Research On The Influence Factors Of Task Performance In Crowdsourcing Contests

Crowdsourcing has attracted much attention because it has a competitive advantage over traditional work methods in mobilizing distributed labor and utilizing innovation.With the development of the Internet,network crowdsourcing has become the main way of open innovation.Crowdsourcing contest is one of the most popular and effective crowdsourcing models.It involves three elements: task dispatcher,answerer and crowdsourcing platform.The dispatcher mostly concerns about the task performance.How to set various factors reasonably to attract many answerers to participate in the task and submit sufficient and high-quality task solutions,so as to maximize the task performance is the core of the dispatcher's attention.Existing research in the field of economic management often measures performance from both behavior and results,while existing crowdsourcing contest research mainly starts from task results,and rarely uses the process and behavior of answerers to measure performance.At the same time,existing studies often only measure performance from one aspect of the task result—the quantity and quality of the solution,and rarely measure both.In addition,there are many factors that affect task performance.Existing studies often focus on only a part of the factors,and lack comprehensive systematic research.Among them,there are some differences in the existing research conclusions about the impact of task rewards and task duration on performance;task descriptions are often measured only by description lengths,and text mining methods are rarely used to analyze the textual characteristics of task descriptions.This study measures the performance of crowdsourcing contests from the perspectives of both results and behavior.The quantity and the quality of the solutions are selected from the results perspective.According to the goal setting theory,this study divided the crowdsourcing contests into two parts: task design and task execution,which include four factors: task design,task description,task process,and competitive environment,and analyzed their impact on task performance.Objective data of 2,545 completed tasks from the Taskcn crowdsourcing website were used as research samples.STATA software and LIWC text analysis tools wereused to conduct empirical analysis and discussion of the 39 hypotheses proposed in the study.There are four main innovations in this research:First,the study found that there is an inverted U-shaped relationship between the effects of task rewards and task duration on task performance in task designs,which explains the differences in existing studies.It is found that there is an inverted U-shaped relationship between the task rewards and the quantity and quality of the solution.According to the expected value theory,when the task rewards is relatively low,as the task rewards increases,the expected return of the answerer increases,the number of participants increases,the amount of effort invested increases accordingly,and the quantity and quality of the submitted solutions will increase accordingly;according to the theory of self-efficacy,continuing to increase the rewards to a certain level will intensify competition and affect the expectation of the answerer,and will inhibit the degree of participation and effort.The quantity and quality of the solutions will decrease.It is found that there is an inverted U-shaped relationship between the task duration and the quantity of solutions.As the task duration increases,the answerer has more time to choose to participate in the task,and the submission increases.After the task duration continues to increase to a certain extent,the longer the payback period,the higher the time cost,and the inhibition of participation,will make the quantity of solutions decrease.It is found that there is a linear negative relationship between the task duration and the quality of solutions.The increase of the duration leads to an increase of opportunity cost,the answerer will not invest too much energy and effort because of the longer task duration.Second,the study found that description length positively affected task performance,and description attachments negatively affected task performance,and this deepens the study of the impact of descriptive characteristics on performance.Text analysis was used to extract such eight textual features as the readability,psychological distance,and cognitive complexity markers,activation levels,causality,humanized details,normative information,and language intensity,and whose effects are investigated to fill the gaps in this area.By dividing the task description into description features and text features,the study found that the length of the description positively affects the quantity and quality of the solution.The longer the length ofthe task description,the more clearly the task requirements can be communicated,and the number of participants and the quality of the solutions will increase.It is found that the description attachment negatively affects the quantity and quality of the solution.Adding pictures or document attachments when the dispatcher describes the task will increase the uncertainty of the information,increase the difficulty for the answerer to understand the task,and reduce the task performance.It is also found that the task description text contains more cognitive complexity markers,higher activation levels,more humanized details,and more normative information will increase the quantity of solutions;more causality is stated in the description text will reduce the quantity of solutions;the more descriptive text and normative information in the description text will increase the average quality of solutions;the poorer the readability of the description text,the greater the psychological distance,and the higher the language intensity,the lower the average solutions quality.Thirdly,it studied the impact of task process and competitive environment on task performance,and found that task communication positively affected the quantity of solutions,negatively affected the solution quality,task density and market price had a negative impact on task performance,which enriches the research content of task performance impact.The study found that the communication during the task process will increase the quantity of solutions,but will reduce the quality of solutions.The supplementary explanation,price increase and duration extension have no significant effect on the solutions quantity.The price increase and duration extension has a negative impact on the quality of solutions,explaining that the dispatcher's intervention in the task process often has an adverse effect on performance.It was found that the task density negatively affected the quantity of solutions and positively affected solutions quality.The higher the task density,the more the number of tasks in a given time,and the fewer the answerers who share the same task,the lower the probability of the answers winning and obtaining the expected benefits,and the quantity of solution submissions will decrease;when the task density is higher,the lower self-efficacy answerers will be squeezed out,the average quality of the solution will be improved,but this improvement does not improve overall performance.The higher the market price,the higherthe opportunity cost for the answerer to participate in the task,and the probability of choosing to participate in the task will be reduced,as a result,the quantity and quality of the solutions will decrease.Fourth,the average participation time of answerers was introduced into the study to measure performance,which filled the gap in measuring the performance of crowdsourcing contests from a behavioral perspective.Participation time reflects the degree of effort that the answerer puts in the task execution process,and the degree of effort will directly affect the final task performance.The study found that task settings,task descriptions,task processes,and competitive environment all have varying degrees of impact on the time spent by the answerers,which in turn affects performance.At the same time,the study analyzed and compared the differences in solutions quantity,solutions quality and participation time to measure task performance,and found that the use of participation time can effectively measure task performance.