I get many e-mails from companies that say they have the secret to picking the best college for their clients. I usually delete those messages, since there is no way to verify their claims. But last week I had one of my very infrequent good ideas, a method for putting such companies to the test.
I invented a time machine.
_____About the Author_____
Jay Mathews, a Washington Post education reporter, writes a weekly Class Struggle column exclusively for washingtonpost.com. He also covers school issues in a quarterly column for The Post Magazine. He can be reached via e-mail at firstname.lastname@example.org.
It wasn't that hard, and it was a rather cranky invention, giving me different results every time I used it. But I thought it had possibilities, and since I was too lazy to take it to the U.S. Patent Office (even though I drive by its big offices in Arlington every night after work), I decided to share it with the world for free.
I got this inspiration while reading the latest press release from one of those aforementioned new and exciting college search companies. It is called Destination-U.com. The Los Angeles marketing agent it hired, Lisa Townsend, said it "scientifically measures a student's personality profile to determine the best college fit for each student, using proprietary research on college satisfaction metrics."
I had thought that since my daughter, the youngest of my three children, had already gone off to college, I could no longer exploit her as a laboratory rat for my various theories on college admissions. Many relatives and friends had said they were glad that I could no longer make such shameless use of her personal life, culminating in my milking every Katie Mathews story I had for my college admissions guide, "Harvard Schmarvard." But it occurred to me, looking at the Destination-U press release, that they were wrong. The Katie well was not yet dry. She could be my time machine. I could even lure some of her friends into the venture if I asked nicely.
Destination-U has its clients fill out an online survey about what they were looking for in a college -- size, prestige, academic orientation, location, school spirit and lots of other stuff. To test its accuracy, all I had to do was have Katie and her high school friends, now college sophomores, pretend to be high school juniors again and answer the questions as they would have then. If the colleges Destination-U recommended turned out to be the same colleges that they found they liked through months of visits and interviews and arguments with their parents, I could say the company was on to something.
And if their ride two years into the future found something quite different than what Destination-U said was right for them, I could safely conclude the concept needed more work. Many companies use this survey-style college-search method. I thought they all might benefit if they subjected their models to similar journeys in time.
Katie, a patient and loving young woman, said she would give it a try, and she asked some of her high school friends to do the same. Her e-mail to them described my request as "dorky and kind of lame," but we unsung genius fathers are used to such abuse.
Here are the results. Destination-U asks a lot of questions, and then, if you give them your credit card number for the cut-rate introductory fee of $39.95, it presents a list of schools which it calls its "FitList." (Destination-U.com, at my request, gave Katie and friends a code that allowed them to use the service for free, in what I thought was a vain but nonetheless courageous hope that the publicity would be worth it.) The colleges that the Web site matches with each client are divided into three categories, Reaches, Targets and Safeties. It also marks those it considers "best values" based on cost.
My six sophomore time travelers all said the Destination-U.com survey described their college likes and dislikes pretty well. But when it came to picking colleges, the Web site did better in some cases than others.
Two of Katie's friends, Caroline Ward and Willa Brown, thought Destination-U did well. "I had actually applied to all four of the reach schools it listed," Brown said, "and am even at one of those reach schools right now" -- Middlebury College. Ward said "when I applied to college two years ago, I chose the University of Virginia for its location, size, and low cost. Both UVA and UNC-Chapel Hill, two of my top-choice schools, were included on my list of matches in this survey. Seeing as I am happy and doing well at UVA two years in, I would say that the survey did well in selecting schools for me to consider."
Three other friends, Ben Weinberg, Naabia Ofosu-Amaah and Michael Drezner, thought the Web site mostly missed its targets. Ofosu-Amaah, who is at Harvard, found none of the schools she applied to on the Destination-U list, which suggested her desire for a mid-sized campus near a city would be best served by UC-Berkeley, UCLA, Cornell, Penn State or SUNY at Binghamton. Weinberg was also told, wrongly, that he would like big state schools, such as Penn State or SUNY at Buffalo. (Although in the Web site's defense, when he asked what it thought of the school he actually chose, the University of Pennsylvania, it said that also was an excellent match for him.) Drezner was given the names of several small liberal arts colleges, which is what he was looking for, but those were not the names of the schools he actually applied to or the one he is now attending, Swarthmore.
Katie's results fell somewhere in between. Destination-U put Stanford on her reach list and Middlebury on her target list. She likely would have applied to both of those schools if she had not been accepted elsewhere early decision. But it did not mention Pomona, which won her heart very early in the process and which she now attends.
I sought out the owners of Destination-U.com and discovered they were a mother and son, Toby and Greg Waldorf. Toby has been a west Los Angeles educational consultant, specializing in college admissions, for 15 years. She developed the business after a difficult search for a college that would suit her daughter Julie, who had great artistic talent but not the traditional academic interests that make finding colleges easy.
Julie found success at a little-known school in Los Angeles, now known as American Intercontinental University, and has prospered as a fashion designer. Her older brother Greg was a strong student in high school and went off to UCLA and then the Stanford Business School. He eventually suggested to Toby that her knowledge of colleges and his familiarity with computers and the Internet could be combined to produce Destination-U.com.
Toby and Greg were too polite to say so, but I think one reason why their system did not do much for Katie and many of her friends was that their college choices were too predictable. They all went for schools that ranked high on the various lists, and there was no real science in that. Programs like Destination-U prove their worth by finding out-of-the-way colleges that fit individual needs and escape the irrationalities of the selective school admissions process. Even if the computer thought that a student like Ofosu-Amaah was a good fit for Harvard, which she was, it would seem to me a bad idea to make much of this, since applying to a school that selective is like playing the lottery. The losers are just as qualified as the winners and the results are largely dictated by chance.
Katie's friends said services like Destination-U are good for identifying good schools they haven't heard of, and I think that is exactly right. For $39.95, or the similar fees charged by other such services readily available when you Google "college search," it might be worth the time and trouble. And all those companies might consider borrowing my machine and giving every potential customer a free chance to test the system by trying it out on somebody who has already gone to college.
Some of you are probably wondering, did I try out Destination-U myself? The answer is yes, but my results were not terribly useful, and for a good reason. My time machine works pretty well if you ask it to take you just two years into the future, but the number of years I had to punch into the dashboard indicator were just too many, and the whole thing went up in smoke.