|
Field
Abstract
|
Before
High employee turnover remains a persistent barrier to productivity in the manufacturing firms. This study investigates the roles of pre-employment information, financial growth incentives, and job assignment mechanisms in mitigating turnover and improving worker-job matches. We conduct a large-scale field experiment with approximately 2,000 newly hired workers in a Chinese electronics factory using a 3x3 cross-randomized design. In the Hiring Stage, we test the effect of information frictions and perceived career growth on selection. Applicants are randomized into: (1) a pure control; (2) an information treatment viewing a video on factory life and strict workplace regulations; or (3) a "growth path" treatment combining the video with information on position-specific allowances and tenure-based subsidies. In the Job Assignment Stage, we analyze the efficiency of managerial decision-making versus data-driven and preference-based approaches. Workers are cross-randomized into three assignment protocols where managers receive different information sets: (1) Manager Discretion (status quo); (2) AI-Assisted Assignment, utilizing an algorithm that predicts retention and productivity based on worker attributes; or (3) Preference-Based Assignment, where managers are informed of workers' stated task preferences. We measure outcomes including retention rates, productivity, and job satisfaction to determine optimal strategies for personnel management in high-turnover environments.
|
After
High employee turnover remains a persistent barrier to productivity in manufacturing firms. This study investigates the roles of pre-employment information, financial growth incentives, and job assignment mechanisms in mitigating turnover and improving worker-job matches. We conducted a large-scale field experiment with approximately 2,000 newly hired workers in a Chinese electronics factory using a 3 × 3 cross-randomized design. In the hiring stage, we test the effects of information frictions and perceived career growth on selection. Applicants are randomized into: (1) a pure control group; (2) an information treatment viewing a video on factory life and strict workplace regulations; or (3) a "growth path" treatment combining the video with information on on factory life, position-specific allowances, and tenure-based subsidies. In the job assignment stage, we analyze the efficiency of managerial decision-making versus data-driven and preference-based approaches. Workers are cross-randomized into three assignment protocols where managers receive different information sets: (1) random assignment; (2) improved matching-based assignment, utilizing an algorithm that predicts retention and productivity based on worker attributes; or (3) preference-based assignment, where managers are informed of workers' stated task preferences. We measure outcomes including retention rates, productivity, and job satisfaction to determine optimal strategies for personnel management in high-turnover environments.
|
|
Field
Trial Start Date
|
Before
March 01, 2026
|
After
March 07, 2026
|
|
Field
Last Published
|
Before
March 05, 2026 05:55 AM
|
After
March 05, 2026 07:36 AM
|
|
Field
Intervention (Public)
|
Before
We collaborate with the in-house factory of a Chinese technology firm manufacturing earphones. The factory currently faces a critical challenge: high turnover rates among general workers and a disconnect between worker capabilities and task assignment. The production line operates under a hierarchy (General Workers, Multi-skilled, Leaders), but recruitment is rapid and non-selective. Managers currently assign tasks based on limited heuristics (e.g., arrival order, gender, perceived energy), leading to potential mismatches.
To address these frictions, we propose a two-stage field experiment involving a projected cohort of 2,000 new hires. The design disentangles supply-side mechanisms (worker expectations and incentives) from demand-side mechanisms (managerial matching quality).
|
After
We collaborate with the in-house factory of a Chinese technology firm manufacturing earphones. The factory currently faces a critical challenge: high turnover rates among general workers and a disconnect between worker capabilities and task assignment. The production line operates under a hierarchy (General Workers, Multi-skilled, Leaders), but recruitment is rapid and non-selective. Managers currently assign tasks based on limited heuristics (e.g., arrival order, gender, perceived energy), leading to potential mismatches.
To address these frictions, we propose a two-stage field experiment involving a projected cohort of 2,600 new hires. The design disentangles supply-side mechanisms (worker expectations and incentives) from demand-side mechanisms (managerial matching quality).
|
|
Field
Intervention Start Date
|
Before
March 01, 2026
|
After
March 07, 2026
|
|
Field
Intervention End Date
|
Before
April 30, 2026
|
After
May 01, 2026
|
|
Field
Primary Outcomes (Explanation)
|
Before
1. Worker Retention (Tenure)
Description: The duration a worker remains employed at the factory. This is the primary measure of whether the Hiring Stage treatments (expectations/incentives) and Job Assignment treatments (AI/Preference matching) successfully reduced turnover.
Construction:
Binary Retention Indicators: A set of binary variables equal to 1 if the worker is still employed at fixed intervals (e.g., Day 7, Day 14, Day 30, Day 90) and 0 otherwise.
Tenure Duration: The total count of days from the hire date to the separation date. For workers still employed at the end of the study window, this variable will be right-censored and analyzed using survival analysis models (e.g., Cox Proportional Hazards).
2. Standardized Productivity Index (Task-Adjusted)
Description: Since workers are assigned to different tasks (some harder, some easier) and different production lines, raw output counts are not directly comparable. We construct a standardized index to compare performance across the whole sample.
3. Assignment Compliance (Managerial Adherence)
Description: This outcome measures the extent to which managers utilized the information provided in the Job Assignment treatments (AI recommendations or Worker Preferences) versus their own discretion. This helps identify the "value of information" and principal-agent friction.
Construction:
AI Compliance Rate: A binary variable equal to 1 if the worker's actual assigned task matches the task recommended by the AI algorithm (highest predicted score), and 0 otherwise.
Preference Compliance Rate: A binary variable equal to 1 if the worker's actual assigned task matches one of their "top 3" stated preferences collected during the hiring stage, and 0 otherwise.
4. Effective Labor Supply (Attendance & Overtime)
Description: Turnover is extensive margin (leaving), but productivity is also driven by the intensive margin (how much they work while employed).
Construction:
Calculated as the ratio of Actual Hours Worked (including voluntary overtime) to Standard Scheduled Hours.
This measure captures both absenteeism (ratio < 1) and willingness to work extra hours (ratio > 1), which is a proxy for worker engagement and motivation.
5. Expectation-Reality Gap (Belief Accuracy)
Description: Specifically for the Hiring Stage treatments, we measure if the video/incentive information reduced the "shock" of factory life.
Construction:
We calculate the absolute difference between Pre-entry Beliefs (surveyed at recruitment: expected hours, expected difficulty, expected earnings) and Post-entry Realization (surveyed at Week 1: actual experience).
|
After
1. Worker Retention (Tenure)
Description: The duration a worker remains employed at the factory. This is the primary measure of whether the Hiring Stage treatments (expectations/incentives) and Job Assignment treatments (AI/Preference matching) successfully reduced turnover.
Construction:
Binary Retention Indicators: A set of binary variables equal to 1 if the worker is still employed at fixed intervals (e.g., Day 7, Day 14, Day 30, Day 90) and 0 otherwise.
Tenure Duration: The total count of days from the hire date to the separation date. For workers still employed at the end of the study window, this variable will be right-censored and analyzed using survival analysis models (e.g., Cox Proportional Hazards).
2. Standardized Productivity Index (Task-Adjusted)
Description: Since workers are assigned to different tasks (some harder, some easier) and different production lines, raw output counts are not directly comparable. We construct a standardized index to compare performance across the whole sample.
3. Assignment Compliance (Managerial Adherence)
Description: This outcome measures the extent to which managers utilized the information provided in the Job Assignment treatments (AI recommendations or Worker Preferences) versus their own discretion. This helps identify the "value of information" and principal-agent friction.
Construction:
Algorithm Compliance Rate: A binary variable equal to 1 if the worker's actual assigned task matches the task recommended by the AI algorithm (highest predicted score), and 0 otherwise.
Preference Compliance Rate: A binary variable equal to 1 if the worker's actual assigned task matches one of their "top 3" stated preferences collected during the hiring stage, and 0 otherwise.
4. Effective Labor Supply (Attendance & Overtime)
Description: Turnover is extensive margin (leaving), but productivity is also driven by the intensive margin (how much they work while employed).
Construction:
Calculated as the ratio of Actual Hours Worked (including voluntary overtime) to Standard Scheduled Hours.
This measure captures both absenteeism (ratio < 1) and willingness to work extra hours (ratio > 1), which is a proxy for worker engagement and motivation.
5. Expectation-Reality Gap (Belief Accuracy)
Description: Specifically for the Hiring Stage treatments, we measure if the video/incentive information reduced the "shock" of factory life.
Construction:
We calculate the absolute difference between Pre-entry Beliefs (surveyed at recruitment: expected hours, expected difficulty, expected earnings) and Post-entry Realization (surveyed at Week 1: actual experience).
|
|
Field
Experimental Design (Public)
|
Before
This study employs a 3x3 cross-randomized design to disentangle the drivers of factory turnover and productivity through two distinct stages. In the Hiring Stage, recruits are randomized into a control group, a "Realistic Job Preview" group viewing a video on strict factory life to align expectations, or a "Growth Path" group receiving the video plus details on long-term wage incentives. Subsequently, in the Job Assignment Stage, managers assign these workers to tasks using one of three protocols: standard discretion, an AI-assisted method utilizing algorithmic predictions of retention and productivity, or a Preference-based method incorporating workers' stated task desires. By interacting supply-side information treatments with demand-side matching mechanisms, the study tests whether retention is better improved by screening for fit and incentives or by optimizing the manager's assignment decisions through data and worker voice.
|
After
This study employs a 3x3 cross-randomized design to disentangle the drivers of factory turnover and productivity through two distinct stages. In the Hiring Stage, recruits are randomized into a control group, a "Realistic Job Preview" group viewing a video on strict factory life to align expectations, or a "Task-specific Compensation Incentive" group receiving the video plus details on long-term wage incentives. Subsequently, in the Job Assignment Stage, managers assign these workers to tasks using one of three protocols: random assignment, an AI-assisted method utilizing algorithmic predictions of retention and productivity, or a Preference-based method incorporating workers' stated task desires. By interacting supply-side information treatments with demand-side matching mechanisms, the study tests whether retention is better improved by screening for fit and incentives or by optimizing the manager's assignment decisions through data and worker voice.
|
|
Field
Sample size (or number of clusters) by treatment arms
|
Before
2000 newly hired workers
|
After
2600 newly hired workers
|
|
Field
Intervention (Hidden)
|
Before
The study employs a 3x3 cross-randomization design. Every recruit is independently randomized into one of three Hiring Treatments and subsequently into one of three Job Assignment Treatments.
1. Stage 1: The Hiring Stage (Information & Expectations)
This stage targets the workers' information set before they fully commit to the production line.
1.1 Hiring Arm 1: Pure Control
Recruits go through the standard hiring process: brief interview, basic screening for criminal records/health, and standard offer extension. No additional information is provided.
1.2 Hiring Arm 2: Realistic Job Preview (Video Info)
Recruits view a specialized video introduction to "Factory Life."
Content: The video explicitly details the daily schedule (morning and evening shifts), the strict "no phone" policy during work hours, and visual demonstrations of various tasks.
Goal: To reduce information asymmetry and align expectations regarding work intensity and discipline.
1.3 Hiring Arm 3: Realistic Job Preview + Growth Path (Video + Financial Incentives)
Recruits view the same video as in Arm 2.
Addition: They are explicitly presented with the "Growth Path" structure, detailing specific subsidies for different positions (position allowances) and the schedule of tenure-based subsidies (wage increases over time).
Goal: To test if highlighting the financial returns to tenure and skill acquisition changes the pool of retained workers and increases the option value of staying.
2. Stage 2: The Job Assignment Stage (Matching Efficiency)
This stage targets the manager's decision-making process. In all arms, the manager retains the final authority to assign the worker, but the information set provided to the manager varies.
2.1 Assignment Arm 1: Manager Discretion (Control)
Managers assign workers to specific stations/tasks based on the status quo method (observables like gender, physical appearance, or random arrival order).
2.2 Assignment Arm 2: AI-Assisted Assignment
We develop an AI algorithm trained on historical data (current employees' resumes, demographics, productivity logs, and turnover records).
For new hires, the AI analyzes their resume, work history, and personal characteristics to generate a predicted score for retention likelihood and productivity for specific tasks.
Intervention: This prediction and a recommended assignment are provided to the manager at the moment of the assignment decision.
2.3 Assignment Arm 3: Preference-Based Assignment
During the Hiring Stage, we elicit "Stated Preferences" from all workers regarding which tasks or positions they desire (e.g., based on perceived difficulty, interest, or status).
Intervention: These specific preferences are explicitly provided to the manager to inform their assignment decision.
|
After
The study employs a 3x3 cross-randomization design. Every recruit is independently randomized into one of three Hiring Treatments and subsequently into one of three Job Assignment Treatments.
1. Stage 1: The Hiring Stage (Information & Expectations)
This stage targets the workers' information set before they fully commit to the production line.
1.1 Hiring Arm 1: Pure Control
Recruits go through the standard hiring process: brief interview, basic screening for criminal records/health, and standard offer extension. No additional information is provided.
1.2 Hiring Arm 2: Realistic Job Preview (Video Info)
Recruits view a specialized video introduction to "Factory Life."
Content: The video explicitly details the daily schedule (morning and evening shifts), the strict "no phone" policy during work hours, and visual demonstrations of various tasks.
Goal: To reduce information asymmetry and align expectations regarding work intensity and discipline.
1.3 Hiring Arm 3: Realistic Job Preview + Task-specifc Incentive (Video + Financial Incentives)
Recruits view the same video as in Arm 2.
Addition: They are explicitly presented with the "Task-specifc Compensation Incentive" structure, detailing specific subsidies for different positions (position allowances) and the schedule of tenure-based subsidies (wage increases over time).
Goal: To test if highlighting the financial returns to tenure and skill acquisition changes the pool of retained workers and increases the option value of staying.
2. Stage 2: The Job Assignment Stage (Matching Efficiency)
This stage targets the manager's decision-making process. In all arms, the information set provided to the manager varies.
2.1 Assignment Arm 1: Random assignment
Managers assign workers to specific stations/tasks based on the random assignment created by the research team.
2.2 Assignment Arm 2: Improved-Matching assignment
We develop an algorithm trained on historical data (current employees' resumes, demographics, productivity logs, and turnover records).
For new hires, the algorithm analyzes their resume, work history, and personal characteristics to generate a predicted score for retention likelihood and productivity for specific tasks.
Intervention: This prediction and a recommended assignment are provided to the manager at the moment of the assignment decision.
2.3 Assignment Arm 3: Preference-Based Assignment
During the Hiring Stage, we elicit "Stated Preferences" from all workers regarding which tasks or positions they desire (e.g., based on perceived difficulty, interest, or status).
Intervention: These specific preferences are explicitly provided to the manager to inform their assignment decision.
|