Student SPHere 2008-09

This semester, both of my finals come in the form of projects, and I couldn't be happier. For Sudipto's spatial class, I'm exploring a dataset he received from collaborators but hasn't yet had time to investigate. This part of research is one of my favorite-- it's like meeting a new person for the first time, getting to know her, learning what makes her tick. Thus far, I've been tabulating variables to get a sense of the design space and plotting the observations to get a sense of the spatial sampling scheme. Eventually, I'll get around to crafting a few models and fitting them, but right now, I'm enjoying uncovering new aspects of the data at every turn.

This is why I love working on research projects. It's a double-edged quality, the open-endedness and freedom of data analysis versus a theory problem. The latter offers a satisfying solution and finite end. But the former yields so much possibility, so much room for innovation and creativity.This semester has offered me more in the way of projects and analysis than proofs, but I know the pendulum will swing back again before too long.

For Cavan's class, the data scope is more limited, so I am focused on model fitting, selection, and validation. It's a small set of repeated measurements on a single variable with no covariates other than time. These sparse data compel me to focus on getting all the power out of each observation, not wasting any variance, and fitting lean models that are carefully tailored. It's another kind of challenge and I'm having fun with that project as well.

Finally, in my research life, I finished a big revision of the recruitment paper from Positive Connections, complete with lovely new figures and a re-vamped intro and discussion courtesy of my colleague Marg. They may come back with further revisions, but at least it's off my desk for the moment!

As an extra bonus, grading is almost finished for the semester. One more homework in the baby Bayes class and then Brad will take care of grading their final projects. Being TA for this class the second time around was seriously a joy. I have enough experience with the material to feel competent and helpful when students have questions. And having all the solutions written up in advance makes posting solutions keys a breeze. It's a shame most students lack the opportunity to serve multiple times as TA for a class, I really think everyone wins.

Posted at 10:14 AM | Permalink | Comments (0)

Spring is finally sprung here, which is buoying my mood for the end of the semester. One great advantage to the warmer temps is that I am back on my bike, after a very wet and cold winter. Last weekend, I biked to the U Rec Center to try out a Zumba class with my friend Julia. What an energizing, fun workout! Although the U's rec center is a fantastic resource, I don't actually use it (this was my second visit in my 4 years here!), preferring to do my sweating in the privacy of my living room. However, this class was a great chance to "try before you buy," as they say. I'm definitely getting the DVDs now.

It's hard to believe that just a week and a half ago, I was trudging through a snow storm in Sioux Falls, SD! We were visiting to celebrate Tom's mom's birthday. All three of our nephews were there, which made for ready entertainment. The twins are so fun and we had a great time with the whole family. We even took in some theater, watching his cousin play Tevye in Fiddler on the Roof at the local high school.

The big news around Biostats lately has been the public phase of our search for a division head. Three candidates have been invited, each giving a public lecture and having tons of meetings with faculty, students, staff, and administrators. My own adviser, Brad Carlin, is one candidate, so I am obviously a little biased. It has been so interesting to see the different approaches that each brings. I attended all of their public talks and was fortunate enough to have lunch with Dr. Dalene Stangl from Duke, as well. Though the decision is ultimately made by Dean Finnegan, the whole process is remarkably inclusive. The faculty have their say at a big meeting with the Dean, and even students are encouraged to submit feedback on the candidates to the search committee.

On the research front, I am looking forward to the re-submission of yet another paper from the work I did with Simon Rosser's group in Epi. We wrote up a little paper on the recruitment strategies used to obtain a sample of folks who are typically underrepresented in epidemiological studies. Our target population was primarily HIV-positive men of color who have sex with men-- not exactly an easy group to enroll. The paper describes the differences in recruitment effectiveness in terms of HIV status and race/ethnicity, as well as figures on retention and the costs of various recruitment methods. My favorite part of the revision was making new figures. We weren't doing any modeling in this paper, instead presenting descriptive statistics and a few Chi-square tests. Thus, my major contribution as an analyst was deciding how to convey the data clearly. I tried about a dozen ways of displaying the data, finally finding inspiration on one of my favorite blogs, Junk Charts. The authors find many examples of truly awful charts in the mainstream media. This post, in contrast, gave kudos for a set of great plots from the New York Times, which I emulated.

I'm trying not to think too much about how little time is left in the semester. I am thankful not to have final exams, and have been able to get a bit of a start on one of my final projects. If I can keep my research going at the same time, the next few weeks will be really productive!

Posted at 1:17 PM | Permalink | Comments (0)

This week was mixed. Let's get the less good news out of the way first. The Bayes midterm didn't go well for me, which was crushing. I hate doing poorly on exams, particuarly in classes that I really like and care about! The test was very long and I freaked out under the time pressure. Even though the whole class had trouble, I was extremely disappointed with my score. Ah well, I never claimed to be good at doing tricky math stat under time pressure. My mind needs a little time to percolate and think.

On to the good news! Friday, I went out with some fellow students for dinner at Bryant Lake Bowl.

We were a diverse group, hailing from India, S. Korea, Brazil, and the US. It was great to hang out with these folks outside of class, and I really enjoyed their stories of "how it's done back home."

Yesterday, I found out that a paper I had been working on since I was in Epi was accepted for publication. Yippee! That manuscript hung around on my to do list seemingly forever, so I'm ecstatic to have it "in the can," as it were. It describes the relationships among substance use and unsafe sexual behavior in the group of men recruited for the Positive Connections study. All these men were HIV-positive men who have sex with men (MSM) and the sample comprised mostly men of color. We found that substance use was very prevalent and that some drugs, particularly amyl nitrite inhalants, were strongly associated with unsafe sex.

Of course, as with all observational studies, there are a lot of caveats! The sample was a convenience sample recruited to participate in a trial of a behavioral intervention, so they may not represent HIV+ MSM in general. We always have to worry about reporting bias when asking people to disclose very sensitive things such as drug use and sexual behavior. And there were a lot of missing items, so we also have to worry about possible biases due to unobserved data, particuarly if the missingness is informative (e.g., if men with most drug use were also most likely to skip drug-related items on the questionnaire).

Now I'm working to finish up the take-home portion of my midterm for Spatial, which is a nice little data analysis problem. As soon as the midterms are completely wrapped up, I'll start working on the final projects for my classes. I totally prefer projects to exams, so I'm grateful for this structure.

Posted at 1:28 PM | Permalink | Comments (0)

With two exams this week, I am feeling particularly grateful for my Spring Break. Tom and I headed to Tampa, FL for three days of rest and relaxation. It was beautiful, exactly what I wanted.

We spent our days at the park and beach about a mile from our hotel, reading and picnicking and enjoying the warm air. The trees and other flora were amazing-- it's so exciting for this Midwestern girl to see oranges growing in a public park! And of course, palm trees, that universal symbol of tropical destinations, were everywhere.

We stayed away from lavish restaurant meals, but we did dig on some delicious cheap Cuban food one night at La Teresita, including Tampa's own Cuban sandwich. In normal life, I'm a vegetarian (ovo-lacto, meaning I eat dairy and eggs), but on vacation, I usually make an exception in order to try the local specialties.

Tampa's historical Ybor City was a center of cigar manufacture in the 1900's. Cuban immigrants working in those factories are credited with this sandwich's invention: ham, roast pork, swiss cheese, pickles and mustard on some fantastic Cuban bread. Credit also goes to the Italian-American workers inspired the addition of Genoa salami. It was delicious!

Now, the trip seems like a very lovely dream, as I study for my Bayes midterm (Tues) and Spatial midterm (Thurs). It's going to be a serious week, but I am glad to get them both over with.

Posted at 7:27 AM | Permalink | Comments (0)

...are death and taxes, as they say. I've had some of both in my life lately.

My mother's father passed away last weekend, so the big family gathered in Western Iowa for funeral services Thursday. Tom and I drove down, stopping as we always do in Ames, IA, the home of our alma mater, Iowa State University. There are numerous restaurants that we miss, and a drive along I-35 gives us an excuse to stop by our old stomping grounds. It's sweet that Ames is on the route to my family's home, otherwise I can't imagine we would ever go out of our way to visit our undergrad university.

This trip, we got a matched pair of window clings, one for U of M and one for ISU to go on the back window of the car. Speaking of the car, our good old Honda Accord rolled over 100k miles on this trip-- here's to another 100k!

Funerals are bittersweet for those not so directly affected by the death. I was delighted to see my family, but the reason for the gathering was sad. My mother has many siblings, some scattered far across the country. It had been too long since I'd see many of them-- many hadn't even met my partner of seven years.

I thought about the circle of life, cheesy as it sounds. Honoring the passing of my grandfather while cooing over the two beautiful new babies in the family, I couldn't help but think of the way humans have been going on like this for so many thousands of years, one generation fading into the next.

Photo by my dad

We knew the trip home would be dicey, as a large snowstorm was predicted for Minneapolis Thursday night. We made it to about 15 miles south of the Minnesota border before the roads turned icy and we had to abandon the drive. Diamond Jo Casino and Hotel turned out to be the first lodging we encountered, so there we stayed. I'm not much for gambling, being a statistician and all, but wanted to check it off my list of life experiences nonetheless. The gambling floor was depressing-- mostly empty of people, but shrouded in cigarette smoke, half-obscuring the endless blinking rows of slot machines.

My only requirements were that we not spend more than the $9 we had on hand and that I get to pull the lever like in the old cartoons. None of this pushing a button on a video screen stuff for me. We ended up winning $6.20. It was a piddling amount, but enough to make me think of the experience as fun and interesting instead of boring and disappointing. Amazing how that works, eh? :)

As for taxes, well, it's filing season and I'm getting more than my fair share of IRS forms. I won't go into detail, but recent experiences are testing my convictions. I have been learning and living on the taxpayer's dime-- attending public schools and universities, conducting NIH-sponsored research-- nearly my whole life, so I take it seriously that everyone ought to give their share. I just wish it would come out of my paycheck steadily the way it's supposed to instead of creeping up all at once!

The semester is flying by. Already I've graded the midterm for the Intro Bayes class and we have one coming up in Spatial next week. That class has been suffering a series of setbacks thanks to Sudipto's travel, trouble with software on the laptops, etc. I hardly feel that we have enough material for an exam, but we'll see. Bayes, in contrast, clips along as ever.

It's unreasonable to think that I would have gotten everything right the first time, but still it's disappointing when we find errors in the Carlin and Louis solutions manual, as we have a couple lately. My TA for Bayes, Brian, found one today as did Brad. Ah well, we'll correct them on the next go round.

Posted at 7:56 PM | Permalink | Comments (0)

Usually, I don't much care for holidays, particularly ones designed to enrich the cards/flowers/candy/schmaltzy junk industries. However, I very much enjoyed a surprise indoor picnic with my sweetie yesterday. A picnic in February in Minnesota-- who knew? He spread out on a quilt in the living room and assembled a delicious noshing assortment from the Wedge. Ahhh, delicious and relaxing.

Given the relatively easier schedule this semester, I've been able to have more fun than usual. On Tuesday, we gathered a team for Science Trivia at the Nomad. Unfortunately, the usual host, the inimitable Dessa of Doomtree, was ill and couldn't attend. I'm going to blame that fact for our lackluster performance-- come on, 5th place? Of course, no matter the score, science trivia is a great time. It's much more fun for nerds like us than usual pub trivia.

Friday night we hit another great venue, Jitters, for a fellow SPHer's birthday (shout out to Ania!). I was a little cranky from working all day, but man, my mood turned completely around when the fabulous Erin Schwab took the stage. She was belting out great sing-along music, telling stories, and generally putting the packed house into a rollicking mood.

Classes are moving at a great pace. Spatial is taking a little longer to get up to speed, since we have made some detours to explain basic hierarchical Bayesian modeling and learn some R. Bayes, in contrast, is cranking along. Though I've seen many of the problems before, Cavan is putting in plenty of additional exercises to keep me challenged. The weekly assignments require a lot of work, but I am grateful for the practice of doing prelim-level problems.

The work on skin cancer with Bruce is moving ever closer to a second manuscript. We've done some very basic modeling thus far, but I'm starting to see glimmers of much more interesting problems in these data. The ionizing radiation exposure profiles of the cohort members are very complex and interesting. So far, simple averaged or cumulative values sufficed, but I am intrigued at the possibilities for more nuanced descriptions of exposure history.

Brad just landed a new project as well that may form the basis of a PhD project in the future. The data come from a clinical trial, but this is secondary analysis. The investigators want to jointly model short-term patient-reported outcomes (quality of life) and longer-term clinically relevant survival outcomes. The literature has several examples of models that use latent variables to link two outcomes that are correlated and share predictors. The challenge is to do this using time-dependent variables and account for censoring and correlation. At this point, I'm getting acquainted with the existing literature, which is an exciting phase.

My final block of school time is devoted to office hours, which have been great. It's so gratifying to be able to help students. I get to feel smart, they feel relieved-- everyone wins.

Have a Happy President's Day!

Posted at 5:54 PM | Permalink | Comments (0)

Sometimes, the vaunted system of peer-review can put a real damper on one's enthusiasm for a research project. Right now, I am struggling with motivation to work on papers from my last RA, in Epidemiology. I submitted this paper back in October (this was its fourth submission, so I've been shopping it around for much longer) and just heard back at the beginning of this month from the reviewers. Some of the revisions were easy re-writes, but for others, I had to write to my old team, ask for data sets and code files to do additional analysis. Which means dragging my brain back into SAS mode. Not an appealing process, as I have been working exclusively in R and WinBUGS for a year or more.

We also heard back on the main paper on that study; the ensuing process will require all kinds of conference calls and a million emails to coordinate amongst the four+ departments from which the authors hail. In the end, it's worth it to notch another one on the old CV, but by the time things actually get out into the world, they often seem like old news.

Physicists don't do it this way, as I've learned from observing Tom and his colleagues. They have an online site called arXiv where papers are "pre-published" before they go into a peer-reviewed journal. It's a way to get thing out there quickly, hear back from people in the comments, and possibly post revisions, all without entirely sacrificing the standard publishing paradigm. I know some journals are trying to shorten the time line, especially the new crop of online journals. There's definitely something to be said for re-examining the way we disseminate findings.

Posted at 9:30 AM | Permalink | Comments (1)

The first day back is usually a good one. It's great to see people again and hear about all the travel over break. Many of my classmates went someplace warm over the holidays-- smart cookies! I'm scheming for a warm Spring Break, but we'll see if it actually plays out. Tom and I have some workaholic tendencies that usually prevent any travel during that time.

I had both of my classes on Tuesday. First up: Spatial with Sudipto Banerjee. I'm really looking forward to this class because I plan to expand and continue doing research in spatial statistics. We're using Sudipto's text (with Carlin and Gelfand), so I have at least skimmed the material before. It's a small class: seven registered, with only five attending yesterday.

We headed out a bit early to catch Obama's inauguration. It was pretty incredible seeing the university community gathering around televisions set up all over campus to watch the ceremonies. I was packed into a conference room in the dean's office with a few dozen other students, faculty, and staff, watching on a big screen. Every time they cut to a long shot of the millions of people on the mall, I couldn't help shaking my head in amazement. What a scene! I was impressed and heartened by Obama's speech-- it really does feel like a hopeful new beginning, despite the dire economic circumstances.

Then in the afternoon, I headed downstairs for Cavan Reilly's Bayes class. More familiar faces there, and a very familiar book-- Carlin and Louis! I know Cavan will be teaching at a higher level than the Bayes course for which I TA, but it certainly is a confidence boost to be using a text that I know so well. He started off with his version of the "Why do Bayes?" talk that every Bayesian eventually learns how to give. I think that I will face far less skepticism in my career than earlier generations of Bayesians, but there is still plenty of educating that can be done of frequentists and those who are outside the debate entirely. One great point he made was that in the sorting of statisticians into a 2x2 table with Bayesian/Frequentist on one axis and Parametric/Non-parametric on the other, it's the latter division that has a much greater impact on actual practice.

Posted at 8:13 AM | Permalink | Comments (0)

Well, I am feeling rested and energized now! We spent a week visiting my family in Kansas City, MO for the holidays. The biggest event by far was my sister having her first baby, my beautiful niece Nithya.

It was such a powerful experience coming together as a family to welcome our newest member. Everyone was excited about the baby, passing her around like an adorable little hot potato all week. I am so proud of my sister and her partner for their teamwork as new parents-- I think they'll be great!

I also spent a day in Lawrence, KS with my other sister, visiting the campus where she is a graduate student in Communication Studies. Her office is so nice compared to mine, I can't say I wasn't a bit jealous.

It was striking how different the academics, assistantships, and culture are in her department versus Biostats. I swore up and down that she wouldn't really be teaching, but lo and behold, she was the sole instructor for two sections of an intro speech class. This is far more intense than any teaching assignments I've heard of at U of M.

And she's friends with everyone! There's precious little socializing among Biostats students, particularly outside of school, but her whole incoming cohort (masters and PhD together) seems to be one big happy family. Also a little jealous of that, to be honest. Watching my little sisters grow up into moms and grad students is crazy, I tell ya.

Another bright spot of the trip was receiving my grades. Despite an unusually challenging semester, all my years of schooling experience paid off again. This is definitely encouraging me for next semester, when I am scaling back to only two classes, but both of them doozies.

With a whole week and a couple of weather-related schedule changes, I even managed to get some work done on the Carlin and Louis solutions manual while at home. We're planning to have it done for spring semester textbook adoptions, so I need to get the last few problems worked out before the new semester. It's a little frustrating to be this close to finished and still struggling with odds and ends, but I know I'll be excited to see it done.

And finally, a little plug for some computer nerdery. My dad helped me update my installation of the open-source operating system Ubuntu, which is a flavor of Linux. I only run one program that requires Windows (WinBUGS; though there is a LinBUGS for Linux, it doesn't seem to be very dependable yet), so I still dual boot. Eventually, I would like to free myself entirely from the evil Microsoft empire. So far, R, LaTeX, Firefox, and Picasa-- all the things I use most frequently-- are working beautifully.

For people not acquainted with the open-source movement, in short, the benefits are free software, better security, and a poke in the eye to "the man." Particularly as an academic, I love the freedom, openness, and spirit of innovation that go with open source.

Have a safe and happy new year!

Posted at 12:11 PM | Permalink | Comments (2)

Only a few days left! Man, I've been waking up thinking about my Linear Models take-home exam and going to sleep thinking about my Linear Models take-home exam. The report is up to 15 pages, but my pace has started to slow. In these last two days before it's due, I need to focus my efforts so I can also do a bit of studying for the Survival final on Thursday. It's incredible how all this anticipation and wishing it would be over turns into anxiety about not having enough time before it's all finished!

My presentation in Probability Models last week went really well. I devoted a lot of time to preparing the talk, hoping to dig out of the hole I got into with the midterm. My partner Tom and friend Julia were both super helpful, listening to my talk, asking questions, and giving great suggestions. It was an interesting paper, and I think I managed to get across all the salient points without losing anyone during the technical hand-waving. Whew!

This Thursday, I took a study break to represent Biostatistics at the annual SPH Winter Reception. This year's theme was something about Hollywood, so we played an adaptation of Hollywood squares, with SPH faculty on the panel instead of C-list celebrities. Our team didn't fare so well, but we had fun and a few laughs at the "creative" answers given by panelists. Here's a picture of Dean Finnegan and me, yukking it up.

Tomorrow, we're getting blasted with some truly awful temperatures (a high of -6 F with windchills as low as -30 F). I'd like to stay in my cozy condo, since I have no classes, but I have to take the bus in to school for my annual tuberculin skin test (TST). SPH, as part of the Academic Health Center, has more serious communicable disease prevention programs than the rest of the U. Everyone must have documented immunizations to register, as well as annual TB screening. This year, two of my public health efforts inadvertently clashed. My first attempt at the TB test was aborted when I discovered I'd had my flu vaccine too recently. For most people, this wouldn't be a problem, but I always get the live attenuated version (FluMist), which would interfere with the TST. Ah well. I won't feel like a real Minnesotan unless I get out there and truly experience the ludicrous temperatures.

Posted at 9:46 PM | Permalink | Comments (0)

Tom and I spent Turkey Day in the wilds of South Dakota (actually, just outside Sioux Falls) with his family this year. When I wasn't working on a final project, I was having a great time. His folks are delightful: so hospitable, always up for card games, talking politics, or getting out and about.

Highlights of the trip were decorating cookies with his cousins, watching Barack and Michelle's interview with Barbara Walters, playing dominoes with Grandma, and getting in way more exercise than usual. Tom's parents keep the house pretty cold, so it pays to keep the blood flowing.

On the school front, I am delighted to report that this semester is going out with a whimper, not a bang. Thursday, I'm presenting a talk on a paper written by my academic uncle Sudipto Banerjee with my academic grandfather Alan Gelfand. It's a reasonably technical paper and uses a kind of stochastic process we never covered in class, so I've got my work cut out for me. Then one more quick HW and I'm done with probability models!

On Friday, we'll get a take-home final for Linear Models, due 1 week later. After that, I have almost a whole week to study for my Survival final-- awesome! After such an intense semester, I am incredibly grateful for an easier-than-average finals season.

Posted at 7:26 AM | Permalink | Comments (0)

I've been a student for a darn long time. Like 18 out of 25 years of my life. Yet every semester there are still valuable lessons to be learned. I'm not talking about the material of the courses; clearly, that changes with each new class. No, I'm talking about the business of being a successful learner.

Over the years, I've learned a great deal about my style of learning. I've learned how to build that perfect crescendo to an exam, winnowing my notes into increasingly concise form as the date approaches. I have perfected my style of reading text books, sometimes reading aloud to keep myself focused, and always underlining key words with my favorite pink pen. I never sit down with a math book without also having pencil and paper at the ready so I can follow along with derivations and examples. My computer files are a paragon of organizational virtue, all filed and neatly labeled by date and course.

Despite all these important bits of self-knowledge, I don't know it all. (Surprise!) This semester has yielded up yet another gem gleaned the hard way. My hard-won pearl of wisdom for Fall 08 is to be proactive about utilizing your resources early in the semester.

My first mistake was failing to get copies of the several "optional" textbooks for the Probability Models class, meaning all I had to read was my professor's handwriting. Not the most legible or clearly organized, and I didn't realize she was directly following different textbooks at various times.

Worse, I was too intimidated to ask fellow students to work with me on homework assignments until very late in the game. Part of my hesitance was my strange status in the department, no longer really an MS student, not yet a PhD student. I regret that now, and having gotten my act together, remembered that it takes very little in the way of collaboration to make a huge difference. You can bang your head up against a problem for a week, when all it takes is a short conversation with a fellow student about what the question is really asking to set you on the right path.

There's a happy ending to this story. I thought I was merely procrastinating working on the latest homework for Probability Models. We've had several weeks for the assignment, but every time I worked, I found myself wanting to spend time on the texts and notes. I wanted to work exercises, look at examples, think about the material. Turns out this was a great strategy. When I finally did start on the problem set, my mind had gelled into a better understanding of the concepts and I was able to make significant headway.

It's too bad that I'm just now really grokking the material, especially after a disastrous midterm, but I'm glad I learned the error of my ways before it was positively too late. This material will show up again in August on the PhD exam, and I have renewed confidence that when it does, I'll be ready.

Posted at 2:07 PM | Permalink | Comments (1)

This election season has been the most intense of my life. Much of the reason is my personal level of interest, but I'm also laying some blame at the feet of modern communication infrastructure. There was a day when you might have to read a few newspaper columns or wait for a candidate's whistlestop tour to visit your town, but today an average citizen is awash in the election for 21 long months. In my personal life, I have practically no exposure to TV or radio ads, but I *love* sites like fivethirtyeight.com, intrade.com, 270towin.com, and of course, Gelman's blog.

The amount of data surrounding an election is tremendous, and the possibilities for longitudinal analysis, predictions, regressions, etc. etc. are endless. The prediction aspect was particularly prominent this year, as Nick Silver found himself helming the preeminent poll analysis site. His approach is Bayesian-flavored, so I have special love for his simulations. So that was my statistical take on the election-- tons of data.

My personal take was relatively blasé until this week. Then on Monday, I suddenly found myself bouncing around like a hyperactive tot on holiday eve, excited to get to bed so that the morning would come faster and I could rocket out of bed and into the voting line. Tom and I were in line by 6:45 and I ended up being the 83rd voter in my precinct. It was a tremendous feeling seeing all my neighbors out with their kids and pets and coffee cups, shuffling along the side of the Whittier Community Center in the wee hours. Rarely am I moved to anything close to nationalism, but I certainly felt a swell of civic pride that morning. The day was a blur of endless blog reading, stat checking, nail biting, and generally failing to get work done.

That night, we gathered with friends for a results-watching party, all of us mentally adding up the electoral votes not only to determine the election, but also to see how we were doing in the betting pools. :) In the end, I didn't give a hoot about my pool guesses, because when Obama gave his acceptance speech, there were tears in my eyes and renewed energy in my heart. It's been a long time since I felt good about this country's leadership. It is a strange feeling, but I am so optimistic about the future, and the potential for President Obama to undo some of the damage in this country.

Coming into class the next morning, I was unable to share much of my enthusiasm with my classmates or professors, all of whom are citizens of other countries. They were interested in the outcome, but it was a stark contrast to the obsessiveness with which my other friends have been following politics this past year.

Ok, enough of the politics. The other fun event was Halloween. Actually.... there's still more politics, because Tom and I both dressed as political stereotypes this year: Joe Six-pack and a Hockey Mom. And given our set of friends, we weren't the only election-oriented costumes: John McCain and Sarah Palin were there, as was Obama's grandmother (may she rest in peace). Halloween may in fact be my favorite holiday. I love to see people getting so creative and not taking themselves seriously at all.

Posted at 11:39 AM | Permalink | Comments (0)

Last week was midterms and it was so miserable, I honestly don't want to write about it. I had two assignments and three exams and it was the opposite of a good time. BUT! At the end of that painful slog, I drove home to Kansas City, surprising my sister Lydia at her baby shower. I got everything I wanted in a visit to my family-- delicious food, seasonal crafts, engaging conversation, beautiful weather, and tons of love. My family is the best; it totally revived me to spend some time with them. A few pictures are below.

Me with my two amazing sisters.

Rachel and I working on jack-o-lanterns.

The finished product!

School hasn't taken much of a break in the aftermath of exams, though professors have been generous in moving due dates to give students some room to breathe. After feeling frustrated with my performance in Probability Models, I have resolved to stop working in isolation all the time. A couple of fellow students and I are going to start meeting to review materials and work on assignments together. Discussing problems with others can be such a boon to solving them.

The Linear Models exam was a take-home and actually kind of cool. It was equal parts calculus (finding MLEs, getting gradient vectors and Hessian matrices) and computing (writing the code to implement Newton-Raphson). In the lecture, we are seeing even more ways to get around that old problem of high-dimensional integration. For generalized linear mixed models, we're discussing how to use Monte Carlo simulation via Metropolis-Hastings, rejection sampling, independence sampling, or Gibbs. It's nice because Markov Chain Monte Carlo is being covered in both Linear Models and Probability Models at the same time. I like hearing about the same material from two very different perspectives simultaneously.

Posted at 11:33 AM | Permalink | Comments (0)

Last weekend, I fully embraced the spirit of autumn. On Saturday, I baked cupcakes using selections from fall's delicious bounty: beets, carrots, maple syrup, and chevre. Ok, maybe that last one's not particularly seasonal. They came out so well! My sweet tooth is strong, but my love of complex flavors is even stronger. These were earthy, rich, tangy, and sweet all at once-- yum! The recipe can be found at one of my favorite food blogs, Apartment Therapy: The Kitchn. I made a full recipe of the cake and a half recipe of the frosting (altered only to incorporate some of the vanilla beans I bought recently) and baked them in cupcake tins instead of assembling as a layer cake.

Sunday, I visited the MN Landscape Arboretum for the first time. What a gorgeous place! It seemed that approximately everyone else in the state had the same idea-- it was crowded with wedding parties, families taking portraits, and couples enjoying the crisp air. The ground are expansive, so we were able to wander off onto quiet wooded paths. This is truly a gem of Minnesota; I only wish I hadn't waited three years to visit. Next time, I will make sure to plan time to visit the apple house for some U of M developed apple varieties. P.S. Students at the U of M always get free admission!

This week has been the usual array of challenges and opportunities at school. My former colleagues in the Division of Epidemiology have kicked my butt into gear to re-submit a couple of manuscripts that have languished since I left last year. Thanks, guys! So I hope that soon I will have a couple more pubs to add to the old CV. Speaking of which, I updated it as part of the application process to the PhD program in Biostats. I'm sure that SOPHAS is a great tool for new students applying to a number of schools. But having used it twice now to apply to just one school (the U of M, where I was already a student both times), it's starting to get on my nerves. None of the information is saved from previous application cycles, so I had to repeat the hours-long process of manually entering every class I've ever taken. Whew-- I've been taking college coursework for 10 years now. Seriously, when people joke about being a professional student, I'm pretty sure I'm exactly what they have in mind.

In Probability and Linear Models, we have embarked on the core material, having finished mathematical preliminaries. It's pretty interesting to learn what happens inside the black box of SAS or R when one fits a generalized linear mixed model, for example. One huge benefit of computing has been to find ways around the hurdle of high-dimensional integrations. This is a concept that keeps coming up again and again. Statistical estimation relies on the likelihood, which is generally a product over observations. If we want to rid the likelihood of some parameters (like random effects in the case of mixed models or the model parameters themselves in the case of obtaining a marginal for Bayesian inference) the necessary integration can get hairy very quickly. With modern computing, we can take a variety of approaches: Expectation-Maximization (EM) algorithm, Monte Carlo integration, quasi-likelihood optimization, Gaussian quadrature, etc. etc. Good stuff!

Posted at 12:02 PM | Permalink | Comments (0)