Chart Ancestry: Rich-->Carl Binder-->Bea Barrett-->Og
CONSECUTIVE HOURLY CHARTING
| I am submitting these charts as an example of scatter
plotting behavior by time of day. While these are very grim statistics
you can use the same tool for looking at any behavior over time. The chart
is changed from "Weekly Count" to "Hourly Count." Consecutive lines represent
consecutive time periods. I have even used consecutive lines for one minute
or ten minute counting periods.
This is a very valuable tool for sensitively and rapidly analying behavior. It will help you understand the most probable times for a particular behavior to occur. You can use this tool to help you determine the best time to intervene in a particular behavior, and to rapidly demonstrate the outcome of an intervention. |
 |
|
|
MOTIVATION FOR THESE CHARTS
Back in 1980 I was renovating our house in Cambridge. We had a wonderful contractor and he had hired his brother, Allan Pierce to do the finish work on our kitchen and new entryway. Allan was an extremely careful, thoughtful craftsman, possessing both creativity and patience for detail work.
Not long after he did the work on our kitchen Allan and his family were killed by a drunk teenager. Allan, his wife, and his newborn daughter, along with Allan's father, were riding in a Toyata Corolla just after midnight on Christmas. All four were killed by a boy driving a very large American car. The teenager had some slight injuries and was released the next day. I don't believe he was even indicted for vehicular homicide.
I felt that the relationship between drinking and driving should be criminalized so that these deaths were treated as a murder. Basically I felt that drinking before driving amounted to premeditation. At that time "murders" in Massachusetts about one hundred per year. Killed by automobiles were above 1,000. Moreover, murders are typically "family affairs" while death in an automobile is more likely to involve strangers. There was little punishment in the system for drinking and driving, and I thought that needed to be changed.
| Among the tools I used to support my argument were
these charts. They show the difference between "all accidents" and "fatal
accidents" over the course of a 24 hour period for different years. Over
ten years of data, the patterns were quite strikingly similar.
The different patterns are very instructive. "Total Accidents" reflects how many vehicles are out on the roads bumping into each other. The highest number of accidents occurs in the 5:00 PM time slot. |
 |
|
|
"Fatal Accidents" climb throughout the day, reaching a peak in the 1:00 AM Time slot. Here in MA, that is when the bars were mandated to close.
A USEFUL TOOL
This technique is a graphic representation of Paul Touchette's Scatter Plot method of viewing behavior across time. The Standard Chart clarifies the data and allows you to quickly infer an extremely strong link between drinking and driving. Back then many alcohol related fatalities were covered up to preserve reputations, etc. In particular drivers who were killed were not tested for blood alcohol. Unless an open bottle was found in the car a death was NOT "alcohol related."
My presentation was unsuccessful but that was NOT the fault of the technique. You can use this approach to study behavior throughout the day as you search for patterns and bursts.
 |
I also used a somewhat different technique to look at "days of the week" and their relationship with drinking. You can again see that Saturday night, Sunday AM are the worst times to be on the road, with Tuesday being the lowest. Again the chart provides a clear picture of the data ten years by day of the week. For what it is worth, fatal accidents are highest in October (here in MA) and lowest in February. I suspect that October is highest because there are two "holidays" (Halloween and Columbus day) and we are changing the clocks. Naturally it is important that it is a 31 day month--that is why February is the lowest. |
|
|
Best wishes,