Agenda - Council - 08/13/2013 - Laserfiche WebLink

Home

Ramsey, Nowthen, St. Francis, Oak Grove, and Bethel, Minnesota Feasibility Study for Shared or Cooperative Fire and Emergency Services Averages should be viewed with a certain amount of caution because the average measure can be skewed if an unusual data point, known as an outlier, is present within the data set. Depending on the sample size of the data set, this skewing can be either very large or very small. As an example, assume that a particular station with a response time objective of six minutes or less had five calls on a particular day. If four of the calls had a response time of eight minutes while the other call was across the street and only a few seconds away, the average would indicate the station was achieving its performance goal. However, four of the five calls, or 80 percent, were beyond the stated response time performance objective. The reason for computing the average is because of its common use and ease of understanding. The most important reason for not using averages for performance standards is that it does not accurately reflect the performance for the entire data set. With the average measure, it is recognized that some data points are below the average and some are above the average. The same is true for a median measure which simply arranges the data set in order and finds the value in which 50 percent of the data points are below the median and the other half are above the median value. This is also called the 504h percentile. When dealing with percentiles, the actual value of the individual cata does not have the same impact as it did in the average. The reason for this is that the percentile is nothing more than the ranking of the data set. The 90`h percentile means that 10 percent of the data is greater than the value stated and all other data is at or below this level. Higher percentile measurements are normally used for performance objectives and performance measurement because they show that the large majority of the data set has achieved a particular level of performance. This can then be compared to the desired performance objective to determine the degree of success in achieving the goal. For this analysis, ESCI was most interested in the ability to respond the appropriate resources to the highest percentage of incidents. For this reason, ESCI analyzed NFIRS data from each department for 2011 and 2012 and generated average and 90`h percentile response performance for emergency incidents only. page 62 .emergcnry Semriccs Cansulang