Senior thesis thoughts

  • Most students' initial thoughts for a topic are almost laughably broad. One girl in my research methods class announced with a straight face that she wanted to measure the impact of foreign aid on country GDP. I asked whether she'd want to consider evaluating the effects of one project in one location. That said, I guess it's an easy mistake to make if you're not used to reading economic papers, or familiar with how this kind of academic bread gets buttered.
  • The other disturbing trend is that students generally have an opinion or conclusion in mind and then want to find evidence to support it. For example, aforementioned girl wanted to show that foreign aid doesn't have much effect on GDP. This is evidence for the theory that reasoning isn't about logic, it's about arguing. Plus I guess if you're only going to produce one thesis in your life, you might as well try to have it say something about you or your worldview.
  • About half of the senior theses last year were sports-related.
  • I lucked into a topic that's ostensibly about economics, but will let me do more web development and user experience testing. I'd been gathering ideas for about a year though; here they are. Most of them are no good because it would be really hard to collect data to test the hypothesis.
********** One sentence summary: Do students at schools with higher alcohol consumption rates enjoy outsize success later on? Examining the effects of drinking in higher education: drinkers generally earn higher incomes than non-drinkers, but do higher-drinking schools enjoy more business success? Drinking may be a way to signal to others that you're trustworthy; do higher-drinking schools enjoy closer alumni networks, more donations back to the school, etc? Do high drinkers have higher status jobs? ************ Libertarians value personal liberty above all else. Among other things, Haidt et al (2010) have found that libertarians have weaker social ties, lower interdependence, etc. However, these qualities are also associated with higher levels of unhappiness and depression. For instance, Buck Schieffelin has tried to find a depression equivalent among members of a New Guinean tribe, with no success. Group ties there are very strong; a member who experiences loss is taken care of by the tribe. Other evidence includes the fact that American women have become unhappier as they’ve become wealthier and more independent. I investigate whether a more libertarian, economically efficient society would be more unhappy, at the same level of income, or whether a more libertarian minded population would be more unhappy, or both. ************* Various ideas related to depression: The amount of Americans indicating they've had a depressive episode has steadily increased over time, and the age of first depression has steadily decreased. Why has this happened? Depression is a reaction to low social status. I want to examine the possibility of treating depression through raising the patient's status - through an MMORPG, hired actors, confederates, a new social group, or some other means. Massively multiplayer online role-playing games are a place where status is entirely earned through hard work - how many points accumulated, kills, etc. They're also engaging, somewhat addictive and tend to distract people from real life. I'm wondering if we could treat depression by having patients slowly accumulate status and recognition in an MMORPG. Would people accept whole-world simulations, e.g. life in a Matrix-like world, instead of suicide, if the option existed? ************* The placebo effect works because we believe that the pill or therapy we are taking will have an effect, so our brains adjust in response even though there is no healing power in the medicine or therapy itself. What if we believed really strongly (eg had a wise person, a religion, or society, tell us) that our brain could produce various chemicals to increase concentration, reduce pain, increase pleasure, etc? It could become a self-fulfilling prophecy, maybe, if people started to expect that their brains would produce more of the chemical in reaction to believing it. Idea from Iain Banks, Player of Games. *********** One sentence summary: Can we convince irrational high discounters to save more money? Testable Question: Develop a model of saving and evaluate legality of offering. Dataset: Purchasing data on lotteries and their odds - do lottery consumers buy based on the odds. Most people should save more, but current forms of saving (CDs, stocks, bank accounts) are either inaccessible to the poor or don’t provide any short term reward. Furthermore behavior such as gambling doesn’t fit into most academic models of saving. What if we encouraged people to save by saving some of the money they use to gamble with? How it works: You purchase a lottery ticket. X per cent of the money goes into the prize pool with everyone else who bought a ticket and is available for immediate distribution. (100-X) per cent is collected and invested on behalf of the ticket purchaser, say by purchasing a 1-year CD. The (100-X) portion can be collected by the purchaser after one year or two years. The effect is to turn lotteries, a negative-NPV investment, into a savings product. *********** One sentence summary: If prediction markets provided accurate estimates of budgets and time to completion for government projects, would contract bidding and completion estimates become more honest? Testable question: What drives decision making and approval of government projects? How accountable are the officials involved to public scrutiny? Dataset: Series of past government projects, a model predicting how government officials will act in each situation I investigate the possibility of introducing a corruption market - each government official is tradeable and the closer his/her price is to 1, the more likely the official is engaging in corrupt behavior. You can buy and sell corruption tickets - perhaps if the price is above 0.9 for a week the official gets investigated by a board, or some other mechanism. Similarly, investigate creating prediction markets designed to make government more accountable: prediction markets for the final cost of government outlays (transportation projects with ballooning budgets, prediction markets for GDP/employment data revisions, etc.) ******** One sentence summary: Can we take advantage of investors fear of underperforming relative to other investors? Testable Question: Can an investor worried about overall return improve upon current market performance? Does the theory suggest arbitrage opportunities? Dataset: CRSP As Eric Falkenstein points out in his paper "Risk and Return in General: Theory and Evidence," there is virtually zero demonstratable premium for taking on excess risk. Falkenstein's theory is that investors are more concerned for their status vs. other investors than they are concerned about overall return. Thus there should be a difference in people's expectations for the return on various risky assets even though there is little difference in the real return. Could an investor unworried by social status, or the yearly returns of his fund, exploit this expectation gap via arbitrage, by setting up a prediction market for returns, or shorting options on "risky" assets while buying options on "safe" ones. ************ One sentence summary: Test the theory that weak employees stay with firms for the longest because their outside prospects are slim, and the resulting implications for companies that offer pensions, healthcare, etc. Testable Question: How does employee turnover affect a firm's profitability? Dataset: List of employees by time with company, and their productivity. Employees who are with a firm for a long time might be the firm's weakest employees. Weak employees have little motivation to seek other jobs, and no one is trying to hire them. There are a number of interesting hypotheses here - Investigate whether this is harming the long-term stay with one company model, or whether high performers feel compelled to signal strength by switching jobs often. ********** One sentence summary: Improve traffic, pollution, quality of life, potentially wages, by placing a tax on commute time. Testable Question: Can we place a tax on commute time? What would be the effects? Dataset: A model that predicts how people choose their homes (price, location, crime, etc) that will be able to estimate the change in total miles driven from an increase in the price of a longer commute. Also a model of how much a shorter commute is worth to employees. Long commutes are correlated with lower levels of happiness among employees. People seek longer commutes because home prices near the place of employment are generally higher and they can purchase more home for the same price by living further away, even though this has deleterious effects on welfare. I investigate the effects of a tax on high commute time and whether such a policy would be welfare-improving. I predict the following effects: workers will move closer to the workplace, workers will be more satisfied with their lives, workplaces will spread out further, municipalities will face more demand to invest in transportation/reducing the cost of traffic jams, wages could rise, for instance, to adequately compensate someone with a longer commute time. *************************************************************** throwaways *************************************************************** Is there a correlation between gross Internet use and student grades? Are students who are on the Internet too much getting worse grades? We could track students using SurplusMeter, an unobtrusive Internet-tracking data application for the Mac. Do the dreams of high and low status people differ in significant ways? How important are people's sources of information to their idea generation and ultimate success? Examine very successful, average success and low success people to see how they collect information, whether their sources are significantly different from their peer groups. Do we pick colleges to impress our high school friends and jobs to impress our college friends? Persistence is a useful tool but in the short term persistence is low status, because sometimes you might try really hard and still fail, an outcome clearly worse than not trying at all. Measure persistence, suggest ways to raise the status of persistent people, etc, try to find a correlation between avg persistence and success.

Liked what you read? I am looking for work.

Leave a Reply

Your email address will not be published. Required fields are marked *

Comments are heavily moderated.