If I’ve learnt anything after 18 years in research, it’s that most scientists have no idea how many hours they actually work, let alone how long they need to complete specific tasks.
Without these data, you can’t make realistic plans to get things done. The inevitable consequence is work creeping into every part of your life, including the weekends. Life becomes like an episode of the US psychological television drama Severance, where you never, ever clock out.Time management for scientists
Once, a manager called me into a meeting with human resources and asked me why I wasn’t spending more time in the classroom. Because I use software to track my time, I could account for every minute of the year. The data revealed I spent much more than half my time in the classroom. We ended up having a very different conversation from the one my manager was expecting — how to reduce my teaching load. Winning!
Gathering data about how and when you work not only helps to settle occasional workplace disputes, but it can also stave off overwork. After five years of time-tracking, for instance, I know that managing my e-mail requires at least 40 minutes per day, and that I need to commit an extra 30% of ‘invisible work’ time to manage any given project. (Invisible work includes meetings, paperwork and other tasks that are unpredictable but nevertheless part of my total time budget.) Yet, when I show other people my time-tracking system, their usual reaction is polite horror. Some people cannot believe that I spend only 15 minutes or so checking data in the system for accuracy and generating reports, or fail to see the point of spending time to measure time. But I would argue that these 15 minutes are well spent if I can avoid working weekends to catch up.
Time is on my mind
I have learnt the hard way that overwork leads to burnout. But I’ve also studied the working practices of academics. In a 2019 analysis, my colleagues and I reported that academics submitted manuscripts to two medical journals at disproportionately higher rates at the weekend1. That’s not surprising — writing is one of the tasks that is most easily pushed out of normal working hours. When multiple people are involved in a writing project, the problem of slipping deadlines intensifies. Other people’s poor planning affects your own schedule, and instead of relaxing, your summer holiday is spent writing, leading to a vicious cycle of overwork.
The solution is obvious: track your time.
My favourite tool for time-tracking is the app Timing for macOS users that sits in the background and ‘watches’ how I work. (RescueTime serves a similar function for Windows users.) These programs cost up to US$10 per month on subscription (which is tax-deductible here in Australia), and allow you to capture time spent both using your computer and away from the keyboard — for instance, in meetings.
The Timing app can be trained to categorize your hours according to the type of program you are using and the keywords inside a document. A dashboard summarizes how you spend your time and the rhythms of your work, and you can decide what constitutes ‘productive’ time at the keyboard to generate a score. You might rate social-media sites, for example, as a poor use of time compared with time spent using Microsoft Word. I can add notes to these time logs; this is important because tasks such as writing can vary depending on context — writing a journal article takes more time than writing an e-mail, and it’s important to know the difference to plan appropriately.
Keeping an eye on your productivity score can help you to stay honest to your goals and arrange your days. My own analysis tells me that I am most productive before lunch, so I make sure that I schedule difficult tasks in the morning. You can also generate reports to inform your planning processes and share them with your colleagues.Burnout and breakdowns: how mid-career scientists can protect themselves
My typical working week is just over 44 hours. About half of that time is spent teaching, with the rest equally divided between service work, research, passion projects and FAT (my cheeky term for invisible work; no prizes for guessing what FAT stands for, but you can find out on my blog). Passion work is work I do for me. I don’t believe in donating extra hours to my employer, so I make sure all my university duties fit into the 37.5 hours a week stated in my contract. I also like to record podcasts and write articles like this one, which are not strictly part of my job. When a previous employer made noises about wanting to take my blog name and content when I left, my Timing data helped to show that I owned the intellectual property. I departed with my ‘Thesis Whisperer’ baby in my arms, but the experience made me realize that my career is more than the job I happen to have at any given time and that I need to protect my assets.
Pre-game analysis
When planning a big project, I like to do an analysis with the US Navy’s Program Evaluation and Review Technique (PERT).
PERT considers three variables — most likely (Tm), optimistic (To) and pessimistic (Tp) time to completion of a task — which you can estimate using your time-sheet data. I do this manually, but Timing can also generate spreadsheets showing how many hours I spent on particular projects, which enables me to compile and compare data from week to week. You can then plug your numbers into the following formulas to estimate the amount of time that you should set aside to complete the task:
PERT time estimate = (To + 4Tm + Tp) / 6
Standard deviation = (Tp − To) / 6
You will be shocked by how quickly your calendar fills up when you plan with PERT! Yet for many academics, keeping time sheets is a non-starter. Perhaps they feel that it undermines the autonomy and flexibility we all value about academia. But really, it’s the opposite: if you know how long typical tasks really take, then you can slot them into your week instead of using the weekend as a work buffer.
And it’s better that we do this rather than our employers. To be clear, I am fiercely against the idea of universities tracking our time for us — in management’s hands, detailed time-on-task data can be used as a stick to drive performance. But for researchers, time-tracking software can lead to greater productivity and clarity around your work practices and habits.
As the saying goes, what gets measured gets managed — and maybe you get your weekends back.
doi: https://doi.org/10.1038/d41586-023-00866-9
采用有道翻译如下：
如果说我在18年的研究生涯中学到了什么，那就是大多数科学家都不知道自己实际工作了多少小时，更不用说完成具体任务需要多长时间了。
没有这些数据，你就无法制定切实可行的计划来完成任务。不可避免的后果是工作渗透到你生活的方方面面，包括周末。生活就像美国心理电视剧《遣散费》的一集，你永远不会打卡下班。
科学家的时间管理
有一次，一位经理叫我去参加一个人力资源会议，问我为什么不多花点时间在教室里。因为我使用软件来记录我的时间，所以我可以记录一年中的每一分钟。数据显示，我一半以上的时间都在教室里度过。最后，我们进行了一场与经理期待的完全不同的对话——如何减轻我的教学负担。赢了!
收集关于你工作方式和时间的数据，不仅有助于解决偶尔的职场纠纷，还可以避免过度工作。例如，经过5年的时间跟踪，我知道管理电子邮件每天至少需要40分钟，而且我还需要额外拿出30%的“隐形工作”时间来管理任何既定项目。(隐形工作包括会议、文书工作和其他不可预测的任务，但它们都是我总时间预算的一部分。)然而，当我向别人展示我的时间追踪系统时，他们通常的反应是礼貌的恐惧。有些人不相信我只花15分钟左右的时间在系统中检查数据的准确性和生成报告，或者没有看到花时间测量时间的意义。但我认为，如果我能避免周末加班来赶上进度，这15分钟是值得的。
我很关心时间
我已经尝到了工作过度会导致精疲力竭的苦头。但我也研究了学术界的工作实践。在2019年的一项分析中，我和同事报告说，学者们在周末以不成比例的高比例向两家医学期刊投稿。这并不奇怪——写作是最容易被挤出正常工作时间的任务之一。当多人参与一个写作项目时，拖延截止日期的问题就会加剧。别人糟糕的计划影响了你自己的日程安排，你的暑假不但没有放松，反而被用来写作，导致了过度工作的恶性循环。
解决办法很明显:跟踪你的时间。
我最喜欢的时间跟踪工具是macOS用户的应用程序Timing，它位于后台，“看着”我如何工作。(RescueTime为Windows用户提供了类似的功能。)这些程序的订阅费高达每月10美元(在澳大利亚是免税的)，可以让你记录使用电脑和远离键盘的时间——例如，在会议中。
计时应用程序可以根据你使用的程序类型和文档中的关键字对你的时间进行分类。一个仪表板总结了你的时间安排和工作节奏，你可以决定在键盘上的“高效”时间是多少，从而得出一个分数。例如，你可能会认为社交媒体网站比使用微软文字处理软件(Microsoft Word)更浪费时间。我可以在这些时间日志中添加注释;这一点很重要，因为像写作这样的任务会因上下文而异——写一篇期刊文章比写电子邮件花费更多的时间，了解两者的区别以进行适当的计划很重要。
关注你的效率分数可以帮助你忠实于你的目标，并安排你的日子。我自己的分析告诉我，午餐前我的效率最高，所以我一定会把困难的任务安排在早上。您还可以生成报告来告知您的计划流程，并与同事共享这些报告。
倦怠和崩溃:处于职业生涯中期的科学家如何保护自己
我典型的每周工作时间只有44小时多一点。大约一半的时间花在教学上，其余时间平均分配在服务工作、研究、激情项目和FAT(我对无形工作的厚脸说法;FAT代表什么没有奖品，但你可以在我的博客上找到答案)。激情工作是我为自己做的工作。我不赞成给老板额外的工作时间，所以我确保我所有的大学职责都在我合同中规定的每周37.5小时之内。我也喜欢录制播客，写像这篇这样的文章，严格来说，这不是我工作的一部分。当我离开时，我的前雇主发出声音说想要拿走我的博客名称和内容，我的计时数据帮助表明我拥有知识产权。我抱着我的“论文低语者”宝宝离开了，但这段经历让我意识到，我的职业不仅仅是我所从事的工作
赛前分析
当计划一个大项目时，我喜欢用美国海军的计划评估和审查技术(PERT)进行分析。
PERT考虑三个变量——完成任务的最可能时间(Tm)、乐观时间(To)和悲观时间(Tp)——您可以使用时间表数据来估计这些变量。这是我手动完成的，但Timing也可以生成电子表格，显示我在特定项目上花费了多少小时，这使我能够每周编译和比较数据。然后你可以把你的数字代入下面的公式，来估计你应该留出完成任务的时间:
PERT时间估计= (To + 4Tm + Tp) / 6
标准差= (Tp−To) / 6
当您使用PERT计划时，您会惊讶地发现您的日历很快就被填满了!然而，对许多学者来说，保留时间表是不可能的。也许他们觉得这破坏了我们都看重的学术界的自主性和灵活性。但实际上，恰恰相反:如果你知道典型任务真正需要多长时间，那么你就可以把它们安排到你的一周中，而不是把周末作为工作缓冲。
这样做的最好是我们自己，而不是我们的雇主。需要明确的是，我强烈反对大学为我们跟踪时间的想法——在管理人员手中，详细的任务时间数据可以用作推动绩效的标竿。但对于研究人员来说，时间跟踪软件可以提高工作效率，让你的工作实践和习惯更加清晰。
俗话说得好，能衡量的就能管理——也许你就能重新享受周末了。