The effectiveness of self-imposed deadlines on procrastination

I often hear of graduate students postponing their research to do other things: play Tetris, read comments on Slashdot, or write a blog. We defer doing something “more important” to do something else and feel guilty and pleased at the same time.

How sweet is it not to do work? Apparently, sweet enough to abate the heavy and bitter costs of procrastinating. Late fines and extra work for missing a deadline seem distant when you can chat online for another 20 minutes right now.

Why do people procrastinate? This is an effect psychologists attribute to “hyperbolic time discounting”: the immediate rewards are disproportionally more compelling than the greater delayed costs. In other words, Procrastination itself is the reward.

However, the eventual cost of neglecting a task has such an impact on people that they learn to impose deadlines on themselves to restrict their own behavior. At what lengths do people do this? This article looks at three questions:

  1. Do people self-impose costly deadlines on tasks in which procrastination may impede performance?
  2. Are self-imposed deadlines effective in improving task performance?
  3. Do people set their deadlines optimally, for maximum performance enhancement?

A few studies are reported in this paper, where students had the opportunity to choose their own deadlines for three tasks they needed to do (write or proofread papers). They were allowed to set separate deadlines for each paper, but they would be bound to the deadlines and be assessed penalties if the papers were submitted late. Logically, the best solution would be to set all the deadlines to be the last day, which would give them the most flexibility and time to work on the three tasks.

However, only 27% of the students chose to submit all three papers on the last day of class. This answers the first question — people are aware of their own procrastination and give themselves earlier deadlines to counter it. The studies show that these deadlines do improve performance over only having deadlines at the very end. Unfortunately, they are still suboptimal because the subjects who were given equally spaced deadlines performed better, thus supporting question two but rejecting question three.

Procrastination Study

But hey, I’ll push myself to start my taxes earlier, but after a round or two of Winterbells.

Ariely, D. & Wertenbroch, K. (2002). Procrastination, deadlines, and performance: Self-control by precommitment. Psychological Science, 13(3), 219-224. [PDF]

December 26, 2006 at 2:42 am 85 comments

Comic Chat

I remember using Comic Chat when I was in elementary school, trying it out with my new Internet Explorer install. It was the first chat program I had used, and thought it was both exciting and scary to be able to talk to complete strangers. Comic Chat is an application which generates comics from online chat, and uses the IRC protocol.

I was surprised to find this paper on Comic Chat written by the authors in 1996. Interestingly, it was published in SIGGRAPH, the top computer graphics conference in academia. From reading this paper, I find that Comic Chat is a lot more complicated than I initially thought.

Comic Chat creates realistic comics, which mainly consist of characters, speech balloons, and panels.


Generating a comic requires placing characters in a panel. Comic Chat used cues present in the text to generate the character’s gesture and expression. Things such as smileys :-), use of “I” or “you”, and punctuation would change the appearance of the character. In addition, the position and orientation of the characters is determined by a greedy algorithm. The following strip has examples of position and orientation issues: the first panel is missing a speaker, the characters in the second panel are not facing each other, and the outer two characters in the third panel are talking over the two middle characters. The fourth panel shows a correctly drawn panel.



Comics generally use four different types of balloons,

  1. Speech balloons for regular text, drawn with a solid outline and tail
  2. Thought balloons for what a character is thinking, with a solid online but a tail of ovals
  3. Whisper balloons for private conversation, with a dotted outline and tail
  4. Shout balloons for shouting text, with a jagged outline (not shown in figure)


Determining a balloon’s dimensions and placement is determined by a complex algorithm, which you can find in the paper. There are many things to take into consideration when placing balloons, such as placing them so they are read in the correct order, so they don’t overlap, so they are located somewhat over the speaker’s head, and to leave room for the tails.


Panel breaks are calculated to accommodate text properly, and to make the comic appear more natural. Breaks can be made when there are too many characters in a panel, or there is not enough room for the text. A break is also introduced when a character speaks twice to ensure a character does not have more than one balloon per panel. Panels are usually close-ups of characters to get a good view of the active character. However, a zoomed out shot is sometimes done to show the surroundings and characters in the scene.

While Comic Chat has become antiquated and few users now use it to chat online, it still has some value today. I realized when reading this paper that the web comic, Jerkcity is constructed from Comic Chat.

You can download a copy of Comic Chat if you want to give it a spin.

Kurlander, D., Skelly, T., & Salesin, D. (1996). Comic Chat. Proceedings from SIGGRAPH ’96: International Conference on Computer Graphics and Interactive Techniques, 225-236. [PDF]

December 4, 2006 at 1:09 pm 11 comments

The First Digit Law

If I were to pick a random city in the world, and tell you its population, what might the first digit of that number be?

You may think there’s equal probability for the first digit to be 1 to 9, but over 30% of the time it’s 1 (one).

Why? Think about it this way: let’s say a stock price doubles every year, starting at $100/share; it would spend a year with a first digit of 1 until it reaches $200, a year as $2xx or $3xx until it reaches $400, a year as $4xx, $5xx, $6xx, or $7xx, and then just a month or so at $8xx or $9xx, and all of a sudden it’s at $1,000 and the first digit is 1 again. Now it takes a long time (a year) to reach $2,000. There is a disproportionate amount of time when the stock price begins with the digit 1.

Many things in nature increase logarithmically. Benford observed this first-digit phenomenon in places including populations, addresses, baseball statistics, area of rivers, specific heats of compounds, and death rates. This rule has been used to identify accounting fraud where made-up numbers don’t match the distribution found in real accounting numbers.

Benford sampled over 20,000 numbers, and noticed the distribution of numbers was as follows,The First Digit Law

Digit Occurence
1 30.6%
2 18.5%
3 12.4%
4 9.4%
5 8.0%
6 6.4%
7 5.1%
8 4.9%
9 4.7%

This can be closely modeled using the log distribution of

F_a = log(1 + 1/a)

where F_a is the frequency that the digit a is the first digit in used numbers.

Additionally, the frequency of the n-th digit of a number can also be calculated using a similar formula, presented in the paper.

This is the law of anomalous numbers. We’ve learned to count 1, 2, 3, 4, … but nature counts 1, 2, 4, 8, …

Benford, F. (1938). The Law of Anomalous Numbers. Proceedings of the American Philosophical Society, 78(4), 551-572.

November 21, 2006 at 12:57 am 6 comments

Customer Satisfaction and Word of Mouth

Do consumers complain or praise more?

Telling other people about your experiences with a company of product is something many people do. This may be evangelizing their newly purchased laptop, or gripe about the slow service at a restaurant. Word of mouth is defined as informally communicating with a third party about goods and services, rather directly telling the company.

There are 3 theories about consumers engaging in word of mouth.

  1. Highly satisfied consumers engage in more word of mouth, because
    • they want to help others,
    • appear well informed or smart,
    • engage in ego defense,
    • reduce cognitive dissonance,
    • present themselves in a positive light;
    • they may have a general bias towards positive things,
    • or want to avoid saying negative things.
  2. Dissatisfied consumers engage in more word of mouth, because
    • they want to vent hostility,
    • reduce anxiety,
    • warn others,
    • or seek vengeance.
  3. Consumers engage in word of mouth more if they are extremely satisfied or dissatisfied, because
    • they will have a greater impact, hence more utility when they engage in word of mouth in a more extreme situation.

Word of mouth activity is also predicted to increase at a greater rate for dissatisfied consumers than satisfied consumers, since negative information about products is less common and has a greater effect.

A study indicates that theory 3 is more likely than the other two, since the resulting data is a U-shaped model, where highly satisfied and dissatisfied consumers engage in greater word of mouth. The hypothesis that negative word of mouth increases at a greater rate is also supported by the data.


The study replicated similar findings in both the United States and Sweden, indicating that word of mouth practices are likely universal.

Anderson, E. W. (1998). Customer Satisfaction and Word of Mouth. Journal of Service Research, 1(1), 5-17. [PDF]

November 20, 2006 at 3:58 pm 5 comments

The Creative Process

What are the underlying cognitive mechanisms for the creative process?

Gabora explains what happens to our minds when we’re doing creative work.

Stages of the Creative Process

Preparation: Becomes obsessed with the problem, collects relevant data and attempts traditional approaches.

Incubation: Does not actively attempt to solve the problem, but unconsciously continues to work on it.

Taking a shower is my most common incubation step. Programming bugs seem to reveal themselves when I’m lathering up in there.

Illumination: Possibilities surface to the consciousness in a vague and unpolished form.

Verification: The idea is worked into a form that can be proven and communicated to others.

Architecture of the Mind

Since the human mind can sense and store any of the combinations of sound, color, etc., the number of memory locations is much less than the number of possible experiences. The mind can therefore be represented as a sparse matrix.Mind Architecture

During the initial, intuitive phase, each thought activates, and potentially retrieves information from, a large region containing many memory locations. Because of the distributed, content-addressable structure of memory, the diverse contents of these many locations merge to generate the next thought. Novel associations often result. As one focuses on an idea, the region searched and retrieved from narrows, such that the next thought is the product of fewer memory locations. This enables a shift from association-based to causation-based thinking, which facilitates the fine-tuning and manifestation of the creative work.

This association-based seems to be the main reason why brainstorming works. Jotting down everything you can think of about a topic lets you visualize the associations between concepts.

Gabora, L. (2002). Cognitive Mechanisms Underlying the Creative Process. Proceedings from C&C ’02: Creativity and Cognition, 126-133. [HTML]

October 22, 2006 at 1:53 am 5 comments

Academic Prostitution: Publishing what the referees want

The system of journal editing existing in our field at the present time virtually forces academics to become prostitutes: they sell themselves for money (and a good living). Unlike prostitutes who sell their bodies for money (Edlund and Korn, 2002), academics sell their soul to conform to the will of others, the referees and editors, in order to gain one advantage, namely publication.

The process of going from lowly undergrad to omniscient professor goes something like this:

  1. Undergrad: Write a senior thesis (for good grad school recommendations)
  2. Grad student: Publish at least 1 paper each year in the top journals (conferences, if in engineering) in your field. Pass Quals
  3. Assistant Professor: Publish several papers (including those you “co-author” with your students); get grant money. Repeat for 6 years
  4. Tenured Professor: You win, game over. Credits roll…

Obviously, getting published is the most important factor of an academic’s career. However, to get into the top journals of your field, your work has to be approved by the editors and referees. If the editor accepts your paper in the first round, several anonymous referees review it and offer suggestions for improvements. They also hold veto power, and your paper can be rejected by any referee. Only once you have made all the “suggested” changes (and this may go for several revisions), will you have a chance of being accepted for publication.

Making the revisions forces you to publish something different from your original work under the demands of an anonymous person. It also costs precious time, and time is always ticking on your academic career.

This presents a dissonance in the system because referees have the power to dictate changes to a paper, but no property rights in the journals. They may appear to act in the journal’s interest, but there is no economic benefit to them for doing so.

To counter this, Frey proposes a modified publication system where journal editors make an accept/reject decision upon receiving the paper, and referees propose suggestions which are up to the author to implement or ignore. This system treats scholars like artists, reducing intellectual prostitution and bolstering creativity in published articles.

The academic publication process is unlikely to change anytime soon, but writing books or putting articles on the web in a working paper series (such as SSRN or arXiv are potential alternatives for some academics out there.

Frey, B. S. (2003). Publishing as prostitution? – Choosing between one’s own ideas and academic success. Public Choice, 116, 205-223.

October 15, 2006 at 5:23 pm 17 comments

The Effect of Brief Naps on Alertness and Cognitive Performance

If you don’t get enough sleep at night, a quick 10 minute nap may help improve your alertness and cognitive performance. However, shortening it to a 30 or 90 second nap won’t cut it.

Sleeping Girl

Psychology researchers performed a study with 16 subjects, each restricted to 5 hours of sleep at night. The subjects were split into 4 groups — no nap, 30-s nap, 90-s nap, and 10-min nap. Subjects that took naps for 90 seconds or less were not found to perform any better on alertness and cognitive tasks. However, subjects that took a 10 minute nap significantly improve performance in multiple post-nap tests. This seems to suggest that only stage 2 sleep helps you recuperate from lack of nocturnal sleep.

Tietzel, A. J. & Lack, L. C. (2002). The recuperative value of brief and ultra-brief naps on alertness and cognitive performance. Journal of Sleep Research, 11, 213-218. [PDF]

October 13, 2006 at 4:26 pm 16 comments

Older Posts Newer Posts


Read any good papers lately?

If you're interested in academic research, I'd love to have additional contributors. Shoot me an email.