Economic Decision-Making and Job Satisfaction of Delivery Workers

2020-10-27

2020-12-23

Uncertainty preferences;
Time preferences;
Social preferences;
Cognitive abilities;
Job satisfaction;
Job search intensions.

Uncertainty preferences include measures of risk aversion (the MPL-elicited certainty equivalence and self-reported risk-taking), ambiguity aversion, loss aversion;
Time preferences include measures of patience between payoffs today versus payoffs 3 months later, as well as that between payoffs 3 months later and 6 months later, and both the dummy indicating present bias and the extent of present bias;
Social preferences include a continuous variable of sharing in a standard DG, and measures of altruism and efficiency concerns derived from choices of several modified mini-DG choices;
Cognitive abilities include measures from 3 cognitive questions;
Job satisfaction is measured by a five-point scale, for the job overall, for the aspect of work load/intensity, as well as for the aspect of wage income;
Job search intension is measured by a five-point scale regarding the likelihood of searching for a different job within a year within the company, outside the company but within the delivery industry, and outside the industry, as well as investing time and money in skill training for jobs other than the current one.

Work load in the past 20 days; stress levels in the past 20 days; stress level at the current moment when answering the survey; levels of positive and negative affect in the past 20 days; mood on that day; wage income most recently received; liquidity constraint in the recent time.

Work load in the past 20 days, both the maximum and average levels, are either self-reported or from the administrative data, if we have get access to it.
Stress levels in the past 20 days are measured using both the Perceived Stress Scale (PSS) and the work-related part of the Trier Inventory for Chronic Stress (TICS) scale;
Stress level at the current moment when answering the survey is measured using a self-reported stress scale from 0 to 100;
Levels of positive and negative affect in the past 20 days are measured by the Positive Affect Negative Affect Scale (PANAS);
Mood on that day is measured by a simple five-point mood scale.
Wage income most recently received is either self-reported or from the administrative data, if we have get access to it.
Liquidity constraint in the recent time is measured by two questions regarding recent needs for credit and the difficulty in raising a certain amount of cash within a week.

We send out the survey experiment at different time.

In this study, we elicit delivery workers' economic preferences and cognitive abilities via an online survey experiment. We randomize the timing workers first receive the survey link: while 1/3 of the workers in our sample answer the survey before the Double-11 festival, a major shopping season in China, another 1/3 answer the survey after the festival but before receiving the corresponding wage income, and the last 1/3 answer the survey after receiving the Double-11 festival wage income. While our main design is a between-subject comparison, we also send a second-time survey at the third time point for those in the first group, so that we also have a within-subject comparison design embedded for this subsample.

Moreover, we also randomize whether the participant receive information regarding the officially reported average wage income of those employed in Tianjin in year 2018 (the latest year available), which is orthogonal to the timing treatments. Therefore, we have in total 2X3=6 treatment groups.

We access a roster of all employees at the Tianjin sorting center of a delivery company. We randomize by assigning a participant code (1-381) to each of the employees, selecting those with a participant code that equals 3n+1 to the Before Group, those with a participant code that equals 3n+2 to the After-1 Group, and those with a participant code that equals 3n+3 to the After-2 Group, where n=0,1,2,...,126.

The randomization of city average wage income information provision is implemented using the “randomized questions” feature of the Tencent Survey website (wj.qq.com). There are two versions of the questionnaire: one with the average wage information, and the other without; the website randomly selects one version when presenting the questionnaire to the participant.

Individual.

No

1 delivery sorting center of a company.

We will send our survey to all the more than 360 employees at the sorting center. We expect the response rate to be above 90%.

110 individuals in the Before Group, 110 in the After-1 Group, and 110 in the After-2 Group.
165 individuals receiving the city average wage income information, and 165 individuals not receiving.