ANALYSIS: Is winter driving around Rossland more dangerous than in the past? ICBC data from 1996 to 2011 spins our wheels

Not a highways accident...but a good picture--Photo A. Bennett

Recent car accidents have stirred up discontent with the state of winter highways and have resurrected concerns that maintenance standards on provincial highways might have fallen.

With all the talk going on in the area, we decided to contact ICBC for data that might help us answer the question, ‘Is winter driving in our neck of the woods more dangerous now than it was a decade ago?’.

We'll get to the results—winter driving is decidedly more dangerous in Rossland than elsewhere in our immediate area and our region, but no more so than it was a decade ago. But first, some caveats on the data.

Susan Floro of ICBC explained that their data is far from perfect. Even though operational systems go back to 1974, all the data prior to 1996 is on "tapes" made for 1950s era IBMs.

"Before 1996, I don't believe 'city' was taken as a [data] field," Floro said. "Everything was just on handwritten notes sorted by claim number."

These old claims are now stored in sequence on microfiche, most without a location noted, "just transaction number and amount," Floro said. Back then, ICBC was only concerned with whether the accident was in or out of the province.

"Originally the system was designed to monitor a claim from when it was called in until it was paid off and signed off," Floro said. Since 2005, when the province put ICBC onto new tasks such as driver licensing, ticket collections, road safety, and others, the system has been "pushing its limits."

"The system's not built for it," Floro explained. Incorporating a digital road atlas has vastly improved their inner city data, accurately pinpointing problem intersections, but ICBC haven't reached this level of sophistication on highways.

The short story is this: data from 1996 to 2011 is available, but location data from 1996 to 2000 is "poorly presented in the database," Floro said. From 2000 forward the data is "pretty good."

The data does not, however, differentiate crashes on city roads from crashes on highways. Our interest here is highways, but this cannot be separated from city accidents in the data.

Nevertheless, we took ICBC's data on the number of crashes in the Southern Interior region—Osoyoos to Cranbrook, Kamloops to Golden, and everywhere in between—and we lumped together "Our Area"—namely  Brilliant, Castlegar, Columbia Gardens, Fruitvale, Genelle, Montrose, Oasis, Ootischenia, Trail, Waneta, Warfield, and Rossland.

Other towns like Paterson weren't in the database; crashes in these areas were attributed to a nearby town by ICBC.

Next, we lumped months together into "Winter"—December to March—and "Other"—April to November.

A more precise analysis would keep months and towns separate within these categories to measure ‘variance’ and determine statistical significance, but we didn’t go there.

If winter maintenance has become a worse problem, we'd expect to see an increase in the number of "Winter" collisions relative to those in "Other" seasons, regardless of changing population sizes or traffic patterns.

That's not what the data shows, however. The percentage of winter accidents hasn't changed very much over the years.

In the Southern Interior as a whole, the percentage of winter accidents has bumped between 31.3 per cent and 36.6 per cent with no discernible trend, with an average of 33.8 per cent of the average annual total of roughly 33,200 accidents.

In our area the variation is greater, likely due to fewer crashes—technically a smaller ‘sample size’—but there is still no trend. The percentage of winter accidents varies between 30.2 and 42.9 per cent, with an average of 37.7 per cent of the average annual total of some 1290 accidents.

In Rossland itself, excluding 1996 and 1997 as anomalies, the percentage of winter accidents varies from 55.5 per cent (1999) to 41.4 per cent (2000) with an average of 46.6 per cent of the average annual total of about 148 accidents.

In terms of absolute numbers of crashes, there has been a slight upward trend in the Southern Interior as a whole, and fluctuating levels in our area, but in both cases the trends in winter matches the trends in other seasons.

What does this mean? Well, one obvious conclusion is that winter driving is more dangerous in Rossland than elsewhere in our region. In the Southern Interior as a whole, you're just as likely to get schmucked in the winter as any other season; in Rossland, however, nearly half of accidents occur between December and March.

But as far as ICBC data can tell us, there's no reason to suspect that winter driving is any more dangerous now than it was in 1996.

Although the trends in crashes haven’t changed, however, that result doesn’t necessarily correlate to equally good road maintenance.

The problem is psychological, variously called the “Peltzman effect” or “risk homeostasis”: roads could very well be in worse condition now, but people typically compensate by staying home, pulling over, or driving slower. Conversely, when conditions are good, people drive faster and more recklessly, so the prevalence of accidents remains about the same in both cases.

We could ask ICBC for additional data fields to pick apart the question, ‘Are the winter accidents now any worse than they were before?’ But this approach is still marred by the Peltzmann effect. What we really need is data to measure whether winter “inconveniences” have increased. This data, of course, does not exist.

Nor does ICBC data exist prior to 1996—unless someone would like to spend weeks scanning microfiche and taking notes.

Consequently, the important questions raised by Ed Davies and others remain unanswered, namely whether Emcon’s tenure on maintenance since 1991 was worse than Bell’s from 1988 to 1991, and the Ministry of Highway’s before.

What’s the best option for now? Drive safely!