Gas station without pumps

2014 October 13

Practice, teaching, or genetics

Filed under: Uncategorized — gasstationwithoutpumps @ 10:04
Tags: , , ,

Mark Guzdial, in The 10K Hour Rule: Deliberate Practice leads to Expertise, and Teaching can trump Genetics | Computing Education Blog, responds to a Slate article claiming that genetics is more important than practice:

Here’s my argument summarized. The Slate authors and Macnamara et al. dismiss the 10K hour rule too lightly, and their explanation of genetic/innate basis for expertise is too simple. Practice is not the same as deliberate practice, or practice with a teacher. Expertise is learned, and we start learning at birth with expertise developing sometimes in ways not directly connected to the later activity. The important part is that we are able to learn to overcome some genetic/innate disparities with good teaching. We shouldn’t be giving up on developing expertise because we don’t have the genes. We should be thinking about how we can teach in order to develop expertise.

Mark’s blog is read (or at least commented on) mainly by teachers of computer science, so he is largely preaching to the choir here. I would like to believe that my teaching makes a difference—I spend almost all my time teaching, grading, or preparing to teach.

I do believe that most students in my classes leave the class with better skills than they came in with.  Whether that is due to my teaching or just to the students being forced to practice is somewhat difficult to determine—to a large extent my teaching style consists of forcing students to practice skills that they’ve generally ignored in the past (like in-program documentation) and providing them detailed feedback on their practice.  I’d like to believe that the feedback (both individual and group) matters, since I give up my weekends to provide the feedback.  If only the practice matters, then I could do as many of my colleagues do and just do I/O testing or delegate the feedback to untrained undergraduate graders.

So I have a bias towards believing Mark’s claim that teaching matters, and that there is a difference between different sorts of practice by students.

But the outcomes for individual students seem to depend more on the students coming in than on what I do.  Those students who come in better prepared or “innately” smarter progress faster than those who come in behind, so the end result of the teaching is that differences among the students are amplified, not reduced. Whether the differences in the students coming in are due to prior practice, prior teaching, or genetics is not really knowable, but also not really relevant.

Mark claims that “Genetics/innate starts at birth, no later”, which is somewhat of a simplification.  Although innate differences are present at birth (by definition), they may not be expressed until much later, either due to the developmental program that coordinates gene expression or due to environmental triggers.  So phenotypic differences may not appear until much later (genes for patterns of facial hair among men generally make no difference until puberty, for example).

He claims that

If you’re going to make the genetics/innate argument, you have to start tracking participants at birth. Otherwise, there’s an awful lot that might add to expertise that’s not going to get counted in any practice logs.

I’ve only had one child that I have taught from birth on (and lots of others also taught him), and we all know the uselessness of sample size=1, so it is not possible for me (and probably for anyone) to track participants from birth for a significant sample size.  But there are certainly ways to estimate the heritability of talent without tracking all activity since birth—the twin studies that he dismisses attempt to do precisely that.  (Some of the twin studies are well done and some are useless anecdotal reports—but there is substantial evidence that some talents have a substantial heritable component.)

Of course, it is always hard to pick apart whether “nature” or “nurture” is responsible for a particular difference in talents, since there is a large feedback loop.  Small differences in initial results can result in differences in how much pleasure practice provides and how much support is given, which can in turn affect how much practice is done and how valuable the practice is.  So small differences in “innate” talent can be amplified to large differences in outcomes.

I’d like to believe Mark’s claim that “Hours spent in practice with a good teacher are going to contribute more to expertise than hours spent without a teacher,” and that I’m a good enough teacher to make that difference.  But I fear that there is a lot of confirmation bias here—I want to believe that what I do matters, so I accept articles and studies that confirm that belief.

Looking back over my own education, I had a few teachers who helped me progress, and a few who probably delayed my learning by convincing me that the subject they were teaching was unutterably tedious, but a lot of my learning was on my own without a teacher. Sometimes the initial learning was with a teacher (often my Dad, when I was child, see Thanks, Dad), but subsequent learning was pretty much entirely from books and solo practice.  It is hard to say whether I would have achieved more expertise with teachers—some of the stuff I learned was esoteric enough that there were no teachers and I had to teach myself.  Other material was more commonly available, but I came at it from an unusual direction, so that the conventional ways of teaching the material would have been a very bad match for me.

Having an expert mentor around can make difference, and structured practice (such as I assign to my students) can make a difference—even just having an externally imposed reading schedule can make a difference.  But most of my learning in the past couple of decades has been without a teacher and without an externally imposed course structure.

So my own experience is that teachers are not the secret sauce to developing expertise.  Good teaching helps, but good learning can take place even in the absence of teachers.

Mark wrote

Look back at that definition of “deliberate practice”—who’s going to pick the activities that most address your needs or provide the immediate feedback? The definition of deliberate practice almost assumes that there’s going to be teacher in the loop.

I think Mark is wrong here.  For example, when I was teaching myself electronics design, I picked the activities based on what I wanted to design.  The feedback came from building and testing the circuits—from the real world, not from the opinions of teachers. I found that some of the simplified models used in the text books and religiously repeated in intro courses were not very useful, while others were very handy and gave good results.  Having a teacher steering me would have probably resulted in less learning, because I would not have been as invested in the examples (so less willing to explore) and the examples would have been chosen to give the conventional results, rather than showing where the conventional models break down.

For example, my post Capacitance depends on DC bias in ceramic capacitors explains how I found out about how ceramic capacitors change their capacitance with DC bias.  The knowledge was out there in various industrial application notes, but it is not generally taught in beginning electronics courses—capacitors are treated as ideal devices.  A teacher would probably have led me to a circuit that did not have a large DC bias on the capacitors, so that they would have acted much like the ideal devices, and I would not have learned a very important (and often overlooked) flaw in the models.  I may be less expert in the conventional models than someone who spent the same amount of time studying electronics with a teacher, but I have picked up odd bits of learning that I would have missed with most teachers.

Similarly, my posts Diode-connected nFET characteristics, More mess in the FET modeling lab, and Mic modeling rethought showed my learning about the characteristics of nFET transistors, where I ended up with a different model from the textbook ones.  Teachers would have almost certainly directed me to learn the conventional model first, and then much more complicated models to patch the conventional model (that’s all I could find in any of the textbooks).  Not having a teacher let me find a useful simple model for the I-vs-V curve that models the entire curve fairly well, without having to switch between models.  (Incidentally, I never did come up with an explanation for the negative resistance in the first nFETs measured in the “more mess” post—that part has been discontinued and other nFETs I’ve measured don’t exhibit the phenomenon.)

Mark might argue that I had good teachers in the past, which allowed me to develop more expertise at self-teaching.  I won’t dispute that, but I think his main point “the definition of deliberate practice almost assumes that there’s going to be teacher in the loop” is refuted by self-teaching with real-world feedback.

2014 August 14

ScratchJr

Filed under: Uncategorized — gasstationwithoutpumps @ 11:42
Tags: , , ,

If you’ve been wanting to teach your child to program, but found Scratch too complicated for your 5-year-old, there is a new option: ScratchJr.

On the  ScratchJr – About page, the developers say

What is ScratchJr?

ScratchJr is an introductory programming language that enables young children (ages 5-7) to create their own interactive stories and games. Children snap together graphical programming blocks to make characters move, jump, dance, and sing. Children can modify characters in the paint editor, add their own voices and sounds, even insert photos of themselves—then use the programming blocks to make their characters come to life.

ScratchJr was inspired by the popular Scratch programming language (http://scratch.mit.edu), used by millions of young people (ages 8 and up) around the world. In creating ScratchJr, we redesigned the interface and programming language to make them developmentally appropriate for younger children, carefully designing features to match young children’s cognitive, personal, social, and emotional development.

ScratchJr is now available as a free iPad app. We expect to release an Android version later in 2014 and a web-based version in 2015.

Unfortunately, I can’t give you a review of the program, as I don’t have an iPad to check it out on (nor do I have easy access to 5-year-olds, now that my son has grown up).  They did a great job on Scratch, though, so I would hope that ScratchJr has extended the concepts to a lower age group appropriately.

2014 May 7

Quiz corrections

Filed under: Circuits course — gasstationwithoutpumps @ 20:36
Tags: , , , , , ,

As I reported last week, students did poorly on the first quiz, which came as no surprise to me.  I had the students redo the quizzes as homework, allowing collaborative work (as long as they acknowledged the collaboration in writing).  They turned in the homework on Monday, a week after the quiz, and I returned them today.  No one aced the redo, with the top score being still only 25/33 (which would have been an A on the first pass, on a redo maybe a B+).

A lot of the students still seem to be having trouble with complex numbers—they got the formulas right when working symbolically, but then the exact same question with numbers instead of letters (which could be done by just plugging into the formulas) came out with real numbers when complex impedances were asked for.  Also, a lot of sanity checked were skipped (several people reported a battery as doubling in voltage when hooked up to a resistor, for example).

These students are not major mathphobes (they’ve all passed a couple of calculus classes and most have done more math past that), but they don’t seem to have any sense for reasoning with or about math—they just want to plug in and grind, even on simple problems like ratios in voltage dividers. This class has almost no memory work (I gave them a one-page handout at the beginning of the year with all the math and physics I was expecting them to memorize), but relies heavily on their being able to recognize how to apply those few facts.  This often requires subdividing a problem, like recognizing that a Wheatstone bridge is the difference between two voltage dividers, or that a 10× oscilloscope probe is a voltage divider with R||C circuits for each of the two impedances.

I spent the entire class today working through each problem in the quiz, to make sure that everyone in the class could understand the solution, and (more importantly) see that they did actually have enough knowledge and math skill to do the questions. Some of the students were feeling overwhelmed on the quiz, because they are not used to doing anything more than 1-step pattern matching for problems, and some of the quiz problems required two steps.  None of the quiz problems were as hard as the prelab they had to do this week, which involved 8 or more steps to get the resistor values to set the gain of the amplifier:

  1. Determine the pressure level of 60dB sound in Pa.
  2. Determine the sensitivity of the microphone in A/Pa:
    1. Convert -44dB from spec sheet to a ratio
    2. Get V/Pa sensitivity for microphone for circuit on spec sheet
    3. Convert to A/Pa given resistance of I-to-V conversion resistor on spec sheet.
  3. Determine voltages needed for op amp power supply.
  4. Determine I-to-V resistor needed to bias microphone in saturation region.
  5. Convert A/Pa sensitivity, RMS pressure level, and I-to-V resistor to RMS voltage out of microphone.
  6. Determine corner frequency and R, C values for DC-blocking filter.
  7. Determine maximum output voltage range of the amplifier as the most limiting of
    1. Voltage range of op amp outputs
    2. Power limits of loudspeaker (10W)
    3. Current limit of op amp (which is a function of the power-supply voltage) into 8Ω loudspeaker
  8. Determine max gain as ratio of RMS voltage into op amp and RMS voltage out of op amp (I’m allowing them to be a bit sloppy about RMS voltage vs amplitude, since we are not looking just at sine waves—the amplitude of a symmetric square wave is the same as the RMS voltage.)
  9. Choose resistor values to give the desired gain.

I’m hoping that pushing them go through these multi-step designs in the lab will give them more practice at decomposing problems into smaller pieces, so that two-step problems on a quiz no longer seem daunting, but routine.

I’m going to be giving them another quiz in about a week, covering op-amp basics and the amplitude response of RC filters.  I’ve got to figure out the best time to do this—possibly a week from Friday, after they’ve done another op-amp lab (using a phototransistor to make a pulse monitor, using this handout).  I think I’ll reorder the labs after that, doing the pressure sensor instrumentation amp lab, then the class D power amp, then the EKG.

 

 

2014 April 26

As expected, students did poorly on the quiz

Filed under: Circuits course — gasstationwithoutpumps @ 15:43
Tags: , ,

I gave about the same quiz as I did last year,  changing the numbers, removing one of the harder questions, and making sure that some of the other questions reflected worked examples we had done in class. The quiz was again on the 12th day of instruction. I had intended to move it to the 10th day, but one of the students was called out of town, so I rescheduled it so that everyone could take it at the same time.

I expected similar distribution to last year’s (last year the range was 3/32 to 12/32), but was hoping for slightly better.  I saw a distinct bimodal distribution this year, with half the class getting scores from 0/33 to 6/33 and the other half getting 11/33 or 12/33. This is a little clearer distribution than last year’s, which spread the students out more uniformly. I was still hoping that some of the better students would get over half the points on the quiz, but they seemed to top out at 36%.

I worked this year’s quiz myself in about 24 minutes (which means the quiz was a little too long still—I want about a 3:1 ratio on time, and the students had only 70 minutes).

I was really depressed after last year’s quiz, because I had not been expecting such dismal performance. This year I was braced for it, but still hoping for better.  Still there were some surprises:

  • There were a few questions that should have been free points (like asking for the impedance of a resistor with resistance R)—I was disappointed that some students missed even the trivial questions.
  • I had a pair of questions which were identical, except that one asked for algebraic formulas for impedance and the other gave component values and asked for numbers. I put the algebraic ones first this year, so the numeric ones were just a matter of plugging the numbers into the algebraic ones (and doing a sanity check).  The algebraic ones had a mean score of 2/4 with a standard deviation of 1.2, while the numeric ones had a mean of 1.22/4 and a standard deviation of 1.2.  I had not expected a drop in performance on the numeric ones, since the received wisdom in the physics education community is that students do better with numeric examples than algebraic ones.
  • No one got any points on the oscilloscope probe example, even though it was identical to an example we had worked in class.
  • The average score on a load-line problem was 1/6 with a standard deviation of 1.3.  This did not look like a normal distribution, but an exponential one, with half the class getting no points.
  • I had two low-pass RC filter questions. One asked for algebraic formulas; the other used the same circuit but asked for numeric answers using specific component values, voltages, and frequencies. The algebraic one was bimodal, with 2/3 of the class getting 0 and 1/3 getting the answers completely right. The numeric one was significantly worse, with only 2 out of 9 students getting any points (1/6 and 3/6).
  • I asked a couple of voltage divider questions that required applying the voltage divider formula circuits in which the voltmeter was connected between two nodes, neither of which was ground.  One asked for an algebraic results (a Wheatstone bridge), the other for a numeric result (voltage across the middle resistor of three in series. Students did very poorly on both,  with only one person getting the voltage for the middle resistor (one got half credit for setting it up right, but computing wrong), and no one getting more than 1/5 for the Wheatstone bridge.

Last year I suggested several ways to handle the poor performance on the first quiz:

  1. I could tell them to study and give them another quiz.  That would be totally useless, as it would just repeat the problems on this quiz.  They don’t know what it is that they need to know, and vague exhortations to study are pointless.  I don’t think the problem is lack of effort on their part, and that’s the only problem for which pep talks are a potential solution.
  2. I could go over the quiz question by question, explaining how I expected students to solve them.  This is classic lecture mode and the approach I used to use. It would be easy to do, but I doubt that it would help much.  I already did an interactive lecture on the material, and another approach is now needed.
  3. The students could get the quiz back and be told to go home and look up in their notes and on-line anything they did not get right.  They would find and write down the right answers, as if this were homework.  (This “quiz correction” is a standard strategy in high school teaching, but not common in college teaching.)  One difficulty here is that they might be able to find answers (say by copying from other students in the class) without understanding how to do the problems.  It is probably a better approach than yet another lecture, but I’m not sure it will work well enough.  If the students were trying to get from 80% understanding to 95%, it might be fine, but to get from 30% to 80%, something more directed is needed.  More time and open notes would help, but maybe not enough.
  4. I could break them into groups and give each group a couple of the problems to work on together in class. This peer instruction technique would be a good one if about 1/2 the students were getting the problems right, but with the top of the class getting only 1/3 right, I may need to give them more guidance than just setting them loose.  For example, on some of the problems there was a fundamental misreading of the circuit schematics that was very common. I could clear up that misunderstanding in a minute or so and have them rework the problems that depended on it.  Then I could send them home to write correct solutions.
  5. I could give out lots of problem sets to drill them on the material.  Of course, since it took me more than all day Sunday to make an 8-question quiz, it would take me forever to generate enough drill problems to be of any use.

I feel the same way this year about the possible teaching strategies, but this year I’m going to try a mix of methods 3 and 4, asking them to redo the quizzes at home, working with others until they are satisfied that they can now do the problems and other similar problems when asked.  I’ll have them hand it in this year as a homework, but not go over it in class until after they turn it in.  They need to take a more active role in trying to master the material, and not rely so much on my telling them what to do.

Monday we’ll cover inductors and loudspeakers, in preparation for the Tuesday measurement lab.

On Wednesday I was planning to do gnuplot analysis of the loudspeaker data, but I think I’ll keep that fairly short, so that we can get an intro to sampling and aliasing also before Thursday’s lab.  I have to decide whether to bring in my son’s stroboscope and a moving object to demonstrate aliasing.

Friday, I’ll introduce op amps, with the intent of developing the block diagram in class on Monday for a simple op amp microphone circuit for the Tuesday lab.  This weekend I need to rewrite that lab from last year—I decided last year to use the dual power supply with a center ground for their first op-amp design, rather than having them build a virtual ground (we’ll get that in the next lab assignment).

 

 

2014 April 7

Feedback on first lab report

Filed under: Circuits course,Printed Circuit Boards — gasstationwithoutpumps @ 17:11
Tags: , , , , ,

Most of today’s class was taken up with feedback on the design reports that students turned in by e-mail on Saturday.  Overall the reports were not bad (better than the first reports last year), but I think that the students could do better.  Here are the main points:

  • Anyone can redo the report to get it re-evaluated (and probably get a higher grade).
  • No one attempted theV1 &V2 problem, so I reassigned it for Wednesday.

    The circuit I had given as an exercise, asking them to determine the output voltage V_out.

    The circuit I had given as an exercise, asking them to determine the output voltage V_out.

  • A lot of reports mixed together two different problems:the 1kΩ–3.3kΩ problem and the optimization to maximize sensitivity of the thermistor temperature sensor.  I encouraged students to use more section headers and avoid mixing different problems together.
  • Figures should be numbered and have paragraph-long captions below each figure.  I reminded students that most engineering reports are not read in detail—readers flip through looking at the pictures and reading the picture captions. If the pictures and captions don’t have most of the content, then most readers will miss it. I also pointed out that many faculty, when creating new journal articles, don’t ask for an outline, but ask for the figures.  Once the figures tell the right story, the rest of the writing is fairly straightforward.
  • A lot of the students misused  “would” in their writing, treating as some formal form of “to be”. The main use in technical writing is for contrary-to-fact statements: “the temperature would go down, if dissipating power cooled things instead of heating them”.  Whenever I see “would” in technical writing, I want to know why whatever is being talked about didn’t happen.
  • A number of the students had the correct answer for the optimization problem, but had not set up or explained the optimization. Right answers are not enough—there must be a rational justification for them. In some cases, the math was incomprehensible, with things that weren’t even well-formed equations. I suspect that in many cases, the students had copied down the answer without really understanding how it was derived and without copying down the intermediate steps in their lab notebooks, so they could not redo the derivation for the report.
  • A number of the plots showed incomplete understanding of gnuplot: improperly labeled axes, improperly scaled axes, plots that only included data and not the models that the data was supposed to match, and so forth. I pointed out the importance of sanity checks—there was no way that anyone ran their recording for 1E10 seconds! I was particularly bothered that no one had plotted the theoretical temperature vs. voltage calibration based on the parameters from their temperature vs. resistance measurements, so I could not tell whether the voltage divider was doing what they expected it to.
  • No one really got the solution for the 1kΩ–3.3kΩ problem perfectly. A number of them set up the equations right and solved for R (getting 2.538kΩ), but then not figuring out what Vin had to be.  It turns out that Vin depends strongly on R, so rounding R to 2.2kΩ or 2.7kΩ results in different good values for Vin, and the 2.2kΩ choice gives a more desirable voltage (around 3.3v, which we have available from the KL25Z boards, as it is a standard power-supply voltage).

I also showed the students how I had expected them to setup and explain the optimization to maximize sensitivity at a particular operating temperature.

After that feedback, I started on new material, getting the explanation of amplitude, peak-to-peak, and RMS voltage. I think that the RMS voltage explanation was a bit rough.  I was deriving it from the explanation that we wanted a measurement that represented the same power dissipation in a resistor as the DC voltage, and I got everything set up with the appropriate integrals, but I forgot the trig identity (cos(\omega t))^2 =\frac{1}{2} (1-cos(2 \omega t)), and ran out of time before I could get it right.  I did suggest that they look up the trig identity and finish the integration.

I had hoped to get at least partway into Euler’s formula, complex sinusoids, and phasors, but the feedback took longer than I had expected. Those topics will have to wait until Wednesday or even Friday, since Wednesday we’ll want to do the modeling of the DC characteristics of the electret mic, and talk about how the mic works.

 

Next Page »

The Rubric Theme. Create a free website or blog at WordPress.com.

Follow

Get every new post delivered to your Inbox.

Join 274 other followers

%d bloggers like this: