An Ounce of Intervention is Worth a Pound of Cure

1
An Ounce of Interventionis Worth a Pound of
Cure

• Ann B. Pudoff
• Introduction to Probability and Statistics
• Dr. Stanley Barrick December 16, 1999

2
Study
In 1995, I implemented an ergonomic program at
one of the locations in my corporation. This
location consisted of approximately 400 male and
female employees that spend 4 to 8 hours per day
performing data entry. The ergonomic program
consisted of training employees on basic
ergonomic principles, performing workstation
evaluations for employees that were experiencing
signs or symptoms of repetitive motion injuries
(RMIs) and following through with recommended
improvements. Supervisors were trained to be
aware of signs and symptoms of RMIs, and to
respond immediately observed.
3
Study (cont.)

Early intervention should prevent symptoms from
becoming costly injuries. The following
statistical data was taken after two years
following the implementation of the ergonomic
program (1997 and 1998).
4
Question?
Will a proactive Ergonomic Program lower
workers compensation costs?
5
Hypothesis

• Ho µ 1 µ 2 Ha µ
  1 gt µ 2
• µ 1 average cost of RMI in 1997
• µ 2 average cost of RMI in 1998
• The null hypothesis (Ho) assumes that there
  will be no change in average cost of RMI from
  1997 to 1998 after implementing an Ergonomic
  Program.
• The alternate hypothesis (Ha) assumes that
  their will be a decrease in average cost of RMI
  from 1997 to 1998.

6
Hypothesis Test
Ho µ 1 µ 2 ? .05 Ha µ
1 gt µ 2 Critical Value 1.645
Reject null hypothesis
Do not reject null hypothesis
.05 ?
t 1.645 (critical value)
0
7
Data
Does the data provide sufficient evidence at the
5 significance level showing no significant
change in average cost of RMI from 1997 to 1998?
Based upon the assumptions listed below, I chose
to use the Pooled t-Test for two population
means.

• Sample size small (lt30)
• Independent samples
• Normal populations
• Equal population standard deviations
• (SL/Sslt2) 22,245/12,876 1.7

8
Statistics
1997 X1 17,858 S1 22,245 n1 17
1998 X2 7,447 S2 12,876 n2 23
9
Calculations
where
Sp (16)(494840025) (22)(165791376)
1723-2
t 1.866 Sp 17,445
10
Results
Reject null hypothesis
Do not reject null hypothesis
t 1.866
.05 ?
t 1.645 (critical value)
REJECT THE NULL HYPOTHESIS
The data shows that at a 5 significance level,
there is a decrease in average cost per RMI from
1997 to 1998. NOTE p.035 ? p lt ? (.05), so
we reject the null hypothesis
11
Observations
The fact that the number of claims increased
can be expected due to an increase in
awareness.
12

Suggestions for Further Study
I. Collect data for another year and see if
frequency ( of claims) decreases. II.
Collect data for another year and see if severity
(direct costs) continues to decrease.
III. Conduct a study that compares job
satisfaction to workers compensation
costs.