This paper tests the idea that petrol prices respond more quickly to price increases than to decreases. We show that the results previously documented in the literature for Australia are spurious due to failure to establish the stationarity property of the price series, and the co-integration relationship between retail and wholesale prices when neglecting to account for a regime shift in the data. Using a robust approach involving a threshold error correction model, we find little evidence to support the contention that retail petrol price reverts asymmetrically to long-run equilibrium. Asymmetric adjustments in retail prices are found only in four of the twenty-eight retail gas stations in Queensland. These results cast doubt on the previously reported pervasiveness of this asymmetric price response phenomenon in Australia. We further caution on erroneous inference with the use of weekly rather than daily data, and when failing to account for a regime shift in the data.