GB892 Unit 10 Results Chapter Draft Example

Two Questions, One Chapter: A Unit 10 Results Chapter Draft on Dock Appointments

[Student Name]

Purdue University Global

GB892: Doctoral Project III: Data Analysis

Unit 10 Results Chapter Draft

[Instructor Name]

[Date]

Haverly Logistics, its centers, participants and every result are composite and illustrative; the methods sources cited are real.

Introduction

This chapter reports the findings of a mixed methods study of an online dock appointment portal adopted by six of Haverly Logistics' twelve distribution centers in month ten of an eighteen-month period. The first research question asked whether waits at adopting centers changed relative to comparison centers. The second asked why waits persisted at centers where appointments existed. Section 4.1 describes the data, Section 4.2 answers the first question, Section 4.3 the second, and Section 4.4 sets the two strands side by side.

4.1 The Data

The yard system extract contained 288,061 inbound gate records. After five exclusion rules applied in a fixed order, removing drop-trailer visits, duplicate scans, missing departures, auto-filled departure times and implausible dwell, 196,518 live-unload visits remained. Excluded records were split between center groups in nearly the same proportions as retained records but were more often night arrivals. The full preparation record appears in Appendix A.

Descriptive statistics show median dwell at adopting centers of 126 minutes before adoption and 117 after, and at comparison centers of 118 and 114. Median minutes waiting for a door fell from 44 to 37 at adopting centers and from 38 to 34 at comparison centers, while unloading time held at 47 to 48 minutes in every group and period. Full descriptive tables appear in Appendix B.

4.2 Research Question 1: Did Waits Change?

Departure noted here. The proposal specified dwell in minutes as the outcome. Because dwell was skewed, the analysis used log minutes; the planned model's estimate is reported below for comparison. Inference used a wild cluster bootstrap rather than conventional clustered errors, because twelve centers are too few for conventional errors to be reliable (Cameron et al., 2008).

The primary estimate. Relative to the comparison group, the adopting centers' dwell dropped by roughly 4 percent, with the bootstrap interval running from 1 percent to 7 percent; at the 126-minute baseline median that comes to about 5 minutes a visit. Fitted in raw minutes as first planned, the estimate rises to 6.2 minutes, bounded by 2.3 and 10.0, because the longest waits weigh more heavily on a mean.

The secondary outcome. Departure noted here. The proposal named detention charges, which could not be linked to visits; the share of visits over two hours was used instead. That share dropped 3.2 points, an estimate whose interval stretches from just above zero to just over six points.

Segments and the negative control. Door wait fell by about 11 percent. Unload time, which appointments should leave alone, barely moved and its interval included zero, consistent with the interpretation that the estimate reflects appointments rather than another change (Lipsitch et al., 2010).

Sensitivity. Departure noted here. Dropping DC-09, whose own staffing changed early in the pre-period, or dropping DC-03 each nudged the estimate to about 5 percent. The proposal had planned a site ranking; it was removed because each site estimate would rest on a single treated center.

4.3 Research Question 2: Why Do Waits Persist?

Seventeen interviews, eleven with dock supervisors and six with carrier dispatchers, were analyzed with a codebook and reflexive thematic analysis (Braun & Clarke, 2019). Departure noted here. Twenty-four interviews were planned; dispatchers were the group most underrepresented.

Four themes were constructed. In the first, appointments scheduled trucks but not crews, so booked trucks waited when their slots fell between shifts or during breaks. In the second, carriers arrived early on purpose, because early trucks were sometimes served ahead of schedule. In the third, floor-loaded freight took far longer than the standard slot assumed and delayed later trucks at the same door. In the fourth, paperwork holds after unloading were untouched by appointments.

Accounts resisting the themes are reported in full. Two dispatchers said unbooked trucks now waited longer than before, and one supervisor at a center with continuous crew rotation said nothing had changed.

4.4 The Strands Side by Side

The joint display sets each theme beside the yard field linked to it in the codebook and reports what the field shows, following the approach Guetterman et al. (2015) describe. For the crew-slot theme, booked trucks arriving on time in the half hour before a shift change waited a median of 49 minutes for a door, against 26 at other times. For the early-arrival theme, early trucks waited a median of 58 minutes for a door, against 28 for on-time trucks, confirming the theme in the opposite way carriers intended: arriving early did not, on average, get them served sooner. For the floor-loaded theme, floor-loaded visits occupied doors about twice as long, and the next booked truck at the same door waited longer. For the paperwork theme, the interval after unloading did not change with adoption, consistent with supervisors' accounts. The walk-up account was confirmed: unbooked trucks at adopting centers waited a median of 52 minutes for a door after adoption.

The interviews explained where the remaining minutes went, and the yard data confirmed most of the explanations and corrected one.

How the Departures Bear on the Findings

None of the departures noted above changes the direction of a finding. Logging the outcome and switching to a bootstrap made the primary estimate more conservative than the planned model. Replacing detention charges with the two-hour share means the study speaks to exposure to detention, not to its cost. Dropping the site ranking means the chapter reports how the portal performed across the adopting centers together rather than which center did best. And the shortfall in dispatcher interviews means the early-arrival theme, which rests heavily on dispatchers, should be read with more caution than the crew-slot theme, which rests on supervisors.

Summary

The portal centers cut waits by roughly 4 percent more than the others did, near 5 minutes a visit, almost entirely through shorter waits for a door. The persistent waits are explained by crews scheduled by shift, early arrival, floor-loaded freight and paperwork holds, and the yard data support each explanation. Early arrival, which carriers described as a strategy, did not shorten their waits. Unbooked trucks bear part of the cost of appointments for booked ones. Chapter 5 discusses what these findings mean for Haverly's dock scheduling, for the carriers who use its portal and for the six centers that have not yet adopted the portal.

References

Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic analysis. Qualitative Research in Sport, Exercise and Health, 11(4), 589-597. https://doi.org/10.1080/2159676X.2019.1628806

Cameron, A. C., Gelbach, J. B., & Miller, D. L. (2008). Bootstrap-based improvements for inference with clustered errors. The Review of Economics and Statistics, 90(3), 414-427. https://doi.org/10.1162/rest.90.3.414

Guetterman, T. C., Fetters, M. D., & Creswell, J. W. (2015). Integrating quantitative and qualitative results in health science mixed methods research through joint displays. Annals of Family Medicine, 13(6), 554-561. https://doi.org/10.1370/afm.1865

Lipsitch, M., Tchetgen Tchetgen, E., & Cohen, T. (2010). Negative controls: A tool for detecting confounding and bias in observational studies. Epidemiology, 21(3), 383-388. https://doi.org/10.1097/EDE.0b013e3181d61eeb