do not text and drive

Despite what you might have read in the news lately, a recent OR paper suggests that you should not use your cell phone while driving.

Alex Nikolaev, Matthew Robbins, and Sheldon Jacobson recently wrote a paper that analyzed traffic accidents in 62 counties in New York before and after a cell-phone-while-driving ban.  The University of Illinois press release summarizes their results:

The team found that after banning hand-held cell phone use while driving, 46 counties in New York experienced lower fatal accident rates, 10 of which did so at a statistically significant level, while all 62 counties experienced lower personal injury accident rates.

They also discovered that the personal injury accident rate decrease was more substantive in counties such as Bronx, New York and Queens, where there was a high density of licensed drivers rather than in sparsely populated areas of upstate New York.
“What that suggests is, if you have a congestion of cars and you’re distracted, you’re more likely to hit someone,” Jacobson said. “If you have a lower congestion of cars, you’re still distracted, but you’re less likely to hit anyone because there are less people to hit. It’s simple probability.”

The results of this paper are particularly interesting, since it counters what we have been hearing in the news the last two weeks (namely, that cell phone bans do not reduce motor vehicle accidents).

Link: the full paper in Transportation Research Part A: Policy and Practice.

Do you use your cell phone while driving?

OR and the Superbowl?

I hope you enjoyed the game last night. How was OR used for the Superbowl? I enjoyed reading about the logistics of the 2005 Superbowl in Jacksonville, Florida–including using cruise ships to provide enough hotel rooms–but I didn’t see anything similar for yesterday’s game.  Please share anything you’ve found.

Are the commercials worth the price tag now that viewers watch their favorites online? What was your favorite commercial?

The numbers behind health care – for a lay audience

In October, This American Life produced two interesting episodes about health interventions and evidence-based medicine that I am just listening to now. These This American Life shows seem fresh and interesting, largely because they try to discuss health insurance/care issues from a quantitative, evidence-based perspective, even though cite few numbers in the episodes.  To illustrate this point, part of the first episode is even titled “Every CAT scan has nine lives,” referring to the side effects of over-using advanced medical techniques such as CAT scans.

The More is Less episode is particularly interesting for OR folks.  It starts with a twenty-minute discussion of some of the cost and effectiveness problems with health care.  The episode steps through one of the problems that grappled the medical community from the 1970s: geographic disparities in hysterectomies and other medical procedures in Maine and Vermont.  This baffled the doctors, since the disparities in hysterectomies across the state could not be explained by demographics, age, religion, or other factors, as was initially suspected.  One of the doctors who performed some of the analysis (Wennberg) concluded that 70% of women would receive hysterectomies in some communities. Maine doctors concluded that disparities were in part based on the doctors choosing to over-perform certain surgeries rather than the patients asking for procedures.

The episode continues to discuss why performing more medical procedures leads to more side effects and potential destruction in some cases, including the PSA tests for prostate cancer, thus exploring the tradeoff between cost and effectiveness.  These issues have been in the news quite a bit lately, particularly with the chance in mammogram screening recommendations.  These discussions in the news have included too much pandering and too little math and analysis, for the most part.   I’ve struggled to find good resources for a lay audience that address the numbers, so this episode was much appreciated.  I’ll leave the rest of the episode a mystery, so you can listen to it yourself.  The second episode (Someone Else’s Money) is not nearly as good, but examines health insurance in more detail.

Podcast Links:  More is Less and Someone Else’s Money.

Link: The Numbers Guy (WSJ) on Mammogram Math is also an interesting read.

Have you found any good references on lay explanations of the numbers behind health care and health insurance?

a combinatorial coupon challenge

Normally I do not get that excited about grocery shopping, but this Sunday was an exception: I used OR to do my grocery shopping.

Many grocery stores have promotions that encourage shoppers to buy 10 items, such as the 10 for $10 sales.  Oddly, there has typically been no penalty for buying fewer than ten items:  all items would be $1 each.  The sale merely suggests to stock up on ten or more sale items.

This past year, my favorite grocery store (Kroger) has periodically offered a promotion that offers $5 cash back for every ten items purchased (but not any ten items, just the items associated with the promotion). The items range in price from $0.99 to $5.99 before the discount.  The promotion essentially offers an a savings of $0.50 per item, which is 50% off of an item costing $0.99.  But no problem so far; I just partition my wish list into groups of ten and stock up on extra rice cakes until a reach a multiple of ten. [Check out the ad]

There is an additional twist this week: a combinatorial challenge. Most of the items in the promotion have coupons in the Sunday paper that require the purchase of two or three items.  For example, I had two coupons for $1 off of three boxes of Kleenex.  The Kleenex are on sale for $1.49 (not the best sale), would be effectively $0.99 after the promotion (not too shabby), and would be $0.66 after the coupon (a bargain!). But I can only buy them at this price if I buy three or six boxes.  Maybe this isn’t exciting to you, but my kids went through two boxes of Kleenex this week when they had the sniffles.

On Sunday morning, I cut out coupons and made a list of promotion items my family could use.  I made sure that (1) the total number of items on my list was a multiple of ten, (2) individual items were purchased in groups of two or three (according to the coupon restrictions), and (3) individual items were within the lower and upper bounds according to our needs (there are only so many cartons of orange juice we can drink).

I bought 30 items with the promotion, saving $15 with the promotion and an additional $12.30 with coupons. I even found some bargain-priced day old bakery bread. Needless to say, it was the most fun I have ever had at the grocery store.  Kroger has a good OR team, and I appreciate some of the savings–and challenges–they offer shoppers.

Related posts:

• Grocery store market share
• Coupon collecting — and decollecting

Have you ever had a combinatorial shopping experience?

do women read this blog?

I was fascinated by a post about science news and gender that I read this week on the Eureka Zone, a Times Online science blog.

Science research councils have increasingly encouraged their grant-holders to engage with the public about their work and for many research grants some form of public engagement is now a necessity. But whom do these scientists end up engaging?

They find–for reasons unexplained in the article–that men disproportionally read science news.

[M]en are more likely to actively choose to consume science in their free time and the bias is much greater than that seen in formal education.

They report that although half of podcast listeners are female, women make up 10-33% of science podcast listeners.  Similar trends exist for reading science news and blogs. The article does not explain why.  Perhaps there are inherent biases in how news is presented to attract men? I would be interested in a reasonable explanation.

The question is, do women read this blog? Based on the comments, I’d say yes.

I started writing this blog without gender in mind.  At some point, I made a conscious effort to occasionally write more about my personal life as a female geek (like the last two weeks, when I blogged about baking and sewing).  I did this in part to address some of the STEM stereotypes that are decidedly male (like this list–the first item assumes that the presumably male engineer has a wife), which at one time made me feel somewhat excluded from the “real” STEM community of engineer/OR geeks in my case.  I have learned learned that I do in fact belong.  I have met many geeky women who have had similar experiences.  This issue weighs heavily on me, particularly as a university professor who wants everyone to feel welcome to the OR table. I certainly hope everyone feels welcome at this blog.

I don’t want this blog to be a blog just for women, but I wholeheartedly support other female bloggers in the sciences, operations research or otherwise:

• new blog Reflections on OR by Patricia Randall
• Güzin Bayraksan’s OR Blog
• RENeW by Anna Nagurney
• Thoughts on business, engineering and higher education by Aurelie Thiele
• ThinkOR by Dawen Peng
• FemaleScienceProfessor

Do the NBC executives use OR?

People of earth:

I always thought that the television networks used OR/MS to manage the uncertainties inherent in running a television network.  Now I’m not so sure.

I have been following the late night changes with the Late Show.  Last week, NBC announced that they are moving Jay Leno’s prime time show (that airs at 10pm) to 11:35pm by moving the Tonight Show with Conan O’Brien from 11:35pm to 12:05pm.  In response, Conan O’Brien announced that he will quit if his show is moved from 11:35pm to 12:05pm, and a deal is in the works for Conan to leave NBC.

There are two aspects of NBC’s move make me uncomfortable.  The first is that NBC was seemingly unprepared for the the uncertainty surrounding Leno’s primetime experiment.  I find it hard to believe that radically changing the late night schedule was the best contingency plan NBC could have formulated.  It is particularly puzzling, since uncertainty is a large party of television programming.  After all, most new shows are not picked up for season-long contracts.  How could NBC be caught so off guard?  To be fair, prime time shows switch time slots frequently, whereas the scheduling in late night are effectively set in stone, so the expectations are different.  The fact that NBC lags in ratings is a symptom that they are not handling uncertainties well.  I expected NBC to have used more advanced analytical methods to plan for these uncertainties.

My second concern deals with the ethics of such a decision.  Maybe ethics is the wrong word here, since NBC may be legally justified in moving the Tonight Show.  However, it is not the right thing to do. The option to change O’Brien’s time slot should have been immediately ruled out as an infeasible alternative.  The ethical concerns weigh on me as a professor.  The academic analog of the NBC-Conan-Leno fiasco are troubling (my take: NBC is the unethical professor and O’Brien is the bewildered graduate student whose expectations have abruptly changed, and Leno is a graduate student who is benefiting at O’Brien’s expense). In my opinion, students should not be punished because they have not been given clear expectations.  Instead, professors should tell students what they expect.  As a result, I strive to convey my expectations for students–particularly student researchers–because students and professors should be on the same page.

Bottom line: despite what was in O’Brien’s contract, NBC should have laid out their expectations more clearly before making any big changes in his show, since the expectation was clearly not to change the time slot.

To be honest, I don’t have all of the facts, or even most of them.  As a result, it’s hard to say how OR could have been used to manage some of the uncertainties and to make better decisions.  Maybe it’s presumptuous for me to rush to judgment, but it is clear that something is rotten in the state of late night decision-making at NBC.  What do you think?

Haiti and OR

Like many of you, I have been following the news from Haiti with disbelief and incredible sadness.  I am interested in hearing about how OR is helping with the relief efforts.  If you know of any such work, please leave a comment!

OR over the holidays: sock monkeys

Last week, I wrote about how I sewed two dolls this past year. That was not my only sewing project. I also sewed two sock monkeys for my daughters after I found a kit at the draft store. The kit was particularly handy, since the type of socks traditionally used for sock monkeys are no longer produced. I’ve always wanted to sew a sock monkey, since it seems to be one of those projects that “real” sewers undertake. And sock monkeys are cool.

Sewing sock monkeys was considerably simpler than sewing dolls, and I completed the sock monkeys almost entirely on a road trip to a family wedding. Here is a picture.

Punk Rock OR Sock Monkeys

Related holiday posts:

• sewing Punk Rock OR dolls
• baking holiday cookies

Why I like twitter

I signed up for a twitter account last year, and so far have tweeted about 471 times.  After taking some time to make up my mind and find my twitter voice, I will explain why I like twitter.

• Most of the time, I don’t have much to say.  140 characters is about right.
• Most of the time, I don’t want to read more than 140 characters.
• I follow a much more diverse group of people on twitter than I do on the blogosphere.  As a result, I read a more diverse set of news articles than I used to.
• A diverse group of people follows my tweets.  This is a great opportunity to evangelize about OR.
• I digress from OR more with twitter.  Twitter is more casual, so I have fewer posting inhibitions.
• There are fewer commitments/obligations with twitter (as there is with email or blogging).  Tweeting does not take up much of my time.  When I get very busy, I stop tweeting and no one complains.
• Twitter now allows users to make lists.  As a result, there are a number of good OR lists, which makes following OR news much easier.
• On two occasions, I asked about specific software questions (for Windows versions of “grep” and “diff”) and received excellent responses within minutes.  Twitter is handy for things that would take awhile to investigate with google.

I like twitter, but I’m not in love with twitter.  There are a few downsides.

• twitter.com is not ideal for following twitter feeds, although you can download better alternatives.
• I have an ancient cell phone and can’t update my twitter status via texting (I have a pay-as-you-go plan that I am loathe to give up for a monthly commitment), so I haven’t had the true twitter experience.
• The trending topics are rarely interesting, unless you’re interested in keeping up with the most recent Kanye West death hoax.  Once, Alan Turing was a trending topic, which was cool.
• I read fewer blog posts now that I follow twitter.
• I have yet to see conferences effectively use twitter.  It’s hard to attend talks and network while twittering, particularly with an ancient cell phone.

What are the pros and cons of twitter?

OR over the holidays: sewing dolls

Earlier in 2009, I decided to sew dolls for my two daughters.  They each had one handmade doll, and I wanted to give them each another.  I do not like hard plastic injection molding dolls, and due to my love of sewing, I decided to sew two dolls.  When I gave the dolls to my daughters for Christmas, my brother-in-law asked me if I used OR to sew them.  I emphatically answered “Of course!”

Sewing projects take time, and there aren’t nearly enough ways to improve the efficiency aside from buying a sewing machine.  To maximize the probability of finishing the dolls by December 24, I started the project in June.  Here are the steps I used to make the dolls:

1. I first designed a pattern for the dolls based on tracing one of the existing dolls.  I already had some leftover material (fabric, yarn for hair, embroidery thread for faces, thread needles, etc.) from other projects.
2. I then cut out the fabric pieces for the body, labeled the pieces and stored them carefully.  This is my least favorite part of sewing, since it requires meticulous attention and careful measuring.  And my cat usually lies on the pattern and fabric.  The upside is that it did not take any extra time to cut pieces for two dolls than it would for one doll.
3. Once the fabric pieces are cut, the advantage to starting early paid off.  I put the pieces aside until I had a some spare time to embroider the faces (eyes, eyebrows, nose, mouth) at my convenience.
4. Once I had some more spare time, I sewed the body parts together.  This took about half an hour, but it required (1) overhead of setting up the sewing machine and then cleaning up the pile of threads and pins and (2) scheduling this sewing around my daughters’ schedules so that I could sew in secret.
5. Once I sewed the body parts together, I had a collection of legs, arms, heads, and torsos.  I took these parts along with a bag of stuffing on a routine car trip.  I stuffed the arms, legs, and heads while my husband drove.  This took a good hour, mainly because it was difficult to stuff the thin arms and legs.  By the end of the car trip, I had a bag of stuffed body parts, which I’ll admit was pretty creepy.
6. The next step was to sew the parts to the torso and stuff the torso.  This part of the process was fairly simple, but it was too hard to do in a bumpy car trip.  I was able to sew the body parts together while watching TV.  At this point, the dolls resembled voodoo dolls, albeit happy voodoo dolls.  Hair fixes that problem.
7. The sewing of hair is a logistical nightmare.  It involves carefully lining up yarn, taping it to some fabric tape, sewing it to the fabric tape, and then sewing it to the doll’s head by hand.  The hair must be treated with care throughout the process so that it doesn’t get tangled or discovered by my cat, who would immediately turn it into a toy (cats discover the yarn the second you take it out).  I took on the hair all at once.  This process also used the sewing machine, so it had the same issues with overhead, but it was fun to watch my dolls go from creepy to cute.
8. The hair was the last step of sewing the dolls, which I completed in August.  I wanted them to be finished before classes began.  But the dolls were naked and required clothes.  I didn’t have the energy to even think about sewing clothes until the end of October.   I created patterns for clothes at this time, by tracing the clothes to another doll.  I decided to opt for simplicity by only sewing a shirt and dress for each doll (skipping shoes, socks, and accessories).
9. I waited until the Thanksgiving break to sew the clothes.  By then, I was intimidated by the task of hemming tiny clothes (it gets harder the smaller you go, at least for me).  I sewed the doll shirts with great frustration.  As a result, I chose fleece and felt for the dresses to avoid hemming (one of the few ways to improve efficiency).  I sewed five dresses in less time than it took to sew the two shirts.  Later, I embellished the dresses with pockets and a few beads, which only took a few minutes per dress.  The clothes were all completed in a single day.

After describing the sewing process and its frustrations, I should probably explain why I like sewing in the first place.  I enjoy sewing because planning a project from beginning to end is challenging yet rewarding.  I particularly enjoy projects that require me to be creative with my own designs.  All sewing projects are aimed at meeting a specific need in my life, and it feels great to solve the problem with a needle and thread (this is a little easier to see when I am sewing curtains or a chair cover).  Sometimes I use OR to solve problems, and at other times I use my sewing machine to solve problems, but it’s ultimately a similar process for both.

A picture of the finished product is below.  My daughters were happy with how the dolls turned out.  My five year old named her doll after herself (the red-headed doll was made to look like her).  I helped my two year old name her doll after her Great-Grandmother Ena, whose name is easy for her to pronounce (the first rule for naming dolls!–I was inspired by Ralph Keeney’s Value Focused Thinking, which mentions several criteria for choosing names).

Related post:

• Baking cookies

If you sew, please let me know!