Probability of Time Loss and Non-time Loss Accidents

From the table below, I am trying to evaluate the probability of the following: time loss, non-time loss and fatalities. How could I accomplish it please?


Year Work Orders Months Unsafe Incidents Time-Loss Non-Time Loss Fatalities
Year 1 8,082 12 3 0 1 0
Year 2 9,000 12 2 0 0 0
Year 3 10,766 12 11 0 0 0
Year 4 11,533 12 8 0 0 0
Total 39,381 48 24 0 1 0

Thank you.

I appreciate very much your help.

Originally Posted by shg

       ----B----- -----C----- --D--- -------E-------- ----F---- ------G------ ----H-----
   1                                                                                    
   2   Confidence         50%                                                           
   3                                                                                    
   4      Year    Work Orders Months Unsafe Incidents Time-Loss Non-Time Loss Fatalities
   5   Year 1           8,082     12                3         0             1          0
   6   Year 2           9,000     12                2         0             0          0
   7   Year 3          10,766     12               11         0             0          0
   8   Year 4          11,533     12                8         0             0          0
   9   Total           39,381     48               24         0             1          0
  10                                                                                    
  11   High Limit                                     (none)           28,407 (none)    
  12   Low Limit                                         56,815        14,625     56,815

You can't calculate probability, but you can calculate the upper and lower limits of a confidence interval for some specified confidence.

In F11 and copy across,

= IF(F9 = 0, "(none)", 2 * $C$9 / CHIINV(Confidence / 2, 2 * F9) )

That says there is no upper limit to the confidence interval that a time loss accident will occur, as none have occurred to date.

In F12 and copy across,

=IF(F9 = 0, $C$9 / -LN(1-Confidence), 2 * $C$9 / CHIINV((1-Confidence) / 2, 2 * (F9 + 1) ) )

That says that with 50% confidence (the value in C2), the low limit of a time loss accident is 56,815 work orders.

---------- Post added at 02:20 PM ---------- Previous post was at 02:19 PM ----------

Super! Your calculations point me the right direction. Thank you.

This is really more of a statistics problem than an excel problem? What I mean is this board isn't really the right place to ask this question--you probably want to go ask some statisticians.

Based on what you've got I've done some simple things and attached them.

Anyway, there's only four data points, but I noticed the relationship between work order and incidents is not linear; ie, the probability of an event doubles or triples once you cross about 10,000 work orders. If I had to hypothesize, then I'm going to guess the push to meet production leads to safety getting compromised.