{"title":"在不同工程任务中的信心和表现","authors":"C. MacRobert","doi":"10.1680/jfoen.23.00001","DOIUrl":null,"url":null,"abstract":"To explore how different tasks inform decisions, geotechnicians were invited to predict the stability of a failed tailings dam. Participants included students, contractors and experienced consultants. The first task involved noting down stability concerns and making a stability prediction. Participants then prepared sketches, suggested material parameters and updated their predictions. The third task involved suggesting strength parameters for a limit equilibrium model, critiquing the model and making a final stability prediction. A 30 min time limit increased the chance of human error. Accurately assessing the failure mode at the note and sketch sections did not translate into correct predictions, despite high confidence in these tasks, particularly from experienced participants. The majority (80%) of final stability predictions were highly correlated with the analytical computer model, despite low confidence in this model, particularly from experienced participants. While this trend in confidence is perhaps expected, most participants, including experienced participants, failed to identify that failure was imminent and were confident in their final predictions. This was largely because many suggested drained strength parameters, rather than undrained strength parameters appropriate for the problem. This highlights the importance of intuitive tasks such as note taking and sketching to understand problems before building analytical models.","PeriodicalId":42902,"journal":{"name":"Proceedings of the Institution of Civil Engineers-Forensic Engineering","volume":null,"pages":null},"PeriodicalIF":0.4000,"publicationDate":"2023-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Confidence and performance in different engineering tasks\",\"authors\":\"C. MacRobert\",\"doi\":\"10.1680/jfoen.23.00001\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"To explore how different tasks inform decisions, geotechnicians were invited to predict the stability of a failed tailings dam. Participants included students, contractors and experienced consultants. The first task involved noting down stability concerns and making a stability prediction. Participants then prepared sketches, suggested material parameters and updated their predictions. The third task involved suggesting strength parameters for a limit equilibrium model, critiquing the model and making a final stability prediction. A 30 min time limit increased the chance of human error. Accurately assessing the failure mode at the note and sketch sections did not translate into correct predictions, despite high confidence in these tasks, particularly from experienced participants. The majority (80%) of final stability predictions were highly correlated with the analytical computer model, despite low confidence in this model, particularly from experienced participants. While this trend in confidence is perhaps expected, most participants, including experienced participants, failed to identify that failure was imminent and were confident in their final predictions. This was largely because many suggested drained strength parameters, rather than undrained strength parameters appropriate for the problem. This highlights the importance of intuitive tasks such as note taking and sketching to understand problems before building analytical models.\",\"PeriodicalId\":42902,\"journal\":{\"name\":\"Proceedings of the Institution of Civil Engineers-Forensic Engineering\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.4000,\"publicationDate\":\"2023-06-30\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the Institution of Civil Engineers-Forensic Engineering\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1680/jfoen.23.00001\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q4\",\"JCRName\":\"ENGINEERING, CIVIL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the Institution of Civil Engineers-Forensic Engineering","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1680/jfoen.23.00001","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"ENGINEERING, CIVIL","Score":null,"Total":0}
Confidence and performance in different engineering tasks
To explore how different tasks inform decisions, geotechnicians were invited to predict the stability of a failed tailings dam. Participants included students, contractors and experienced consultants. The first task involved noting down stability concerns and making a stability prediction. Participants then prepared sketches, suggested material parameters and updated their predictions. The third task involved suggesting strength parameters for a limit equilibrium model, critiquing the model and making a final stability prediction. A 30 min time limit increased the chance of human error. Accurately assessing the failure mode at the note and sketch sections did not translate into correct predictions, despite high confidence in these tasks, particularly from experienced participants. The majority (80%) of final stability predictions were highly correlated with the analytical computer model, despite low confidence in this model, particularly from experienced participants. While this trend in confidence is perhaps expected, most participants, including experienced participants, failed to identify that failure was imminent and were confident in their final predictions. This was largely because many suggested drained strength parameters, rather than undrained strength parameters appropriate for the problem. This highlights the importance of intuitive tasks such as note taking and sketching to understand problems before building analytical models.