About jeanyang

Ph.D. student in computer science at MIT.

Explainer: Font Size and Aviation Regulation

What looks like a disagreement over font size has roots that run deep in the discourse of civil liberties, industry regulation, and the importance of commercial aviation to the American economy.

On April 1, the Supreme Court declined Spirit Airlines’s appeal, ending years of legal battle between the discount airline and America’s Department of Transportation about how prominently airlines must disclose the total price of a ticket [3]. Since 1984, the Department of Transportation has required airlines to advertise the “entire price” of tickets, including the taxes. Airlines previously could have advertised ticket prices and taxes separately, but according to the new rules, the tax component cannot be listed “in the same or larger size as the total price.” The US Court of Appeals says that this prevents airlines from confusing customers about total cost [1].

This represents a victory for those who want to protect the traveler from deceptive advertising. Not unlike other discount airlines, Spirit has built its business around advertising $9 fares and then charging additional fees for checked and carry-on bags, advance seat assignments, and a “passenger usage fee” of up to $17 each way for tickets booked online. Spirit is not required to include the online booking fee in advertised prices. For reasons like this, the government has been imposing a growing number of fines. In 2011, the Transportation Department assessed 21 penalties for fare advertising regulations with total fines more than $1 million, compared to $379,000 in 2001. Those fined include Spirit, LAN Airlines, South African Airways, Orbitz, Virgin Atlantic, Thai Airways, JetBlue, and Air Canada [6].

Precedent against the Spirit Airlines was heavy. As UCLA law professor Eugene Volokh explains, “The Supreme Court has said in the context of commercial advertising [that] the government has a very broad right to mandate speech that is reasonably aimed at preventing people from being misled” [6]. And thus, Spirit was not even able to convince four justices they should among the 75 cases annually (out of about 8,000 petitions received) be allowed to make their arguments to the Supreme Court [1].

But why, then, did Spirit think this fight was worth the legal fees? Several airline trade groups and the American Society of Travel Agents raised concerns about the new regulation, saying that they violate the First Amendment [6]. Airlines also stand to gain ground in the regulation battle. Spirit’s attorneys write, “Such a government effort to micromanage how speakers communicate the burdens of taxation would raise serious First Amendment concerns in any industry, but they are doubly problematic in an industry Congress specifically chose to deregulate” [1].

The back story is an ongoing discussion between airline executives and the government about increasing regulation. The question of restructuring aviation taxes and fees has been the subject of ongoing debate, especially as declining fares are causing tax revenues to drop [7]. In 2011, President Obama suggested implementing a $100 “take off” tax and also increasing the Transportation Security Administration (TSA) per-passenger security fee on top of the typical 20% tax percentage on a ticket [2]. Nicholas Calio, president of Airlines for America, responded, “We’re one of the most heavily taxed industries in the country. There are other places to go” [4].

Indeed, the airline industry is not only heavily taxed but also struggling. According to the National Airline Policy says that in an industry that has already lost $55 billion and 150,000 due to recession and rising fuel prices between 2000 and 2010. And they make some good points, stating that the commercial aviation industry creates more than $1 trillion per year in economic activity, helps drive nearly 10 million American jobs and 5 cents of every dollar of US GDP. They post a policy wish list that includes reduced taxes, reformed regulations, investment in modernizing air traffic, protection against competition from foreign carriers, and more stable energy prices [5]. The latest rules obviously do not go towards achieving any of these goals.

Spirit’s rejected appeal may represent a victory for those who wish to protect consumers from price confusion, but it is a loss for civil libertarians and airlines, who want the government to engage with them by investing rather than taxing. Looking at all sides of the story, it becomes less clear whether this decision is good for customers in the long run.

[1] Doyle, Michael. Supreme Court won’t hear Spirit Airlines’ appeal of ad price policy. The Miami Herald, April 1, 2013.

[2] Erb, Killy Phillips. Obama’s Plan for Higher Airline Taxes Sees Not So Friendly Skies Ahead. Forbes, September 28, 2011.

[3] Gulliver Business Travel Blog. Problems in the Spirit World. The Economist, April 2, 2013.

[4] Jansen, Bart. Airlines oppose higher taxes to fix federal budget. USA Today, December 5, 2012.

[5] National Airline Policy. America Needs a National Airline Policy: Five Policy Priorities. Viewed April 2, 2013.

[6] Stellin, Susan. Airfares with Less Fine Print. New York Times Business Day, December 26, 2011.

[7] Yamanaka, S., J. Karlsson, and A. Odoni. (2006). Aviation infrastructure taxes and fees in the United States and the European Union. Airlines, Airports, and Airspace: Economic and Infrastructure Analysis (Transportation Research Record No. 1951), 44-51.

Fact Checking: Pristiq

Http iframes are not shown in https pages in many major browsers. Please read this post for details.

(Transcript at end of post.)

Perhaps this speaks to the kinds of TV shows I watch (I neither confirm or deny that I am in a betting league for The Bachelor), but I am often disturbed by the pharmaceutical commercials that run. Too afraid to read more about the lash-growing product Latisse whose possible side effects include respiratory compromise and blindness, I decided to fact-check the commericals for the anti-depressant Pristiq. Given the extreme side effect warnings that include worsening of high blood pressure and suicidal thoughts, it is important that Pfizer is not just manipulating us into think Pristiq is effective. The manipulations include employment of “common people” and glittering generalities.

The central claim I wanted to investigate was the one that Pristiq was a key in helping treat depression. The Pfizer website claims that according to 8-week studies, 50mg Pristiq yielded “significant improvement” in depression symptoms. On the 17-point Hamilton Rating Scale for Depression, patients reported a 52% improvement from the baseline (-12 adjusted score) as opposed to a 43% reported change from using a placebo (-10 adjusted score) [1]. For large n this can be significant improvement over a placebo [2]. Something to watch out for, however, is that this commercial suggests that Pristiq is useful for treating depression of all severities. According to this 2010 Wall Street Journal article about a study in the Journal of the American Medical Association, researchers at the University of Pennsylvania found that antidepressants seemed to help the severely depressed (improving them as much as 4 points) while having little effect on the mildly depressed over a placebo.

In fact-checking the effectivness of Pristiq, a point of concern was the eight-week period of the study: there is a bit of media buzz about the “poop out” effect (antidepressant tachyphylaxis). According to a 2011 article in the Journal for Psychiatry and Clinical Neuroscience [4] as well as a Johns Hopkins Health Alert [5], this tends to be observed in selective serontonin reuptake inhibitors (SSRIs) rather than serotonin-norephinerphine inhibitors (SNRIs). According to PubChem, Pristiq’s major active ingredient desvenlafaxine is a selective serotonin and noreprinephrine reuptake inhibitor [6]. Thus Pristiq seems less likely to lose effectiveness and the short trial period should not decrease our confidence in the long-term advantages of Pristiq.

The reported side effects of Pristiq, which include suicidal thoughts and the worsening of high blood pressure, seemed alarming. It seems, however, that these side effects are common to SNRIs [7][8][9][10].

In conclusion, it seems valid to tell TV audiences to talk to their doctor about this drug.

[1] Pfizer description of Pristiq.
[2] E-mail exchange with Adeeti Ullal, Ph.D. candidate at Harvard-MIT.
[3] Effectiveness of Antidepressants Varies Widely, WSJ, January 2010.
[4] Katz, Gergory, MD. Tachyphylaxis/tolerance to antidepressants in treatment of dysthymia: Results of a retrospective naturalistic chart review study. Journal for Psychiatry and Clinical Neuroscience, August 2011.
[5] Johns Hopkins Health Alert: Antidepressant Medication “Poop Out.
[6] PubChem: Pristiq – Substance Summary.
[7] “Pristiq Prescription Information”. Wyeth Pharmaceuticals Inc. April 2011. Retrieved March 12, 2013.
[8] “Effexor XR Prescription Information”. Wyeth Pharmaceuticals Inc. November 2011. Retrieved March 13, 2013.
[9] “Cymbalta Prescription Information”. Eli Lilly and Company. September 2011. Retrieved March 13, 2013.
[10] “Savella Prescription Information”. Forest Pharmaceuticals Inc. December 2009. March 13, 2013.

Commercial transcript
Woman: Depression is a serious medical condition that can take so much out of you. I feel like I have to wind myself up just to get out of bed. And, well, I have to keep winding myself up to deal with the sadness, the loss of interest, the trouble concentrating, the lack of energy.
Male voiceover: If depression is taking so much out of you, ask your doctor about Pristiq. Pristiq is a prescription medicine proven to treat depression. Pristiq is thought to work by affecting the levels of two chemicals in the brain, serontonin and norepinephrin. Tell your doctor right away if your depression worsens or if you have unusual changes of mood, behavior, or thoughts of suicide. Antidepressants can increase suicidal thoughts and behaviors in children, teens, and young adults. Pristiq is not approved for children under 18. Do not take Pristiq with MOAIs. Taking Pristiq with NSED pain relievers, aspirin, or blood thinners may increase bleeding risk. Tell your doctor about all medications, including those for migraine to avoid a potentially life-threatening condition. Pristiq can cause or worsen high blood pressure, high cholesterol, or glaucoma. Tell your doctor if you have heart disease or before you reduce or stop taking Pristiq. Side effects may include nausea, dizziness, and sweating.
Woman: For me, Pristiq is a key in helping to treat my depression.

My Aspirations for Journalism: Help Us Navigate the Science Branding Game

In fourth grade science class, I learned that taste buds were divided into regions of the tongue. We even conducted a confusing failed “experiment” in which we were supposed to confirm this with taste strips. Year later, in junior year biology, I learned that this experiment failed because taste buds are, in fact, distributed across the mouth. This taught me that textbooks are insufficient for teaching us the science by which we should live.

The alternative, unfortunately, is a volatile scientific understanding that comes from sources of questionable trustworthiness. One day, we read that coffee is bad for health because it increases blood cortisol levels. A few weeks later, we read that coffee is, in fact, good for health because it is correlated with a decreased risk of Alzheimer’s. Most of us do not read the journal articles or talk to the experts. We are left to throw up our hands and browse Reddit instead, resuming our previous aggressive coffee consumption habits.

Branding motivations play a major role of knowledge obfuscation. It is well known that pharmaceutical companies and tech companies try to skew public understanding to increase consumption. There is also another, more insidious branding at work: that of researchers trying to increase their influence. Even the most scrupulous researchers are susceptible to branding pressure: mainstream acceptance is influential in faculty hiring and tenure cases. As a Ph.D. student in computer science, I have learned not to tell you why the ideas I propose may never work. If I am lucky enough to get that media interview, I am not going to say why my problem is not the most important problem or why my solution is not the best solution. Because of the public’s short attention span, presenting catchy sound bytes and oversimplified explanations is good for both my interviewer and me. No wonder the science we get from the New York Times, TED talks, and blog posts is so fragmented and inconsistent.

My aspiration for journalism, then, is for journalists to provide context and curation for scientific knowledge. Journalism can publicize not just results but potential motivations, from funding sources to a scientist’s track record of stances on a topic. Rather than presenting stories in a one-sided way, journalists can solicit multiple expert opinions, including experts outside the area who may not have as much reputation and political capital at stake. Rather than pander to public desire for simple panaceas, journalists can teach the public to embrace complexity by giving people tools to help engage with conflicting opinions. The newest research on whether coffee is good for us may be different than what we heard last week, but if we know how this compares with the whole research trajectory on the topic, as well as expert opinions on the validity of the research, we are able to form educated opinions. It is only with this context and curation that non-experts can have any hope at navigating this science branding game.

A Tribute to the Stars, with Star Pasties

A grainy tape plays in the background. On the stage, a stunning brunette in a blue sequin dress mouths passionate words to a fedora-clad man. He has gone too far–he has found out her secret. She pulls out a gun–and pulls off her dress.

This is Film Strip!, the sexiest tribute to film across the decades. For two weeks only, Boston’s Rogue Burlesque and their guests grace the Oberon stage with this sexy, offbeat, and undeniably funny show. You know you are in for a good time when hostess Liz Fang opens by yelling, “This is a burlesque show… Get off the stage until your clothes are off!”

Film Strip! delivers on burlesque’s promise of sexy glamour: Lily Bourdeaux uses a knife to cut off her sheer black thigh-highs before donning a sequined floor-length black-and-white dress. Film Strip! is funny: the audience could not contain its laughter when Ms. Sassypants strips down from a four-legged sequined preying mantis costume, passionately consumes mad scientist Dan Prior, and emerges with his severed head dripping red-sequined blood. Film Strip! is geeky: Brandy Wine cautiously strips from a hobbit costume to a portrait shrine of Orlando Bloom in The Hobbit. The pieces have a deliciously irreverent attitude towards traditional sexual icons: in its homage to “Some Like it Hot,” Kitty Spanks plays a Marilyn Monroe who, increasingly frustrated with the props in her dressing room to blow her dress upward, decides to take everything off. A true celebration of the female body, Film Strip! features the pregnant Polly Surely in “Satan’s Little Helper.” The show also celebrates the male body: the men of Sirlesque have several numbers, the most memorable of which is the ever-playful “Presidential Undress.”

Film Strip! is what burlesque should be: a many-layered, playful celebration of art and sexuality. With only one week left to see it, buy your tickets now:

Media Diary: My Week in Conversations

This is my response to the first assignment to keep a diary of my media consumption.

This week, I spent 78.6 hours in conversations. (Yes, I realize this is about 70% of the 112 waking hours in a week. This provides some evidence to my hypothesis that the internet facilitates higher volumes of communication by allowing, for instance, simultaneous chat.) The runner-up category of “entertainment” had 22.5 hours. (Something to take into account is that I spent four days with friends in Puerto Rico. A normal week may be more news-heavy or reference-heavy, but conversations would likely still dominate.)

Thus my media diary assignment turned into a closer analysis of my conversations, especially once I realized that the broad categories of “e-mail” and “social networking” that RescueTime provided me did not tell a satisfying story. I was interested in exploring how many people I interacted with over the course of a week, how much of this interaction was one-on-one, how much of this interaction was virtual, and the diversity of this group. As I did not know of a tool to help track this, I reconstructed this information into a spreadsheet detailing my interactions with each person from hand-written notes and message archives.

I defined conversations as any interaction with someone whom I can consider a personal acquaintance (as opposed to a one-way conversation with a journalist or celebrity) that was memorable to me–this includes in-person conversations, chat, and social media interactions such as someone I have met before following me on Facebook or a friend “liking” one of my posts. (I did not include people whose posts I had “liked.”) I spent about half of my conversations one-on-one and half in groups. My one-on-one conversations were split evenly between virtual and in-person; most of my group conversations occurred in person.

This week, I had nontrivial correspondence with a total of 127 people. (This is an underestimate, as I am already remembering other people.) To give some context, I have 996 Facebook friends, 414 Twitter followers, and 514 LinkedIn connections. Most of my correspondence was through direct virtual communication: e-mail, chat/SMS, or phone/Skype.

Here are some interesting numbers from the week. For 55 (43%) of the people I interacted with this week, it was my first interaction with them this month. Of all of these people 21 were completely new; 19 of them I met in person and most of them are outside of my social network. A little less than half (57, or 44%) of these people are associated with work: either at MIT with me or working in my research area. Half (63, or 49%) of people I corresponded with live in the same city; most (121, or 95%) live in the United States. A majority (92, or 72%) are within five years of my age. About a third (46, or 36%) are female. I interacted with 87 (69%) of these people strictly virtually and 18 (41%) strictly in person.

The number of interactions are about what I expected: roughly 10% of the people in my broader social networks. I had more in-person interaction than I expected and met more new people than I expected, though both were probably a bit higher than normal due to travel. I was pleasantly surprised that I balanced work-related correspondence with other correspondence. The percentage of women is lower than I expected–this may be because I have higher-volume correspondence with individual women. From this week’s data, at least, it seems that while virtual communication is useful for helping establish connections with people I have already met who are within my social network, the more serendipitous connections came from traveling. Over the course of a year, however, it may be that how many people I meet virtually (through AirBnB, Twitter, my other internet presence) outside of my social network may balance out with how many people I meet while traveling.

If I had more time to do this assignment, I would be interested in looking more closely at the following:

  • Whose links am I “liking” on Facebook? Whose links am I clicking through across media? Whose links am I sharing across media?
  • How diverse is the information I am accessing through these people? How connected are these people to each other?
  • It would be interesting to have a notion of how much within my social network/sphere of interests these people are so I can have some concept of how likely someone is to change my opinion.
  • It would be interesting to measure different categories of conversation: work, gender, and productivity, to name a few.
  • How does my level of serendipitous interaction look over the course of a year and what does it correlate with?
  • How can I measure attention with respect to conversations?