The Pitfalls of Comparing Standardized Test Scores

August 23rd, 2011

The US is falling behind other countries in standardized test scores.

Surely this story is familiar to you by now. I’ve been hearing it most of my life. So much so that you have to wonder how US students can ever hope to compete with the test-taking all-stars in China, Korea, and India.

My own experience teaching a few courses in Computer Engineering at a moderately prestigious US university has been that students from overseas did not perform notably better in my classes than domestic students.

At first, this surprised me. After all, these were students from countries with extraordinarily competitive educational systems that had the talent and motivation to study abroad. My expectation was that they would be much better than the Americans. I discussed my observation with several colleagues, and they agreed that international students didn’t do any better in their classes.

I asked a few Indian and Chinese grad students for their take on it. I was reminded of that discussion by a discussion I was having with Robert M. on a message board recently:

RM: 3rd world here, and I was 2 years ahead when I started Middle School in US. Our education system was brutal; no multiple choice, meticulous memorization, and teachers don’t give you any breaks on homework or any other assignments. Definitely run like a bootcamp, but I was half-witted dumbass so to me it was always a nightmare.

What they told me was that in China and India, their education was heavily focused on memorization. Partly this was due to issues of practical logistics - with a graduating class of 1000 students, it is difficult to give a personalized education - and partly it was a cultural issue. When these grad students came to the US, they felt that US students were much better at abstract thinking, reasoning, and analysis than they were.

In the field of educational psychometrics there is a pretty common analytical framework known as Bloom’s Taxonomy that classifies different types of learning by their cognitive function. In Bloom’s terminology, the grad students I interviewed were expressing the idea that at the undergraduate level, Indian education focused on the “lower-order” functions - memorizing facts, understanding concepts, and applying problem-solving techniques, whereas the American system focused on “higher-order” functions - analysis, evaluation, and creation of new ideas.

The lower-order skills are the ones that produce good scores on the common standardized tests. This is because it is typically quite difficult to test higher-order functions in an inexpensive way. Memorization and application map well to multiple choice questions; evaluation and creation do not.

Naturally, this provokes us to ask the questions “is learning targeted at Bloom’s lower order functions better than learning targeted at higher order functions”? It might seem from the ‘higher order’ moniker that the answer is supposed to be “no”, but it’s actually more nuanced than that. The lower order learning is probably better suited to the majority of the people who will be applying their knowledge in industry, where application of ideas is prized. Conversely, higher order learning is better suited to research fields, which prize creativity and analytical abilities.

Now, I want to be careful to say that I’m not trying to make a blanket statement that Americans are better abstract thinkers than people from India or any other country; clearly there are first-class researchers from these countries. I emphatically do not want to claim that every (or even most) American has a razor-sharp analytical mind. Nor that all universities and teachers emphasize the same things. And I most definitely don’t want to give the impression that it isn’t important for students to achieve the lower-order thinking being evaluated on standardized tests. The sharpest analytical skills are of little avail in any scientific discipline without a base knowledge of theory and the ability to apply mathematical tools. That said, the point I’m trying to make is that there are limitations in what we can extrapolate about educational success by comparing test scores, and I say that as someone who is probably a much bigger fan of standardized testing than most of the people here. Mastery of memorization and application can take you to the heights of test scores, but a high test scores doesn’t guarantee a great education.

If it appears paradoxical that the US scores abysmally in standardized testing and at the same time has the best research universities in the world…well, it isn’t. It’s a natural result of a system that targets a different type of learning.

Bumble Bee Moth

June 18th, 2011

Found in the community garden near my apartment.

E3 2011 - Nintendo Conference

June 7th, 2011

Nintendo’s conference is last this year - I’d like to say that they are batting cleanup. With new console hardware to release and the newborn 3DS to promote, we can expect lots of new games. And although a few hardware details have leaked, Nintendo has kept many of the cards close to its chest, both hardware and software-wise. Still, Nintendo’s legacy with the Wii has left a bad taste in the mouths of many enthusiast gamers. I’m willing to be won over by Nintendo, but they’re going to have to show a lot of games, especially those from third parties. Anyway, the show is starting, so let’s see:


Zelda’s 25th Anniversary is the first topic of the conference, and Miyamoto is out right away. We had rumor that there would be a previously unannounced Zelda at this conference. First Miyamoto reminds us of the Zelda 64 remake for 3DS. Next he announces that Four Sword Adventures will be a free download on 3DS (great announcement, one of my favorite Gamecube games). Then Miyamoto spends a lot of time talking about symphonic Zelda concerts & CDs, and the the development team comes on stage to bow - but no footage shown of Skyward Sword and no new Zelda announced (unless it’s later as part of the new console reel).

Next Nintendo brings Iwata out - different flow to the concert this year as Nintendo’s heavy hitters are out right away. Normally at these things it starts out with 30 minutes of back-patting over their sales results. Maybe there is less to be happy about on that front this year, or maybe they got the memo from the other companies.

Iwata teases the “new platform”, but says that we’ll talk about that later in the conference. First a demo reel of 3DS games. We see references to Star Fox, Mario Kart, Mario, Luigi’s Mansion, and Kid Icarus. Reggie arrives on stage to put each of these games under the microscope.

First we see a long clip of in-game footage of Mario Kart 3DS. Reggie notes that in the conference we can only see these games in 2D rather than 3D, but even so the graphics look fairly disappointing. I guess I had higher expectations for the 3DS hardware, which I though was supposed to be similar in power to a Gamecube. This game looks significantly worse, graphically. Game will be out for the holidays.

Star Fox is apparently just a 3D port of Star Fox 64. Don’t get me wrong, that was one of my favorite N64 games, but another full-price port of an N64 game is not what I was hoping for when I saw that Star Fox logo come up. Available in September.

The Super Mario 3DS game is thankfully new. It looks like a mix of NSMB, Mario 64, and SMB3. Probably fun, although nothing jaw-dropping. Game will be available “before the end of the year”.

Kid Icarus is probably the least anticipated of these games - the impressions I’ve heard have not been that positive Out later this year.

Lastly, the sequel to Luigi’s Mansion, the one-off Gamecube game that I have never played. No comment on release date for this title.

The next demo reel is of third-party games. I’m not going to list everything here, other than to say that there was a lot of enthusiast-targeted games.

I make a cup of tea as Nintendo discusses their Pokemon Altered Reality stuff.

OK, with have the conference left, it’s time to discuss the new console: “Wii U” (We You, get it? I guess they wanted an even more embarrassing name to say out loud). Reggie describes the system as “Utopian”. The new controller, as expected, has a large screen on it. It also has dual analog sticks (? Well, Nintendo refers to them as “Circle Pads”) and touch capabilities - it looks kind of like an especially chunky handheld or a tablet from the 1990s. The screen is 6.2″ so it is fairly sizeable.

The controller screen can seemingly do a lot of things - it can stream your game screen while your TV is off. It can run its own touch-controlled games, it can act like a magnifying glass for the main screen, it can show menus or items not on the main screen - neat stuff. This sort of feature was used in games like Pacman Versus, Four Sword Adventures, and Crystal Chronicles to great effect on the Gamecube. That said, the controller doesn’t look particularly comfortable to use for traditional games due to its large size.

We get to see several examples on the demo reel - including an HD Zelda game - but it’s unclear if these are real games or just mock ups for demonstrating the tech - that was the case with most of the games from the original Wii reel.

Iwata mentions that Smash Bros will be developed for both Wii U and 3DS, which embarrassingly gets loud cheers from the audience despite Iwata saying that development has not even started.

Reggie is also quick to dampen hopes that games are far along on the system, refusing even to call the demos on the show floor “game prototypes”, but rather something to the effect of “tech demos” - not entirely promising. We get to see one of these demos in the form of a non-interactive video of carp and sakura blossoms. We hear about 3 more “not actual games” that the press will be able to play, but Nintendo seemingly lacks the confidence to show video of these demos since we only get screenshots. We get one game that Reggie will commit to being real, but it’s an underwhelming Lego game - again with screenshots and not video. What is going on Nintendo?

More talk,as our next reel is respected developers talking about being excited about the system. Many of them mention hardcore games like Batman, Ghost Recon, and Darksiders but no footage of said games. Thankfully, we finally get some video of a few of these games, but mysteriously none of the videos actually show the Wii U controller being used. They could be games on any system. Reggie further lowers expectations by stressing that everything we see is a “first pass” and that the final games will be more impressive.

EA arrives on stage to talk about (rather than show) the ideas EA has for the console. I use this time to put on sunscreen.

Reggie arrives and wraps things up without a big finish.

But that’s not all - Nintendo has an online-only set of demo videos afterward. The first is a Mario-based game that is a blatant rip-off of Pacman Versus, but I guess if you’re going to rip off a game, you may as well choose one as brilliant as Pacman Versus. They should have shown this in the conference.

Two more demos that aren’t very interesting are shown next, and then things are really over.


I have a bus to catch, so I’ll be brief:

On the 3DS front, Nintendo mostly hit its mark. It needed to demonstrate that there would be games coming this year that make it worth getting a 3DS this year rather than waiting for the inevitable 3DSLite. For people wanting typical Nintendo hardcore games, they should be well satisfied. For people waiting for the next Kirby Canvas Curse or Phoenix Wright - we’re not seeing them yet.

Wii U didn’t bring too many surprises that weren’t in the leaks. The concept is not entirely new; well, not new at all for Nintendo. They were doing almost the exact same thing with the Gamecube + GBA link. Fortunately those experiments produced some great games, so I’m looking forward to what I will see in the future from this system. Disappointing that there was almost nothing that Nintendo would commit to as a “real game” and no idea of price or even release date, so this system seems further off than many might have liked.

Still, best Nintendo conference in many years, and probably best of the 3.