More Recent Comments

Tuesday, February 14, 2012

Exam Question #1

It's that time of year again. My students have a mid-tern test on Feb. 28th so I giving out a list of questions that will be on the exam. Here's one.
Here are two different trees depicting the evolutionary relationship of various classes of animals. Which one is better? Why?


The question is based on the following assigned paper:

Meisel, R.P. (2010) Teaching Tree-Thinking to Undergraduate Biology Students. Evo. Edu. Outreach 3:621-628. [doi: 10.1007/s12052-010-0254-9]


[Image Credit: The tree on the left is from Campbell Biology Chapter 32 Activities Quiz (2002)]

48 comments :

Schenck said...

How many of them have a problem with this one?

Larry Moran said...

We'll see. I expect all of them will get a good grade on this question. (They will be able to interpret "better.")

Divalent said...

Well, although both present the same information, IMO the right one is much better because it more clearly illustrates the important point that the goal of evolution of the animal kingdom is the perfection of the nematodes, and that all other phylums (phylii?) are mere degenerate cousins (almost certainly due to random genetic drift, as no one has ever proposed a plausible adaptational explanation to account for the bizarre morphologies of those other groups).

Paul Clapham said...

"Phyla", I think.

Anyway I'm in the habit of answering questions on programming forums. Whenever somebody asks "Which is better" I always have to respond by asking "What do you mean by 'better'?"

So I was interested to see what R. P. Meisel meant by 'better'... but not $34.95 worth of interested.

anomalous said...

I think that they're both equally bad. AFAIK both molecular and (and some) morphological data support Mollusca and Annelida as sister taxa in the larger group Lophotrochozoa which is in turn sister to Ecdysozoa containing Nematoda and Arthropoda. As for rotating on the nodes and demonstrating that both contain the same information, well that is a lesson worth learning.

DK said...

They are equivalent and thus neither can be "better". "Better" for what purpose?

John Harshman said...

Obviously, the one you didn't show, the one that makes Ecdysozoa and Eutrochozoa both monophyletic, is better. But if we have to pick, the left one is better because it consistently ladderizes right. And yes, a lot of people have trouble with rotated branches.

konrad said...

Not fair, the question contains an incorrect claim!

Anonymous said...

Ooooh, I know the answer!

Anonymous said...

Both are woefully out of date. Ecdysozoa and Spiralia are nowhere to be seen.

jaxkayaker said...

I haven't read the paper you require, though I downloaded it for eventual reading. I'd opine that the tree on the right is better because it doesn't present the relationships in the usual, assumed Chain of Being sequence that students expect forcing them to examine the tree to parse its meaning and to examine their assumptions.

jaxkayaker said...

The plural of phylum is phyla, by the way.

Peter said...

Better? Better lay-out? Only different in lay-out. Not different trees. Why ask?

Joachim Dagg said...

I'd argue that the left one is superior, because the stretch between the third and fourth node in the right one shows the poor photoshopping abilities of the 'creator'.:-)

Joe Felsenstein said...

.. and that all other phylums (phylii?) ...

It's "all other phyla",

ergaster said...

Gee thanks, Larry--I am *so* using this image for my own test for my Grade 11 class on Friday! They should be able to identify that the relationships shown are not different.

And the answer I would expect them to give me is "neither is 'better'". I might ask them which one could be misleading, though....

Deb

Larry Moran said...

You get 2/10 for recognizing that the trees are equivalent as far as the relationships between clades is concerned. Unfortunately, that's all you get (20%) because in my class we go far beyond that point.

gillt said...

I claim great ignorance here, but doesn't the one on the right create a redundant node, unnecessarily complicating the interpretation?

gillt said...

How so? As others have pointed out both trees are equally accurate as far as they go. Since when did style and personal preference become relevant in science? Isn't that your preferred mantra for hammering science writers over the head....accuracy accuracy accuracy?

DK said...

The trees are the same, period. Of course I realize that you want the right one to be "better" because it does not have us/chordata "on top". But unless there is another compelling reason to think it is better, it's just silly (and I pity students forced to conform to that silliness). Care to explain for us, unwashed masses, who just see two identical trees? :-) (Our campus does not subscribe to that journal, so I can't see what the author meant).

Joe Felsenstein said...

Larry's example reminds me of a tree I showed in a semi-popular talk I gave at our University. I showed the tree of primates. I hadn't expanded any group other than the apes, so it was very ladderlike. I was about to place humans on the extreme end of the ladder, with chimpanzee as the sister group, when I decided not to pander so much to orthogenesis. So as I got to drawing that part of the tree on the blackboard, I flipped the order of branches and put humans and then, at the extreme end, chimp. After all, this was just as correct.

After the talk someone came up and told me that I had drawn the tree incorrectly.

DR said...

Since the image has no caption and no description, it is difficult to interpret. Is the vertical axis a timescale? Is it to scale? Is the width of the line the error bar? Should the answer be yes to those questions then the graphs are different:

http://imgur.com/rtE1f

And in that case the better graph is the one closer to reality.

If we are to base our interpretation on the paper (which I skimmed) and the context of an exam, then we are probably expected to answer which graph is better understood by students with preconceived notions of biology. In that case the right one might be better, since it is less ordered, and thereby gives less opportunity for false learning by pattern seeking students.

Disclaimer: Not a scientist, didn't read the paper, but happy to post. :D

Neil Schipper said...

With the right one, there is much less ambiguity about which clade to use for the shamash candle.

DR said...

http://goo.gl/oocuW

Larry Moran said...

There are some "unwashed masses" who think that science is a purely objective discipline, unaffected by human emotions, values, and misconceptions. Those people have a great deal of difficulty seeing how anyone could view the two trees differently. They will also be completely baffled by some of the major controversies in science such as adaptationism vs pluralism, the existence of junk DNA, and the importance of gradualism.

They'll never actually be able to understand theistic evolution or intelligent design creationism.

When it comes to teaching, those "unwashed masses" tend to believe that all you have to do is present correct science and all student's misunderstanding and misconceptions will vanish.

I try to teach that science is a very human enterprise and that most of the interesting controversies stem from very human biases and prejudices. I try to teach that students need to recognize the places where misunderstanding and misconceptions can get in the way of a correct interpretation and they need to take steps to avoid or correct those biases whenever possible. Above all, they need to learn to recognize their own misconceptions and biases.

For example, if your audience believes that evolution produces a great chain leading ultimately to humans, then by always presenting trees like the one on the left you are doing nothing to counter that bias. This is an important error of omission, a missed opportunity. It's not a very effective way of teaching to pretend that such biases don't exist.

DK said...

In that context, although I agree with your reasoning, if I were a student, I'd still be MAD, totally mad for getting only 20% for answering that none is better. Because, you see, science is not purely objective discipline, and the left one is only better in your mind.

gillt said...

As an instructor it's okay to encourage your students to care about the controversies you care about, such as theistic evolution or junk DNA, but docking most of the points for answering a trick question correctly is rather heavy-handed and uninspired.

DR said...

How about a proper introduction into how to interpret the data? Like: Vertical axis represents time, horizontal represents genetic variation, neither axis shows cardinal (to scale) values, only the ordinal sequence of the species depicted and their most recent common ancestors, represented by a junction. The timescale is large, very many transitional species have been omitted.

The tree can also be fully characterized by: (Por, (Cni, (Pla, (Nem, ((Mol, (Ann, Art)), (Ech, Cho)))))) with commutativity, but not associativity.

Ex: Iff (a, (b,c)) then ((b, c), a); but not (b, (a, c)). Brackets contain a (subset of a) monophyletic group.

Assignment: Write down a correct and an incorrect variation of the above tree.

Doesn't that sound succinct?

Larry Moran said...

gillt says,

...but docking most of the points for answering a trick question correctly is rather heavy-handed and uninspired.

As far as I'm concerned, your statement exemplifies most of what's wrong with university education. This is not a trick question and no points were "docked." You get points for recognizing the problems with the trees and why there's a difference between them. If you can't do that, especially after reading the assigned paper, then you don't deserve to get the marks.

That doesn't mean you have to agree with me or with the author of the paper. You can easily get full marks by explaining why you think that both trees are just fine in an introductory biology textbook.

Larry Moran said...

DK says,

... if I were a student, I'd still be MAD, totally mad for getting only 20% for answering that none is better.

Then if you were my student I'd take you aside and explain that university isn't kindergarten. When you answer a question like that you are supposed to understand it in the context of what we cover in class and in the assigned reading. I have no tolerance for student lawyers who try to talk their way into getting full marks for an answer that's obviously incomplete. That's not how adults are supposed to behave.

Imagine that you were interviewing for a teaching job and the interviewer asked you the same question. You reply, "The trees are exactly the same. Next question."

The interviewer would say, "Thank-you very much, DK, we'll be in touch."

... Because, you see, science is not purely objective discipline, and the left one is only better in your mind."

Any student who could defend a statement like that would get full marks. But they won't get any marks at all for avoiding the question.

gillt said...

Really? My statement is all that's wrong with the current state of university education?

LAM says, You can easily get full marks by explaining why you think that both trees are just fine in an introductory biology textbook.

I wonder why your students need to be graded on how intro bio textbooks are written. And as far as that's concerned, the article you referenced says,

Unfortunately, many, if not most, of the trees in textbooks are diagonal (Catley and Novick 2008), even though rectangular trees dominate the scientific literature

Based on this one would conclude that the author of the trees repeats the same error that most biology text books make.

DK said...

Sounds to me that you are encouraging conformism over independent thought. 80% for pandering to teacher's favorite pet peeve is quite ridiculous.

Neil Schipper said...

OK, just in case the joke was too obtuse (rather than just lame..)..

During the 8 nights of Chanukka, Jews light N candles on night N, where N=1..8.

The candelabra, or Chanukkya, consists of 8 + 1 candleholders, where the extra candle is for the shamash, which is the only one directly lit and then used to light the N candles for that night.

Typically of the 9 candles of the Chanukkya, one needs to be visually distinct, often by height or position.

In the above trees, only one has the feature "one clade is distinct from the other 8 in having no next-door siblings".

Chas Peterson said...

The trees portray identical information. (As others have pointed out, equally wrong information in light of a ton of more recent data.)
What's the criterion of "better"? To somebody who knows how to correctly read phylogenetic trees, there is absolutely no difference.
If one of them encourages misinterpretation, that's only a problem if you're trying to communicate to people who don't know how to read trees (i.e. they read across the tips, not the nodes).
Because there are no misinterpretations actually portrayed, in an objective sense.

John Harshman said...

If you're looking for a topologically equivalent but harder to misinterpret alternative, then what you really need is the "monkey puzzle tree" conformation, in which the branches are rotated to allow no possible interpretations of a scala naturae in any part of the tree. Why present an alternative that makes only minor changes to the ladder?

Theo Bromine said...

Oh good, I'm not the only one who thought about which would make a better menorah* when I saw these graphs.

* I'm too old to call it a Chanukkya

J. Christian Somody said...

Without a definition of the axes in the question, these are completely equivalent trees. Saying that the correct one is the one where Chordata is not on top, in hopes of removing bias, merely introduces a new bias. This is the phylogenetic equivalent of affirmative action.

Larry Moran said...

You would have received a decent mark for such an answer. For full marks you would have to explain why one bias is worse than another. And you would have to explain the "bias" in the right-hand tree. (Is it nematode-centric?)

You would also have to defend your claim that using one of the trees is analogous to "affirmative action." Good luck with that if you really think the two trees are completely equivalent!

Larry Moran said...

You would have received a decent mark for such an answer. For full marks you would have to explain why one bias is worse than another. And you would have to explain the "bias" in the right-hand tree. (Is it nematode-centric?)

You would also have to defend your claim that using one of the trees is analogous to "affirmative action." Good luck with that if you really think the two trees are completely equivalent!

Steven N. Severinghaus said...

I don't understand all the outrage here. Larry made it clear in the post that the question was directly related to a paper the students would have read. Looking at the abstract for the paper, it seems utterly reasonable to expect that a student who had done their homework would understand the point of the question. It is "objectively" true that these trees are not equivalent because it is clear that different algorithms were used to lay them out.

Schenck said...

I'm with SNS on this, the question is for a class, and if Prof. Moran never discussed the tree or didn't associate it with the paper above, then his only giving partial credit for noting that there can be rotation at nodes would be unreasonable.

And notice that clearly Prof. Moran isn't deducting points for disagreeing with him, he's giving points out for people presenting cogent arguments and demonstrating their understanding of the relevant material.

The Vicar said...

Which chart is "better" depends a great deal on what the purpose of having such a chart is.

Obviously, in light of the reference to the article, if the purpose is "to represent the idea in an abstract way while preventing laymen from getting hold of the picture and using it as evidence that evolution is tending towards a pinnacle", then the chart on the right is better, because it removes that particular bias by mushing the more recent divergences towards the center of the tree.

If, on the other hand, one were going to use a chart of this information strictly as a reference to speed up access to the data -- unlikely in this case, both because of the already-pointed-out outdated research and because these charts don't show enough information to be worth using as a reference instead of memorizing -- then one could argue that the chart on the left is more useful, because it minimizes the number of junctions between any given branch and the main stem. If this were routing data and I were in need of a chart to reference in order to direct packages/data, the chart on the left would win, hands down.

The Vicar said...

Oh, an addendum: if this were a routing chart, even the chart on the left would not be optimal. Swapping the 3-branch sub-tree with Mollusca as its stem and the 2-branch subtree (out of the 5 branches at the extreme right) would allow the chart to become almost nothing but single branches off the main stem.

Chas Peterson said...

"you would have to explain the "bias" in the right-hand tree. (Is it nematode-centric?)"

No, it's Larry-Moran-centric. The 'right answer' is opinion, not knowledge.

Larry Moran said...

There is no "right" answer but you are living in some kind of dream world if you think that opinions aren't important in scoence. Teaching students to recognize the difference between betwen opinion and knowledge is one of my main goals in the course. That doesn't mean that "opinions" are always bad. "Models," "hypotheses," and "speculations" are all opinions of one sort or another. In many courses these knid of opinions are taught as though they were facts.

Most of my exam questions ask the students for their own opinion on controversial topics. Perhaps you are unfamiliar with that kind of question in a university course?

John Harshman said...

The problem with that it that it perpetuates the illusion that there is indeed such a thing as a main stem. That is exactly why Larry doesn't like the tree on the left.

Allan Miller said...

The one on the left is better because it more clearly represents the Great Chain of Being...

Shawn said...

@Schenck, SNS
Yes, it is often the assumption of students that the default score on a question is 100% and there are only marks to be lost. In reality, the default is 0%, and there are only marks to be gained.