Accident predictor

In the United States in 1996, Wu and Yan-Go (31) also reported a higher accident rate in patients with OS AS. The best predictors for car accidents were falling asleep at the wheel and driving past destinations. This driving risk appeared so important that the American Thoracic Society published recommendations about driving safety among apneic subjects (32). 

A number of different approaches have been taken by researchers examining the relationship between accident statistics and employee job tenure (how long the employee has worked in the job). In some studies, researchers have formed groups of employees based on their job tenure and compared accident rates across the groups. Unfortunately, not all studies that have used this group comparison approach to study the relationship between job tenure and accidents have attempted to control for employee age across the groups. Other studies have used job tenure as a predictor variable in regression analysis or simply correlation analysis in an attempt to find associations between an employee s job tenure and accidents. 

Leigh, J. P. (1986). Individual and job characteristics as predictors of industrial accidents. Accident... 

Probst, T. M., Graso, M., Estrada, A. X., Greer, S. (2013). Consideration of future safety consequences A new predictor of employee safety. Accident Analysis and Prevention, 55, 124-134. 

In summary, selection processes can help to ensure new employee safety if they clearly define the knowledge, skills, and abilities that are required to perform a job, and obtain or develop accurate predictors of these. Put simply if an organization selects an individual for a job that does not have the knowledge, skills, and abilities which are necessary to perform the job in a safe manner, there will be an increased chance that the individual (the new employee) will be involved in an accident. Of course, working safely is also partly dependent on the new employee s attitude toward safety and on their personality (see Chap. 5). Unfortunately, attitudes and personality are not easy to measure in an error-free way. In this regard, an organization should not assume that they have very much ability at all to predict safety-related attitudes or to determine much in the way of safety behavior based on personality profiling. 

Causal factors for low-probability/high-consequence events are rarely represented in the analytical data on incidents that occur frequently, and the uniqueness of serious injury potential must be adequately addressed. However, accidents that occur frequently may be predictors of severity potential if a high-energy source was present (e.g., operation of powered mobile equipment, electrical contacts). 

They are particularly limited for assessing the future risk of high consequence, low probability accidents. (A fatal accident rate based on data from single fatalities may not be a good predictor of risk of multiple fatal emergencies.)... 

Begg, D. J. and J. D. Langley (2004). Identifying predictors of persistent non-alcohol or drug-related risky driving behaviours among a cohort of young adults. Accid. Anal Prev., 36(6) 1067-1071. 

Staplin, L. and K. W. Gish (2005). Job change rate as a crash predictor for interstate truck drivers. Accid. Anal Prev., 37, 1035-1039. 

Li, L., K. Kim and L. Nitz (1999). Predictors of safety belt use among crash-involved drivers and front seat passengers adjusting for over-reporting. Accid. Anal. Prev., 31, 631-638. 

Simple models of acddent frequencies were estimated using five predictor variables AADT, posted speed, curve, roadway width and shoulder width. Traffic volume is used as an exposure to accidents and is therefore present in all models. The prediction equation estimated is given in eqiration [6.3] ... 

It was also tested whether or not adding a variable is overall significant. To determine if curve and roadway width are overall statistically significant, we can compare a model with and without them. Table 6.6 shows that the two degree-of-freedom chi-square tests indicate that both curve and roadway width ate statistically significant predictors of the number of accidents. 

The importance of construct validity. It is possible that a direct relationship between a predictor (such as the measure of accident proneness) and a criterion (such as the number of at-risk behaviors or recordable injuries) can be found (predictive validity) without supporting the underlying principle(s) or theory. This would indicate the absence of construct validity. Suppose, for example, an individual could figure out how to answer the survey questions in order to receive a favorable score. Then, construct validity would be questionable, even if criterion validity were high. 

Hatakka, M., Keskinen, E., Katila, A. and Laapotti, S., 1997. Self-reported driving habits are valid predictors of violations and accidents. In Rothengattei T. and Vaya, E.G. (eds). Traffic and Transport Psychology. Pergamon, Amsterdam. 

Hale, A. R., J. Stoop, and J. Hommels. Human Error Models as Predictors of Accident Scenarios for Designers in Road Transport Systems. Ergonomics 33, no. 10-11 (1990) 1377-1387. 

The results of the regressions are shown in table 20.5. The variables are strong, statistically-significant predictors of the occurrence rate. However, the regressions only explain about a fifth of the variation in the occurrence rates, which implies that there are other factors at work. The regression results are used to predict the number of accidents that a railroad should expect to have. For example, the expected number of collisions and derailments for railroad i is given by ... 

As accident retest reliabilities between two periods of time turned out to be low, high correlation coefficients between predictor and accidents could not be expected. Newbold and Cobb (cited in Thorndike 1949) proposed a mathematical model which represents the maximum correlation of reliability that can be obtained between two sets of accident scores and an infallible set of predictive scores. As McBride et al. (1965) pointed out, application of this formula to the accident reliabilities of the California Driver Record Study, which yielded a correlation coefficient of 0.06 over a three year period, produced a maximum reliability coefficient of 0.29 for accidents. In other words, it is... 

The final 15 predictors of the construct sample multiple regression equation resulted in a multiple R = 0.59 which subsequently shrank to an R = 0.47 upon cross-validation. None of the simulator event variables had even marginal significance and were excluded, therefore, from the final regression equation. The concurrent prediction equation correctly classified 68.9% of the accident-free drivers and 71.2% of the accident repeaters. This was approximately 20% better than chance prediction. Because of the contrasted criterion group design, however, these validities were overestimates of what would be attained on a normal population of drivers. Results on biographical data and psychomotor functions are listed in Tab 5.15. 

The Socioeconomic Cluster was by far the most significant of the biographical predictors. Accident repeaters were associated with inferior education, low socioeconomic index scores, poor vocabulary scores, and high social deviance scores (CIDAO - part C). Traffic convictions were found to be one of the best predictors of accident involvement in other studies it was confirmed in this study, too. CIDAO - part A items appeared to measure maturity, risk-taking, driving attitudes, and emotional stability. Drivers who tended to agree with the accident-keyed items were more apt to be in the accident repeater group. 

---