iSixSigma

Selecting the Right Test

Six Sigma – iSixSigma Forums General Forums General Selecting the Right Test

Viewing 2 posts - 1 through 2 (of 2 total)
  • Author
    Posts
  • #54023

    Lisa FaJohn
    Participant

    I am trying to determine if there is a correlation of injuries to the amount of overtime worked by an employee. The data I have collected is the I have collected data on employees with injuries, as well as, employees without injuries. My data consists of the Business Unit (Construction & Operations), Date (Year 2008-2011/Quarter), total hours worked 7 days prior to the injury, 14 days prior to the injury and 30 days prior to the injury. For employees with no injuries I have collected a random sample of dates for each year/quater, as well as, total hours worked 7 days prior to the sample date, 14 days prior to the sample date and 30 days prior to the sample date.

    0
    #193008

    Robert Butler
    Participant

    The simplest approach would be a two way ANOVA with the interaction term. You would run multiple pairwise comparisons with corrections (Bonferroni, Tukey-Kramer, etc). How many comparisons would depend on your interests. For the first iteration it sounds like all you care about are significant overtime hour counts at 7, 14, and 30 days between injured and non-injured. The only major concern with ANOVA would be the issue of unequal population variances. If the test for unequal variance came back significant you could run the means comparisons using Welch’s test. This analysis would be across business units and there would be no differentiation with respect to time of year.

    If business unit is also an issue then the above could be expanded to either include business units or to run the same analysis on business units separately.

    0
Viewing 2 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic.