Tuesday, August 04, 2009

Thinking about textbooks

In preparation for this morning's meeting of the genetics-revision committee, each committee member has reviewed one of the candidate textbooks. I had prepared a checklist of content and presentation issues to guide this, but it was a quick-and-dirty list and I now realize that I overlooked some big issues.

I realized this because I spent yesterday afternoon reviewing the textbook I'd taken on, and yesterday evening reviewing the first (draft) chapter of a new genetics textbook. I filled in my checklist for the former, and sent the publisher a lot of detailed comments on the latter, but now I think I need to add to my checklist and send a second email clarifying the big issues.

One is the difference between information and science. I really like the first-year textbook I've been using (Scott Freeman's Biological Science) because of its explicitly scientific approach. Each topic is introduced as questions: What do we (students and researchers) want to understand? What are the hypotheses? How have they been/are they being tested? Neither genetics textbook I reviewed does this. Instead they present lots of information, but as facts and history, not science.

The history aspect is a big problem. Classic experiments are described in detail, but it's not at all clear why students should know these. The strongest original evidence that DNA carries genetic information came from two now-famous studies. Both textbooks explain these well, but doing so requires explaining a lot of technical details about the experimental systems used (pathogenic bacteria and bacterial viruses) that does nothing to advance students understanding of DNA's function. If we just wanted to convince students that DNA does carry the genetic information, there are many simpler experiments available now, such as transforming bacteria with a plasmid carrying an antibiotic resistance gene. If we want students to learn history, we need to know why.

Another problem is telling the students why particular information is being presented, how they are expected to use what's in the chapter. The draft first chapter I read was densely packed with information on an enormous range of topics: the history of genetics, the structure of DNA, how gene expression works and how it is regulated, how evolution happens, the first organisms, how evolutionary relationships are inferred from DNA sequences. The authors' Prospectus indicated that they think students will already know a fair bit of this, but the students aren't told how they should use this information. Is it meant to be a review? Will they need to know this in order to understand the following chapters?

The techniques students are taught are strangely archaic. Nobody does Southern blots any more, or scores restriction-fragment-length polymorphisms! Instead, modern genetic analysis is based on direct determination of DNA sequences. The technologies that do this are complex, but analysis using sequences is (should be) much more intuitive for students to understand than the indirect methods we used to rely on. I can appreciate how an old textbook such as IGA might be conservative in the methods it describes (laziness, partly, and lack of imagination) but this is inexcusable in a completely new book.

Thursday, June 18, 2009

Results of the textbook meeting

Yesterday's meeting with the textbook rep clarified several tasks (listed below in no particular order):

1. The textbook publisher could create a composite textbook for us, made up of chapters taken from two or more different sources. But this may be more suitable for a survey course than for one that gradually builds expertise The rep will find out whether instructors have used 'composite' textbooks for courses like ours, and if so will put us in touch with the instructors.

2. The textbook rep will find out about on-line genetics activities provided to students with the various textbooks. And we will go through the activities she discovered in one of the textbooks, to find out how suitable they would be for our students (we want interactive activities, where students have to make decisions about what to do and predictions about what will happen).

3. I will dig out our old autotutorial genetics material to see if some of that could be repurposed for this new course.

4. We will look more carefully at the available textbooks, to determine what might be suitable. In particular, is there a textbook that would be OK if it were supplemented with one or two chapters from other sources, or with a week or two of material we have written specifically for our course?

5. We will prepare an email to the authors of existing genetics textbooks, explaining the approach we want to take and asking if they know of any suitable textbooks or other resources. (We'll also give this to the rep.) Here's a draft for us to start with:

Dear genetics textbook author,

We are planning a new genetics course for second-year biology majors, but we haven't been able to find any textbook that uses the approach we think best (described below). So we're contacting the authors of respected genetics textbooks to ask if they might know of something suitable.

After many years of teaching genetics, we feel that understanding the core of genetics has three main components. First, students must understand meiosis and its genetic consequences - how parental genotypes give rise to gamete genotypes, and how random gamete fusion creates offspring with new combinations of parental genotypes. Second, students must understand how genotypes produce phenotypes - how genes and proteins work, the role of environmental variation, how changes to DNA sequences change gene activities, and how, in diploid organisms, different alleles of the same or different genes interact to produce phenotypes. (This last point is perhaps the most important: students need to understand the molecular basis of dominance and epistasis.) Finally, once students have some mastery of both inheritance and phenotypes, they must learn to put these together to understand how phenotypes are inherited.

We have been unable to find any textbook that takes this approach. Traditional (Mendel-first) texts throw students in at the deep end, asking them to start applying Mendelian principles without any explanation of their causes. Even the simplest Punnett square implicitly requires students to figure out parental genotypes from parental phenotypes, to predict the gamete genotypes and proportions these parents will produce, to predict the offspring genotypes and proportions that random fusion of these gametes will produce, and to predict the offspring phenotypes from these genotypes. It's not surprising that students cope by blindly memorizing rules and patterns. Later chapters in the textbooks do teach meiosis and gene action, but most students treat these explanations as independent facts to be memorized, and never really make the causal connections between them and the rules and patterns they began with. (If you doubt this, try asking students why we see dominance.)

DNA-first textbooks give students all the facts of molecular biology before introducing Mendel, but students are unable to use this information to predict phenotypes because the text spent no more than a paragraph on the critical issue of what happens when two different alleles are present. And meiosis is again treated largely as patterns to be recognized, with no emphasis on using it to predict gamete genotypes.

Here's what we think is missing from the textbooks we've examined:

  1. Material that teaches students to predict gamete genotypes from parental genotypes. This needn't be a chapter in itself, but the basics should be introduced when meiosis is introduced, and extended when each new complication is brought up. For example, when crossing over or chromosome rearrangements are taught, students should also be taught how to predict the gamete genotypes that crossovers and rearrangements will produce. This teaching needs to be accompanied by appropriate problems. For example: "A man has genotype a1 a2 b1 b2. What gametes will a single meiosis produce? What gamete genotypes will the pooled products of many meioses contain, and in what proportions?"
  2. Material (probably a chapter) that teaches students about the causal relationships between diploid genotypes and diploid phenotypes, explicitly incorporating the molecular basis of each effect. (Haploids could also be in such a chapter.) What if one allele produces a functional enzyme but the other produces no enzyme at all? What if one allele of a repressor gene is defective?

We really don't want to write our own textbook, or even our own supplementary chapters. Might you know of any textbook that takes the approach we're looking for. Or perhaps just a chapter or other written material that could fill in the gaps we see?

Thanks very much in advance for any suggestions,



Wednesday, June 17, 2009

Preparing for a meeting with our usual textbook rep

The Genetics planning committee (or whatever we are) is about to meet with a textbook rep to discuss options for getting a textbook that fits the kind of course we want to teach.

I'll set aside for now the issue of whether we want a textbook at all. I think we need a set of required readings (and maybe activities) that we expect students to complete BEFORE they come to class. These could be chapters of a standard textbook, a collection of chapters from different textbooks that a publisher has put together for us, stuff we wrote ourselves, or ???

Why a typical genetics textbook isn't suitable:

Our approach to teaching genetics seems very sensible to us, but it's certainly not the one most courses take. Genetics courses and textbooks usually start either with some combination of Mendel's discoveries, meiosis and DNA/gene expression, introducing the basics of what's called 'transmission genetics'. They may do Mendel first, or DNA first. Students learn to predict phenotypes of offspring from phenotypes of parents, and vice versa. To do this they need to have memorized 'Mendel's laws'. In this context our current understanding of molecular biology and meiosis is presented as explaining what Mendel found.

We instead want to separately teach the two components of transmission genetics. We will thoroughly teach how DNA sequences determine phenotype, building a solid molecular foundation for such concepts as ploidy and dominance. Separately we'll teach how DNA sequences are inherited (meiosis, gamete fusion, chromosome reassortment and crossing over, etc.). Only once both components have been solidly established will we combine them to teach about the inheritance of phenotypes.

But we can't find a textbook that does a proper job of teaching how genotype determines phenotype. This deserves at least one full chapter, maybe more, but textbooks usually gloss over it, assuming that students who understand how DNA makes RNA makes proteins and what proteins do will automatically grasp the implications for phenotypes, especially in diploids.

So maybe the solution is to use a standard textbook but somehow create this anomalous chapter ourselves, or find it somewhere outside of the usual textbooks.

Wednesday, April 15, 2009

Improving the match between objectives and assessment

Yesterday one of the biology instructors presented to the rest of the first-year instructors the results of an analysis she'd done.  She wanted to find out whether the 'learning objectives' we developed and are trying to follow match what we actually assess in our midterms and final exams.  The issue wasn't so much about their content as about their level of difficulty, which she scored using the 'Bloom's Taxonomy' scale.

What she found was that our exams ask more of our students than they would expect from the learning objectives we give them.  Even our multiple choice questions are quite challenging, mostly requiring a lot more than simple regurgitation of factoids.  This is good in that we're assessing learning at the level we want, but bad in that we're not telling students the truth about our expectations

The cause of the discrepancy is that we are all relatively new to writing learning objectives. When we wrote them (as a committee) we focused more on content than on what we wanted our students to be able to do with the content.  We knew enough to use 'performance' verbs like describe, list, and explain rather than 'state' verbs like know and understand, but we needed to also use words like predict, interpret and deduce.

The instructor fixed the objectives for us - she went through all of them (about 50!) and rewrote them to reflect what we are actually assessing.