Item: Pitfalls in the analyses of accident records: the meaning of missing values
Title: Pitfalls in the analyses of accident records: the meaning of missing values
Proceedings: International Snow Science Workshop, Davos 2009, Proceedings
Authors: Bob Uttl1,* and Kelly Kisinger2 1 Mount Royal College, Calgary, AB, Canada 2 University of Calgary, Calgary, AB, Canada
Abstract: Accident records are frequently used to learn about conditions and behavior before, during, and after accidents. However, accident records frequently do not mention the conditions and behavior of interest (e.g., thaw or melting snow) leading to missing values and to uncertainty about their presence or absence. In the development of the Avaluator Avalanche Accident Prevention Card, Haegeli & McCammon (2006) analyzed avalanche accident records for the presence or absence of obvious clues to avalanche danger (e.g., avalanches, thaw) to determine the prevention values of limiting one's travel to slopes with a certain number of clues. However, they were unable to determine the presence or absence of all clues in 1,148 accidents out of 1,400 and deleted them from their data set. Thus, the Avaluator's Obvious Clues are based on only 252 selected accidents. The critical question arises: What is the meaning of these missing values? If the missing values occurred because eyewitnesses did not mention absence of absent clues, the missing values mean absence and the exclusion of accidents was inappropriate. To investigate the meaning of missing values, study participants were shown video clips of accidents and later asked to recall as much information as they could remember about the accident. The results show that eyewitnesses are far more likely to report the presence of present conditions and behaviors (e.g., snow) than the absence of absent conditions and behaviors (e.g., absence of snow). The results highlight that the exlusion of 82% of all accidents by Haegeli and McCammon (2006) prior to the computation of the Obvious Clues prevention values was inappropriate and resulted in inflated prevention values.
Keywords: avaluator, obvious clues, accident records, missing values, avalanche prevention
Digital Abstract Not Available