GB891 Unit 1 Protocol Document Example

Everything Behind One Monday Sheet: A Unit 1 Protocol Document for a Supervisor Safety Feedback Study

[Student Name]

Purdue University Global

GB891: Doctoral Project II: Project Design and Implementation

Unit 1 Protocol Document

[Instructor Name]

[Date]

Corliss Baking Company, its plants, crews and every procedure are composite and illustrative; the methods sources cited are real, and no review board has acted.

Section 1: Roles

The candidate is Corliss's corporate safety manager. The candidate designs the study, trains observers, analyzes data and writes the reports, but never sees an individual associate's answers.

The research assistant is a graduate student hired for the study, reporting to the candidate's doctoral chair rather than to anyone at Corliss. The assistant collects all pulse answers, holds the key linking crew codes to crews and prepares the feedback sheets.

Two trained observers, contracted from a regional occupational safety consultancy, conduct structured observations at the four study plants.

Twelve intervention supervisors, six at each of the two intervention plants, Ashford and Brierly, receive weekly feedback. The twelve supervisors at the comparison plants, Calder and Dunmore, receive none until the study ends.

Section 2: The Twelve-Week Schedule

Week zero. The baseline climate survey is administered to all 24 study crews on all three shifts. Intervention supervisors attend a sixty-minute briefing that explains the feedback sheet, the goal of raising safety in routine exchanges and the rule that no sheet will be used in performance reviews.

Weeks one through twelve. The weekly cycle in Section 3 runs at the intervention plants. Pulse questions are also asked at the comparison plants, with answers recorded but no sheets delivered, so that the two sets of plants can be compared on the same measure.

Week twelve. The follow-up climate survey is administered to all 24 crews.

Week thirteen. The research assistant interviews each intervention supervisor about the feedback.

Section 3: The Weekly Cycle

Step 3.1. On Wednesday and Thursday, the research assistant draws four members of each crew at random from that week's shift roster, using a seeded random number list prepared before the study.

Step 3.2. At the crew's scheduled break, the assistant asks each drawn member three pulse items, in English or Spanish as the associate prefers. The first asks how many times in the past week the supervisor raised safety outside the start-of-shift meeting. The second asks whether the supervisor mentioned a specific hazard on the associate's own station. The third asks whether the associate would feel comfortable reporting a near miss to that supervisor. The items take under three minutes.

Step 3.3. If fewer than four crew members are present, the assistant asks everyone present who has not been asked that week. If fewer than three answer, no score is reported for that crew that week, so that no supervisor can infer an individual's answer.

Step 3.4. Scores close on Friday at 5 p.m. Over the weekend, the assistant prepares a one-page sheet for each intervention supervisor showing the crew's average on each item that week and the previous three weeks.

Step 3.5. By Monday at 10 a.m., the assistant emails each sheet directly to its supervisor. No copy goes to plant managers or to the candidate.

Section 4: Observation

Observers conduct fifteen-minute samples at five fixed stations in each plant: slicer guarding, jam clearing on the packaging line, mixer lockout, oven loading and pallet movement at the dock. Each observer completes two samples per station per week at each plant, recording for each fifteen-second interval whether the required safe practice was performed, using definitions in Appendix C. Every fourth sample is conducted jointly by both observers, and their agreement is calculated with the kappa statistic Cohen (1960) introduced, interpreted against the conventional benchmarks of Landis and Koch (1977).

Step 4.1. When an observer must stop. If an observer sees an associate about to be injured, such as a hand moving toward an unguarded blade, the observer stops observing and intervenes immediately, by calling out or signaling the line lead. The observer records the event, without names, on the decision table in Appendix E, and the sample is discarded. Safety takes precedence over data in every case.

Section 5: Fidelity Checks

The protocol records three things each week to show whether the intervention was delivered as designed: whether each sheet was sent by Monday at 10 a.m., whether each supervisor opened it, using a read receipt, and whether supervisors' pulse scores changed. Carroll et al. (2007) describe fidelity as adherence to the intervention's content, frequency and duration, and these records address all three.

Every step in the weekly cycle exists to get one accurate sheet to one supervisor without letting that supervisor learn who said what.

Section 6: Changes to the Protocol

Any change after the study begins goes into the deviation log, dated, with the reason and who approved it, and reported in the implementation chapter beside the result it affects.

Appendices

Appendix A holds the pulse scripts in English and Spanish. Appendix B shows the feedback sheet. Appendix C holds the observation form and definitions. Appendix D holds both versions of the climate survey. Appendix E holds the decision table for irregular situations, such as a crew merged for a shift or a supervisor on leave.

References

Carroll, C., Patterson, M., Wood, S., Booth, A., Rick, J., & Balain, S. (2007). A conceptual framework for implementation fidelity. Implementation Science, 2, Article 40. https://doi.org/10.1186/1748-5908-2-40

Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20(1), 37-46. https://doi.org/10.1177/001316446002000104

Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159-174. https://doi.org/10.2307/2529310