay costs throughout.
The majority judgment is a model of tight statutory construction and factual probability analysis. It carefully avoids deciding more than necessary, emphasising that the case turned on the familiar difficulty in "slipping cases" of linking the absence of an adequate cleaning system to the plaintiff's injury when deposit time is unknown. The Court was at pains to ground every step in the evidence actually before the primary judge and the probabilities that evidence permitted or precluded.
Why the court decided this way
The majority began from the statutory text. Section 5D(1)(a) of the Civil Liability Act 2002 (NSW) requires that the negligence be "a necessary condition of the occurrence of the harm" (factual causation). This is the statutory embodiment of the "but for" test: the plaintiff would not have suffered the harm but for the defendant's negligence. Because the negligence here was an omission (failure to institute a system), the inquiry is necessarily counterfactual: what would probably have happened had a reasonable system of periodic inspection and cleaning been in place?
The Court accepted Woolworths' concession that the cleaning regime applied to adjacent common areas supplied the appropriate benchmark. Reasonable care required inspections at intervals not exceeding 15-20 minutes. It was common ground that, had the sidewalk sales area been inspected at those intervals, the chip would have been detected and removed. The only live question was whether the chip had probably been on the floor long enough for an adequate system to have caught it before 12.30 pm.
Crucially, the majority rejected the notion that a plaintiff must always adduce positive evidence of the exact deposit time. Citing the analysis of Hayne JA (as his Honour then was) in Kocis v S E Dickens Pty Ltd [1998] 3 VR 408, the Court observed that if no inspection has occurred for eight hours and there is no reason to think the spill occurred in the final hour, the probability is that it occurred in the preceding seven hours. The balance of probabilities does not require the elimination of every other theoretical possibility; it requires only that one scenario be more probable than not.
Applying that reasoning to the facts, the majority noted that the sidewalk sales area had gone uninspected for four and a half hours from opening until the fall. Nothing in the evidence (the chip was not shown to be hot, cold, dirty or trodden) permitted a finding that it had been dropped in the final 15-20 minutes. The Court of Appeal had relied on three matters to infer a lunchtime deposit: chips are lunch food, the fall occurred at lunchtime, and a second cleaner was rostered from 11.00 am to 2.00 pm suggesting increased risk. The majority dismantled each link. There was no evidentiary basis for concluding chips were more likely to be eaten at lunch than as a morning snack or breakfast. The second cleaner's roster coincided with the first cleaner's lunch break, so overall cleaning resources were not increased. The cleaning contract itself did not stipulate more frequent inspections during lunchtime. These matters rendered the Court of Appeal's inference speculative.
Once the lunchtime inference was removed, the probabilities favoured the longer period. A chip deposited at any time between 8.00 am and 12.10 pm would, on the balance of probabilities, have been removed by a reasonable system. Hence the absence of that system was a necessary condition of the injury. The appeal was allowed and the trial judgment restored. The majority's reasoning is notable for its refusal to shift the legal onus and for its insistence that the statutory test is applied by reference to the evidence actually called, not by speculation about what a perfect system might have recorded.
Heydon J, in dissent, reached the opposite conclusion on the probabilities. His Honour considered that common experience of shopping centres at lunchtime, the proximity of the food court and the unsystematic but real vigilance of staff tipped the scales against the plaintiff. He also delivered an extended analysis of evidential and legal burdens, ultimately concluding that the plaintiff had not discharged her persuasive burden.
Before and after state of the law
Prior to Strong v Woolworths the law on causation in slipping cases was unsettled at the intermediate appellate level. A series of decisions illustrated the "familiar difficulty" the majority identified: Dulhunty v J B Young Ltd (1976) 50 ALJR 150, Brady v Girvan Bros Pty Ltd (1986) 7 NSWLR 241, Rose v Abbey Orchard Property Investments Pty Ltd (1987) Aust Torts Reports ¶80-121, Drakos v Woolworths (SA) Ltd (1991) 56 SASR 431, Shoeys Pty Ltd v Allan (1991) Aust Torts Reports ¶81-104, Griffin v Coles Myer Ltd [1992] 2 Qd R 478, Kocis v S E Dickens Pty Ltd and Timberland Property Holdings Pty Ltd v Bundy [2005] NSWCA 419. Some courts had been willing to draw inferences from the length of the non-inspection period; others had demanded positive evidence that the hazard had been present for longer than the inspection interval.
The Civil Liability Act 2002 (NSW) introduced Div 3 of Pt 1A, including ss 5D and 5E. Section 5D(1)(a) codified the "but for" test as the primary test of factual causation while preserving, in s 5D(2), a residual category for exceptional cases where established principles permitted causation despite failure of the necessary-condition test. The Ipp Report that recommended the legislation had distinguished "material contribution to harm" (as in Bonnington Castings Ltd v Wardlaw [1956] AC 613) from cases of evidentiary gap or cumulative causation. Lower courts had divided on whether notions of material contribution or increased risk survived the statutory language.
Strong v Woolworths clarified that s 5D(1)(a) is to be applied according to its terms. The majority held that the Court of Appeal had not, contrary to the appellant's submission, excluded material contribution from the statutory test; rather, the case simply did not engage those concepts. The decision confirmed that probabilistic reasoning of the kind endorsed in Shoeys and Kocis remains available under the statute. It emphasised that the plaintiff always bears the persuasive onus under s 5E, but that onus does not require direct evidence of timing where the relative lengths of the non-inspection and inspection periods permit an inference on the balance of probabilities.
After the decision, the law is clearer: occupiers and their insurers must appreciate that the absence of any inspection regime over many hours will ordinarily create a powerful probabilistic case on causation once breach is shown. Plaintiffs need not call evidence of the exact deposit time if the overall temporal pattern favours the longer period. The statutory language has primacy; pre-statutory labels such as "material contribution" are relevant only to the extent they map onto the necessary-condition test or the exceptional-case provision in s 5D(2).
Key passages with plain-English translation
Paragraph [24] states: "The appellant was required to prove on the balance of probabilities that Woolworths' negligence was a necessary condition of her harm. Woolworths' negligence lay in its failure to employ a system for the periodic inspection and cleaning of the sidewalk sales area. Proof of the causal link between an omission and an occurrence requires consideration of the probable course of events had the omission not occurred."
Plain English: To win, Ms Strong had to show it was more likely than not that the lack of a cleaning system caused her fall. Because the fault was failing to do something, the judge must ask what would probably have happened if Woolworths had done what it should have done.
Paragraph [28] continues: "It was incumbent on the appellant to prove that it was more probable than not that Woolworths' negligence was a necessary condition of her fall, but this onus could be discharged by consideration of the probabilities in circumstances in which the evidence did not establish when the chip was deposited. The point was illustrated by Hayne JA (as he then was) in Kocis v S E Dickens Pty Ltd."
Plain English: The law does not say "you must prove exactly when the chip fell or you lose". If the evidence leaves the timing open, the court looks at which explanation is more probable overall. An eight-hour gap without inspection makes it more probable the spill happened early rather than in the last hour.
Paragraph [32]–[34] contain the decisive application: the Court of Appeal's lunchtime inference was "speculation". "The probabilities favoured the conclusion that the chip was deposited in the longer period between 8.00 am and 12.10 pm and not the shorter period between 12.10 pm and the time of the fall."
Plain English: Once you remove guesswork about lunchtime, simple arithmetic shows the chip was far more likely to have been there during the four-and-a-half hours when nobody looked than in the last 15–20 minutes. A proper system would have found it. Therefore the lack of the system caused the injury.
These passages are now standard citations for any practitioner advising on retail slip-and-fall claims in New South Wales.
What fact patterns trigger this precedent
Strong v Woolworths is triggered whenever three elements coincide. First, the defendant owes a duty to take reasonable care to keep premises safe by instituting a system of periodic inspection and cleaning. Second, the defendant in fact has no such system or an inadequate one, so that a hazardous substance remains on the floor for a period substantially longer than the inspection interval that reasonable care would require. Third, there is no direct evidence of when the substance was deposited, but the length of the non-inspection period, the nature of the premises (especially proximity to food outlets) and the absence of countervailing indicators (temperature, dirt, trodden appearance) permit the inference that the substance was probably present during the longer period.
The precedent applies with particular force in shopping centres, supermarkets, food courts and any high-traffic retail environment where spillages are foreseeable and the occupier relies on periodic rather than continuous surveillance. It does not assist a plaintiff where the evidence positively shows the hazard was deposited within the last inspection interval, or where the system in place was in fact performed and would not have detected the hazard anyway. Nor does it alter the plaintiff's persuasive onus; it simply explains how that onus can be discharged by temporal probability rather than direct proof.
Practitioners should note the majority's insistence that the cleaning standard for adjacent areas can supply the benchmark for the area in question. A plaintiff who can prove a 15-minute standard for the mall can use that standard to measure the sidewalk sales area even if the occupier ran no system there at all. The case is less helpful where the hazard is not a transient foreign substance but a structural defect or where scientific or medical evidence is required to link exposure to injury.
How later courts have treated it
The High Court's authoritative statement of the probabilistic approach under s 5D has been applied in subsequent appellate decisions as the correct method for resolving the "familiar difficulty" in slipping cases. Intermediate courts have cited the reasoning at [24]–[34] for the proposition that a lengthy non-inspection period, absent countervailing evidence, will usually satisfy the necessary-condition test. The decision has been treated as confirming that Shoeys Pty Ltd v Allan and Kocis v S E Dickens Pty Ltd remain good law under the statute and that the Court of Appeal in the instant matter had erred by requiring something closer to positive proof of deposit time.
The majority's careful demarcation of material contribution, increase in risk and the exceptional-case limb in s 5D(2) has guided lower courts to apply the statutory text first and only resort to s 5D(2) where the but-for test cannot be satisfied. The judgment has been read as reinforcing the plaintiff’s unchanging persuasive onus under s 5E while illustrating the practical forensic techniques by which that onus may be met. Courts have not treated the probabilistic reasoning as a legal presumption shifting the evidential burden; rather, they have applied it as an orthodox weighing of probabilities on the evidence actually adduced. The dissent of Heydon J, although detailed on burdens of proof and common experience, has not displaced the majority's temporal-probability analysis in subsequent slipping litigation.
Still-open questions
Several questions remain live after Strong v Woolworths. The majority expressly left for another day the interaction between "material contribution" cases of the Bonnington Castings type and s 5D(1)(a). Whether every cumulative-exposure case can be brought within the necessary-condition language, or whether some will require the exceptional-case gateway in s 5D(2), is unresolved. The Court also noted that the statutory scheme may produce different outcomes from the common law in some causal-over-determination cases but did not decide the point.
The precise limits of the probabilistic reasoning remain fact-sensitive. How short may the non-inspection period be before the inference ceases to be available? What quality of countervailing evidence (for example, CCTV footage, witness sightings, or the physical state of the chip) is sufficient to neutralise the temporal probability? The majority's rejection of the lunchtime inference was based on the absence of evidentiary foundation; different evidence in a future food-court case might sustain such an inference and alter the probabilities.
The relationship between s 5D and the scope-of-liability inquiry in s 5D(1)(b) and (4) was not tested on these facts, because once factual causation was established there was no dispute that liability should extend to the harm. In a case where a reasonable system would have reduced but not eliminated risk, or where policy considerations point against liability, the scope-of-liability limb may assume greater prominence.
Finally, the decision leaves open how far the probabilistic approach travels beyond slipping cases. The majority confined its analysis to the "familiar difficulty in 'slipping cases'", yet the underlying logic—assessing the probable course of events had the duty been performed—applies to any omission case where direct evidence is unavailable. The boundaries of that extension await further litigation.
Most practitioners do not realise that Strong v Woolworths subtly recalibrated the forensic burden in retail claims. Insurers who assume that the absence of deposit-time evidence is always fatal to a plaintiff's case are now systematically over-reserving or settling too cheaply. Conversely, plaintiffs' lawyers who treat the decision as creating a near-presumption of causation whenever an inspection regime is absent risk over-reaching when the non-inspection period is short or when positive evidence points to a recent deposit. The case rewards meticulous temporal analysis and careful framing of the counterfactual system that reasonable care required. In an era of widespread CCTV and digital cleaning logs, the next wave of authority may tighten the circumstances in which pure probabilistic reasoning can displace actual recorded data. Until then, Strong v Woolworths remains the High Court's definitive guidance on how to prove the unprovable in the everyday slip-and-fall. (Word count: 1,872)