NOTE: Every English paragraph is followed by the corresponding Chinese translation
注：每段英文后为该段的中文翻译
Summary | 摘要
In this blog post, I analyse my time spent on the PhD work in the year 2020, identify what goes wrong and propose some solutions to solve the identified issues in the future.
In the year 2020, I worked 1298 hours in total, much lower than the 2080 hours expected for a normal nine-to-five job. From this perspective, I fail my PhD this year.
I fall short of the 2080-hour benchmark mainly because I either worked less than 6.5 hours or did not work at all in over 50% of the working days.
The working duration of previous days is contagious. If I shirked or did not work yesterday, I will be very likely to do so again today.
The time I start my work affects how long I will be working on that day. If I did not start the work on time (i.e. later than 9am), I will be less likely to reach the expected 8-hour working duration for that day.
The biweekly meeting with my PhD supervisors greatly influences my working duration. If I finish a meeting with them recently, I tend to work less than 6.5 hours on the day when the meeting is held and the day after it.
To achieve the 2080-hour benchmark next year, I should (1) start the work on time, (2) set up some goals that are due soon after the supervisory meeting to keep me engaged during the between-meeting period, (3) create a dashboard to monitor my working duration on a daily basis.
To better understand my work productivity, I should keep a daily record of the tasks I complete next year as well.
我在这篇文章里分析了我2020年花在博士工作上的时间，分析了我哪里没做好并提出解决方案。
2020年我总共工作了1298小时，远低于一个朝九晚五员工应有的2080小时。从这个角度来说，今年我的博士工作没达标。
我没能达到2080小时的标准主要是因为有超过50%的工作日，我工作少于6.5个小时或者完全没在工作。
前段时间的工作时数会影响我现在的工作时数。如果我昨天偷懒或者完全没做任何工作，那么我今天有很大概率会再次偷懒或者完全不工作。
我从几点开始工作也影响了我当天会工作多久。如果我开始工作的时间晚了（晚于9点），我当天就不太可能会工作到8个小时。
我和我导师每两周一次的会议也会影响到我的工作时长。如果我最近刚和他们开过会，那么会议当天和之后的一天，我一般会工作少于6.5小时。
如果我想要在明年达到2080的工时目标，我需要：（1）按时开始工作；（2）设定一些很快会在导师会议结束后到期的目标，以此让我在下一次导师会议前保持动力；（3）建立一套系统来监测我每日的工作时长。
若要更好的了解我的工作效率，我还需要在明年收集我每日完成的任务。
1. Introduction | 引言
It is near the end of another year. When it comes to this time of year, we usually reflect back upon the entire year. Oftentimes, our first reaction is: "OMG, time flies so quickly, but I haven't done much!" Another question we commonly hear is: "Where has my time gone?"
又到了一年的年末，每年到了这个时候，大家都会回顾这一年。通常，大多数人的第一反应都是：“天呀，时间过得真快，但我好像都没做什么事情！”我们也常听到另外一个疑问：“我的时间都去哪了？”
As a PhD student, I am also interested in how I spent my time. Since my PhD is funded by the university, the relationship between me and the university does not, essentially, differ from an employer-employee relationship. I am expected to work on a daily basis as well. However, unlike a normal nine-to-five job, I have total freedom in managing my work; no one is monitoring whether I go to the office on a given day or how long I work per day. Under this special setting, I am very curious about how disciplined I am when it comes to working duration.
作为在读的博士生，我也很好奇我是如何安排我的时间的。由于我的博士学习是学校赞助的，我和学校的关系在本质上和一般雇主与员工的关系差别不大。每日工作也应该是我的日常，但和大多数朝九晚五的工作不同，我可以全权安排自己的工作。没人会来检查监督我是否去了办公室或我每天工作了多久。在这种特殊的设定下，我也很好奇我会自觉地工作多长时间。
I happen to keep a record of my time use in my Outlook calendar. With this data, I can analyse my time use and answer the following three questions in this post:
碰巧的是，我一直用微软的Outlook来记录我的时间分配。通过分析这个数据，我将在这篇文章中回答以下三个问题：
[Question 1] How many hours did I work in the year 2020? Does my annual working duration hit the required amount for the full-time employee?
[Question 2] If I fail to reach the required the working duration of full-time employees, what is causing this failure?
[Question 3] What actions can I take to avoid this failure in the future?
[问题1] 在2020年里，我总共工作了多少个小时？我今年的工作时长是否达到了全职员工的水平？
[问题2] 如果我的工作时长没达到全职员工的水平，是什么原因导致的？
[问题3] 若要减少今后工作时长低于全职员工的情况，我可以采取什么措施？
2. Compare My Working Duration & Benchmark | 比较我的工时数和标杆值
Answering the first question simply needs to compare two numbers. The first number is the benchmark - the total number of hours a full-time employee should work in a year, and it can be calculated as follows: 8 hour/day * 5 day/week * 52 week/year = 2080 hours.
回答第一个问题仅需要对比两个数字。第一个数字是标杆值 —— 一个全职员工一年里应工作的总小时数。它是这么算的：每天工作8小时 * 每周工作5天 * 一年里共有52周 = 2080个小时。
The second number is the total number of hours I worked this year, which is approximately 1298 hours. Comparing this number with the benchmark (2080), there is no doubt that I have failed to reach the target by 782 hours (equivalent to 13 days). As most people, after seeing this big gap, I also wonder what is the reason behind it.
第二个数字是我个人今年工作的总小时数，大约是1298个小时。对比标杆值（2080），我的工时数显然没有达到目标值。我比目标值少工作了782个小时（相当于13天）。和大多数人一样，当看到这个巨大的差距后，我也想要知道背后的原因。
3. Explain My Failure | 解释我为什么没达标
To figure out the reason, I first visualise my working hours throughout the entire year in the figure below. The x-axis is the date, and y-axis represents the number of hours. Each point in this plot corresponds to the number of hours I worked on the select day. For example, on 1st January I worked for over 11 hours. The red dashed line represents the 8-hour benchmark. To make sense of this figure, I need to categorise the days according to the working duration.
为了找到这背后的原因，我首先把我一整年的工作的小时数呈现在下方的图表里。图表的x轴代表日期，y轴代表工作小时数。图表中每一个点则代表了某天我工作了多少小时。例如，2020年1月1日，我工作了超过11个小时。图中还有条红色的虚线，它代表了8小时的标杆值。想要更好地解读这个图表，我还需要根据工作时长把每一天分类。
Points that are above the red dashed line mean that I worked overtime on those days. I use the yellow shading to enclose them. On the other extreme, for points that have a y-axis value of zero, it means I did not work at all on those days (purple shading). Meanwhile, even though we are required to work 8 hours per day, we don't actually work that long if we consider the lunch time, Internet surfing etc. Thus, for days when I worked between 6.5 and 8 hours (both ends are included), I called them a normal day (green shading). Finally, the rest of days (i.e. 0 < working hour < 6.5) are classified as shirking (blue shading).
在红色虚线以上的点表示我当天加班，我用黄色的阴影区域来表示它们。那些y值为0的点表示我当天没做任何工作，它们属于另一个极端（紫色区域表示）。虽然我们的工作制为每天8小时，但如果我们把午饭时间，网上闲逛等时间算进去，我们通常不会工作8小时那么久。因此，我把那些我工作6.5到8小时的日子（包括这两个数字）归类为正常工作日（绿色区域）。所有剩下的还没归类的日子（工作时长大于0但小于6.5小时的日子），我将它们归类为偷懒的日子(见蓝色区域)。
A quick glimpse at the new plot would reveal that most of points fall in the areas shaded by blue and purple, meaning that I either shirked or did not work at all for the majority of weekdays in 2020. This impression is indeed confirmed by the raw count shown in the table below.
快速一瞥这个图表，我们会发现大多数的点似乎都落在了蓝色或紫色的区域。这意味着在2020年的大部分工作日里我不是偷懒了就是压根没做任何工作。下方表格里的数据也证实了这个感觉。
As can be seen from the above table, I shirked or did not work in more than 50% of the weekdays in 2020. They are undesirable as the more they occur the fewer hours I could work, taking me further away from the benchmark of 2080 working hours. For brevity, hereinafter, I decide to the days when I shirked and did not work as unproductive days.
如上方表格所示，2020年里的所有工作日里，有超过50%的日子我偷懒了或者完全没在工作。这两个情况是我不希望发生的，因为它们的发生意味着我的工时会下降，最终结果就是我离2080的工时标杆越来越远。为了行文流畅，接下来我会把我偷懒或者没工作的日子称为效率低下的日子。
Some actions must, therefore, be taken to minimise the occurrence of unproductive days in the future. However, I first need to figure out the reason of me shirking and not working before coming up with any meaningful solution. I have three hypotheses regarding what is causing both unproductive days.
若要减少这些效率低下的日子，我必须采取相关的措施。但找到任何有意义的措施前，我必须知道是什么原因导致了我在某些日子效率低下。对此，我提出了三个假设。
Hypothesis 1 - past effect | 假设 1 —— 过去的影响
I assume that the productivity today is positively associated with that of the day before. For example, if I shirked yesterday, I tend to shirk again today. Because this hypothesis involves events in the past, it is the past effect.
我的第一个猜想是今天的效率和昨天的效率是正相关的。例如，如果我昨天偷懒了，那么我今天就会继续偷懒。因为这个假设涉及到了过去的事情，所以也可以理解为过去的影响。
Hypothesis 2 - current effect | 假设 2 —— 现在的影响
Given a normal working hour is from 9am to 5pm, if I start the work late (i.e. later than 9am), I will never be able to reach the eight-hour target for that day. Hence, I assume that starting the work late will make that day unproductive. Since it is at what time I start the work that matters the most, the second hypothesis focuses on the current effect.
一个正常的工作日从早上9点到下午5点，那么如果我开始工作的时间晚了（晚于9点），那当天我绝不可能工作到8个小时。因此我猜想开始当天工作的时间晚了的话，这天就会成为效率低下的一天。在这里，因为是我当天什么时候开始工作起着最大影响作用，所以第二个假设主要探究的是现在的影响。
Hypothesis 3 - future effect | 假设 3 —— 将来的影响
As you can guess, the final hypothesis is about the future effect. I usually have a meeting with my PhD supervisors every two weeks. Just like any meeting, I have to get something done and give it to my supervisors so we can have things to discuss during the meeting. In this case, the date when I have the supervisory meeting is like a deadline to me. Thus, I assume that if I meet my supervisors recently (defined as today or yesterday), then I will relax on both the meeting day and the day after it and work fewer hours (or do not work at all).
正如你能猜到的，最后一个假设是关于将来的影响。多数情况下，我和我博士导师每两周开一次会议。和任何会议一样，我需要在会议之前完成一些任务并把成果发给导师以便我们会议的讨论。因此，每次有会议的日子对我来说相当于一次项目的截止日期。因此我猜想，如果我最近刚和导师开完会（“最近”指会议当天或前一天），那么我会选择在会议的当天和会议后的一天休整，减少工作时长（或者干脆不工作）。
Which of these three hypotheses is true? To answer this question, the time use data from my Outlook calendar is analysed.
这三个假设里，哪一个是对的呢？要回答这个问题，就需要分析一下我Outlook日历的数据。
Test Hypothesis | 验证假设
The following bar chart tests whether the productivity yesterday affects my performance today. There are two interesting patterns in this chart. Firstly, productivity yesterday indeed positively associates with the performance today. For instance, if I did not work yesterday, I also tend not to work today (with a probability of 34%). Similarly, if I shirked yesterday, there is a 40% chance that I would shirk again today. On the other hand, if I worked overtime yesterday, I would be productive today (30% + 23%). The second interesting finding from this bar chart is related with the probability of shirking. Regardless of how I was doing yesterday, there is always an unignorable chance that I would shirk in the next day.
下方的柱状图验证了我昨天的效率是否会影响今日的效率。我们可以从这个柱状图里发现两个有趣的规律。第一，昨天的效率和今日的效率具有正相关性。例如，我昨天没有工作的话，那么我今天也不太可能工作（这个事件发生机率是34%）。如果我昨天偷懒了，那么我明天就有40%的机率会再次偷懒。另外，如果我昨天加班了，那么今天有较大的概率成为高效率的一天（30% + 23%）。第二个有趣的发现和我偷懒的机率有关。从图中可以看到，不管我昨天的效率如何，第二天我偷懒的机率都存在且占有不可忽视的比例。
This bar chart supports the past effect; that is, unproductive days are likely to lead to another unproductive day. An intuition would be to break this vicious cycle. Nonetheless, from this chart, it is still unclear why some days were unproductive at the very beginning. Without the knowledge of the root cause, there is no way to propose any solution to break the cycle. The finding from the second hypothesis, however, provides a possible direction.
前面的柱状图证明了过去的影响是存在的。也就是说，效率低下的日子有较大机率导致紧接的日子也变得效率低下。一个正常的想法自然是设法阻断这个消极的反馈链条。然而，我们无法仅靠上面的图表了解最初是什么因素导致某些日子效率低下。若不知道这个根源，就没法找到打破这个反馈链条的对策。幸运的是，第二个假设提供了一些思考方向。
Another bar chart is displayed below to test the second hypothesis about the relationship between starting the work late and working duration of the day. Again, shirking is an unignorable issue. Apart from it, what is more interesting is the supporting evidence for the current effect. Compared to starting the day on time (or earlier than 9am), starting to work late would see my chance of shirking increase from 38% to 54%. Once combining this finding with the one from the past effect hypothesis, there is one action I could take to break the negative feedback loops between unproductive days. That is, start the work on time or earlier. As such, I not only have a higher chance of being productive today but also reduce the likelihood of being unproductive tomorrow.
下方是另外一个柱状图，它验证了第二个关于开始工作的时间点和当天工作时长的假设。这个图仍然反应出偷懒是一个不可忽视的问题。除此之外，另一个更有趣的发现是它证明了现在的影响的存在。相比于按时开始工作（或早于9点开始工作），开始工作晚了会让我偷懒的机率从38%上升到54%。把这个发现和前面关于过去的影响的发现结合起来考虑，我就可以找到一条对策来打破前面提到的消极反馈链条。那就是：按时或提早开始每日的工作。这样做不仅提高了今天变得有效率的机率，同时还能降低我第二天低效率的可能性。
Findings from testing the third hypothesis are also useful. The bar chart below seems to indicate that I am a deadline-driven person. For example, when the deadline passes (i.e. supervisory meeting is finished recently), I have a high chance (63%) of shirking. People who are deadline-driven tend to be most productive when the deadline is approaching and then spend most time relaxing after meeting the deadline. However, looking at left bar in the plot (i.e. the period when I do not meet my supervisors recently and I should be preparing for the next meeting), I am still be very likely to be unproductive (24% chance to be not working + 28% chance to be shirking). Given that all my supervisory meetings are about two weeks apart, it implies that I might have already completed the tasks for the next meeting within a week and then choose to relax in the remaining period prior to the next meeting. In other words, the goal I set for the between-meeing interval may be too easy to achieve.
第三个假设的检测结果同样很有用。如下方的柱状图所示，我似乎是一个以截止日期为导向的人。例如，一旦过了截止日期（我的情况里指刚和导师的开完会），我就很有可能偷懒（63%的概率）。通常，以截止日期为导向的人在临近截止日期时会最有效率。而当截止日期一过，他们会把大多数时间拿来休息。然而，如果我们把目光聚焦在图中左边的柱子（这个柱子代表了我最近没和导师开过会，但正在准备下一次会议的时间段），这个时间段的我仍然没有效率（24%的机率不工作 + 28%的机率偷懒）。考虑到我和导师的会议差不多两周一次，这意味着我可能已经在其中的一周内就完成了下次会议所要求的任务，然后就拿剩余的时间休整去了。换句话说，我给两次会议间隔的时间段设置的目标太容易达成了。
5. Recommendations | 推荐对策
Based on all the findings, there are three possible actions that I could take in the upcoming year to improve my time use and hit the 2080-hour benchmark:
基于以上所有分析结果，在即将到来的一年里，我需要采取三个对策来改善我对时间的支配以及达到2080的工时标杆：
- Action 1: start the work on time.
- Action 2: set up some goals that are due soon after the supervisory meeting to keep me engaged during the between-meeting period.
- Action 3: create a dashboard to monitor my working duration on a daily basis.
- 对策1: 按时开始工作。
- 对策2: 设定一些很快会在导师会议结束后到期的目标，以此让我在下一次导师会议前保持动力。
- 对策3: 建立一套系统来监测我每日的工作时长。
6. Limitation & Future Work | 本文的不足之处 & 将来的计划
This post simply uses the total number of hours I work this year to evaluate whether my PhD of the entire year fails or succeeds. However, we all know that working longer hours does not necessarily imply the success of the year. We can work productively and get things done within a short amount of time. Alternatively, we can multi-task between doing the work and checking our own social media, resulting in a lengthy working duration.
这篇文章只用了总工作时长这一个指标来评价我一年博士工作成功与否。然而，我们都知道工作时间越长并不等同于这一年就成功了。我们有可能很有效率可以在很短的时间内完成既定的任务。我们也有可能一边工作一边刷社交媒体，这就拉长了我们的总工作时。
A more suitable measure for evaluating our work performance might be the average number of tasks completed per day. By taking into account both the output the input (i.e. working time) and (i.e. completed tasks), the calculation of productivity is more meaningful. Therefore, in addition to take the three actions above, I will be keeping a daily record of the completed tasks next year.
日均完成任务的数量也许是一个能更恰当地衡量我全年工作表现的指标。这个方式不仅考虑了投入（即工作时间）还考虑了产出（即完成的任务数量），因此更有参考意义。所以我决定，在即将到来的新的一年，除了执行前面提及的三个对策，我还要收集每日完成的任务量。
