Transcript Class 24 - University of Virginia

Class 24
Case: Edgcomb Metals
Edgcomb Metals
• 21 service centers doing $500 million in sales
• Supplied by large steel companies
• 35,000 customers (any firm using some kind of
steel)
• Services include cutting, shaping, and daily
delivery.
Troy Plant
• 72K plant serving customers in Virginia
• Seven trucks, eight trailers
• Seven drivers
– $9.50 and hour with 50% for overtime
– Customized uniforms with American flag patch
• Daily deliveries to seven “sectors”
– Product for each “run” was loaded at the plant each
morning in “optimum” order.
– Customers unload.
– Drivers assigned “randomly” to sectors.
Spencer versus Williams
• Spencer spoke up at the most recent drivers’
meeting:
– We get paid by the hour with time and a half for
overtime.
– Some of us hustle throughout the day, finish early,
and help in the shop.
– Some of us don’t hustle and end up with overtime
as a result.
– I want to work hard…..and it’s not right that
others get rewarded for NOT working hard.
Exhibit 2
Month
Jun-83
May-83
Apr-83
Mar-83
Feb-83
Jan-83
Dec-82
Nov-82
Oct-82
Sep-82
Aug-82
Jul-82
TONS
3043
2889
2384
2500
2312
2678
1678
2209
2382
2315
2624
1745
MILES
34907
38799
33367
35288
29876
35056
27171
29917
30143
34771
36523
34693
STOPS REG HRS OT HRS
719
1182
362
728
1259
360
695
1230
382
763
1345
283
613
1205
257
709
1253
306
568
962
187
624
915
276
713
962
239
617
1091
272
724
1108
283
640
1030
249
Exhibit 4 Williams
Miles(W)
331
206
221
193
Stops(W) Hours(W)
3
10.17
2
8.00
4
8.25
4
10.00
.
.
.
.
.
.
.
.
.
261
292
200
292
7
3
7
4
11.40
10.75
10.50
11.33
Exhibit 4 Spencer
Miles(S) Stops(S) Hours(S)
182
3
8.25
364
9
12.50
227
9
10.50
188
4
6.50
.
.
.
.
.
.
.
.
.
346
320
181
369
9
9
9
7
10.00
11.50
9.50
11.75
Notes
Task A
• Calculate summary statistics for both the
Williams and Spencer Data.
– Be prepared to present and comment BRIEFLY on
the results.
Task B
• Test the hypothesis that mean hours is equal
for S and W.
– Formulate you own alternative hypothesis
– Do not use regression
– Be prepared to report and interpret the results.
Task C
• Combine (Stack) the Williams and Spencer
Data.
• Create a dummy variable which designates
driver
• Regress hours on the dummy variable.
• Be prepared to interpret the results and to
test the statistical significance of the results.
Task D
• Test the hypothesis that Spencer’s mean miles
(per run) is equal to Williams’ mean miles.
• Ha: mean miles for Spencer is greater than
mean miles for Williams.
• Be prepared to interpret the results.
Task E
• Test the hypothesis that Spencer’s mean Stops
(per run) is equal to Williams’ mean Stops.
• Ha: mean Stops for Spencer is greater than
mean Stops for Williams.
• Be prepared to interpret the results.
Task F
• For the Williams Data
– Regress Hours on both Miles and Stops.
– Be prepared to report and interpret the results.
• For the Spencer Data
– Regress Hours on both Miles and Stops.
– Be prepared to report and interpret the results.
• Based on your comparison of the two models,
who is the better driver?
Task G
• Combine (Stack) the Williams and Spencer
Data.
• Create a dummy variable which designates
driver.
• Regress hours on the dummy variable, Miles,
and Stops.
– A multiple regression with three X variables.
• Be prepared to interpret the results and to
test the statistical significance of the results.
Edgcomb Metals
What Happened