Midterm 3 results

I was e-mailed the mean and the results for each answer for the final midterm but not the histogram. I will get mailed a copy of that but I thought some of you might want to see the rest of the information right away.

The mean was 82% (vs. 81% last year)

I have looked at the answer distribution for all the questions to check for problems and everything looks fine.

Lowest % correct was 57% for Q12 (self-thinning question) and 58% for Q 27 (cutting a shoot question).

Highest % correct was 98% (one question: the fresh fruit question - I'm glad this course has some practical uses!).
I have looked closely at all the questions where either less than two thirds of you got the correct answer (three questions) or more than 20% of you picked a particular incorrect answer (four questions) and don't see any problems.

This time last year: Midterm 3 etc

Midterm 2 results

The second midterm has been graded and your GSI's will give you your individual results next time you meet with them. Please don't e-mail me - I don't have the individual results, just the overall summary.

Overall there were 248 students for the second midterm, the mean was 86% (vs 80% for the first midterm) and the median was 89% (vs 84% for the first midterm).

Overall 50% of the class scored in the A range (>90%), 28% in the B range (>80%), 13% in the C range, 5% in the D range and 4% in the F range.

These results are slightly higher than last year(see link below for last years results). Like the first midterm twelve people obtained perfect scores - although I have no idea if they are the same twelve people.

As usual I have looked at the answer distribution for all the questions to check for problems.

Lowest % correct was 49% for Q23 (demographic transition question) and 54% for Q 34 (density dependent mortality in the beetle population).

Highest % correct was 98% (two questions).
I have looked closely at all the questions where either less than two thirds of you got the correct answer (just the two questions mentioned above) or more than 20% of you picked a particular incorrect answer (three questions) and don't see any problems.

Section means were similar and, again, no section was more than 4% away from the overall mean of 86%.

FYI the mean and median for the final midterm are usually closer to the first midterm (ie a mean of around 80-82%) than the second.

This time last year: Midterm 2 Results and Lichens in Space

Midterm 1 Results - 2007

The first midterm has been graded and your GSI's will give you your individual results next time you meet with them. Please don't e-mail me - I don't have the individual results, just the overall summary.

However some of you may be interested in this, especially when you have your own grade.

Overall there were 259 students, the mean was 80% and the median was 84%. This difference is typical for skewed distributions and is why the median is a better descriptor. If you got 84% (the median) then half the class scored more than you and half the class scored less.

Overall 37% of the class scored in the A range (>90%), 25% in the B range (>80%), 18% in the C range and 10% each in the D and F range.

These results are similar to previous years (see link below for last years results) with one small exception - a much higher number of students than usual got a perfect score. This year an impressive 12 students got 100% on this midterm. Well done.

As usual I have looked at the answer distribution for all the questions to check for problems.

Lowest % correct was 57% for Q27 (polydactyly in cats - despite the word DOMINANT being in bold capitals, a significant number of people provided the allele frequency that would have been correct if it was a recessive trait) and 50% for Q 37 (the calculation on the seed discarding plant breeder that was meant to be a fairly difficult question).

Highest % correct was 98% (one question).
I have looked closely at all the questions where either less than two thirds of you got the correct answer (6 questions) or more than 20% of you picked a particular incorrect answer (4 questions) and don't see any problems.

Section means were similar and no section was more than 4% away from the overall mean of 80%.

This time last year: Midterm 1 Results and Web Sight

Midterm 3 etc.

Here are the results of the third midterm. Please don't e-mail me for your individual scores, I don't have them!

The overall mean was 81% and the median was 82%.

Nobody managed to get all fifty questions correct this time but four students, in three different sections, got all but one question correct.

Section means were very similar and, yet again, no section mean was more then 3% away from the overall mean of 82%

The histogram of scores is above and the breakdown by grade is below.

A (ie >=90%) - 31%
B - 37%
C - 18%
D - 8%
F - 6%

I have looked closely at all the questions where either less than two thirds of you got the correct answer (I only have this data by section, but it looks like maybe 4 questions) or more than 20% of you picked a particular incorrect answer (again, maybe 4 questions) and don't see any serious problems.

Well it's been fun. I hope some of you have enjoyed reading the blog. Google's blogger is certainly convenient but I have to say it has the most bizarre spell checker. It's not surprising it doesn't have technical terms in its database but it tries to replace them with amusingly inappropriate terms. Here's a couple I particularly enjoyed:

one of a whole family of defoliants used during the Vietnam war.
is much improved as:
one of a whole family of diplomats used during the Vietnam war.

and humongous fungus.
is alliterative but, again, is much improved as
unionized fungus

John

Campbell 7th

It was brought to my attention that there is a fairly serious error in some copies of the Campbell 7th edition. It doesn't appear to be present in the 6th edition.

In figure 38.4 (The development of Angiosperm Gametophytes), earlier copies of the 7th edition have the haploid and diploid labels reversed. It SHOULD be labeled so that the microsporocyte, megasporangium, megasporocyte and integuments are DIPLOID, and all the other structures are HAPLOID (except the micropyle, which is a gap and so is neither haploid nor diploid!)

As an aside, it can be tricky in the US to tell what printing a book is. However for many books there is a simple code on the rear of the title page that gives you the necessary info. In Campbell you'll find this on page ii just above the Benjamin Cummings logo. You are looking for a string of numbers like:
1 2 3 4 5 6 7 8 9 10 -VHC - 08 07 06 05 04
By taking the lowest numbers you get the printing and year (for some odd reason these numbers are not always in order). So the book I'm looking at was the first printing in 2004. Another copy I looked at in the Bio1B office did not have this mistake and was labeled:
4 5 6 7 8 9 10 - VHC - 08 07 06 05
or the 4th printing from 2005.

So now you know. I suppose if you wanted to get maximum value for your dollar in a used textbook store you would look for later printings hoping that mistakes had been corrected.

As requested

I do not make my powerpoint slides available to you for one very simple reason - I have not obtained permission to use all the images. To try to make the lectures interesting I pull up all sorts of images from the web and numerous other sources. As long as they don't explicitly request that the image not be copied I don't worry too much for a single in lecture use. However I would not feel comfortable making all these images available to the whole class by putting the powerpoint slides online.

The legally correct solution would be to just use images I have taken myself (the only pictures I take are of Scooter, my dog) or to use those from Campbell (which would make for a fairly dull lecture since you've seen all these images) or to get permission to use all the images I want to use (which would take forever).

So instead I use what I want but don't make the slides available. That is why I have copied much of the text on the slides in the form of the lecture outlines and indicate, wherever possible, equivalent figures in Campbell.

But someone did point out that I could go one step further and make available all the text slides I use and those very few images I make myself. I'm not sure it provides much more than you already have but here are the text slides from the Plant Section powerpoints. (There is also one other image snuck in there that someone requested so I hope you are reading this).

Final Jeopardy

When I was younger I was fascinated about how other people's minds worked. I was intensely curious about whether other people had a mind full of junk all the time and whether they were constantly making the same sort of apparently random links and jumps.

I was recently reassured to see this animated cartoon by Scott Bateman which illustrates what happened inside Bob Harris's head when faced with a Final Jeopardy question. (Bob Harris is an American radio commentator, writer, stand-up comedian, Jeopardy contestant and blogger.)

I'm always reminded of this when I try to write exams. It is so difficult to try and step back and see questions as others might see them. I've also been talking with Mike Moser and some others about ways to help people who learn the material, feel they understand it, and yet are disappointed by their midterm results. (There was an earlier blog posting on this if you missed it, and a list of advice is on the Bio1b website.)

Speaking with a couple of separate people this week I realized I need to add a section here on time management. I'll illustrate this for a Bio1B midterm but the principal applies to lots of other exams.

Let's assume that you find 35 of the 50 questions straightforward, they only take you a minute or so, let's say 40 minutes total.

There are another 10 questions that you find harder, maybe they are problems, or take extra reading or calculation. You take two minutes over each of these for a total of 20 minutes.

You now have 20 minutes left for the last 5 questions which you are really puzzling over. It is pretty easy to spend most of this time puzzling over these questions and only leave a few minutes for a cursory check of your answers to the earlier questions. You become obsessed by the harder questions at the expense of checking for errors on the easier questions.

However, if you are anything like me, you have probably made at least half a dozen mistakes on those 35 'easy' questions because you misread the question, filled in the wrong entry on the scantron, failed to notice a negative qualifier etc. etc. Fifteen minutes spent checking your answers will probably bring you much better returns than agonizing over the questions you find hardest.

A better strategy, in this example, would be to allot about 1 minute per question to work through the entire 50 questions. Some questions may take less time but try not to go beyond 1 minute on any question. Cross out answers you know to be wrong as you go along. If you run out of time then put a question mark by your best guess and move on. You should finish this in under 50 minutes. Now split the remaining time in two parts. Spend half the time (15 mins) checking your answers. Make sure you read the question correctly. Look for negatives and qualifiers. In doing this you can take another look at the questions you didn't complete the first time. Something new might come to you. Finally, divide the last 15 minutes between the questions that remain.

Midterm 2 results

Here are the results of the second midterm. Again, I can't post your individual results but here is an analysis of the complete data set.

The overall mean was 84% and the median was 86%.

The highest score was 100%, and, again, two students in different sections managed to achieve this.

Section means were very similar and no section mean was more then 3% away from the overall mean of 82%

The histogram of scores is above and the breakdown by grade is below.

A (ie >=90%) - 38%
B - 37%
C - 17%
D - 6%
F - 2%

There are two issues with individual questions. These changes will be made by your gsi and will not show up on your scantron reports.

Due to the conflicting information presented in Campbell 7th I will accept A or B as answers to question 27 (the island biogeography question).

There was an error on the answer key (my fault, it was due to a last minute change in a question). The correct answer for question 15 (inefficiency of food chain transfer) was B (productivity). Answer C (endotherm/ectotherm) is incorrect. For 152 of you this means your score will go up. My apologies for this mistake.

I have looked closely at all the rest of the questions where either less than two thirds of you got the correct answer (4 questions) or more than 20% of you picked a particular incorrect answer (5 questions) and don't see any other problems.

John

Midterm 1 results

I thought some of you may be interested in the results of the first midterm. I obviously can't post your invidual results, and, in fact, don't have them, but I do have an analysis of the complete data set.

The overall mean was 80% and the median was 83% (for skewed distributions like this the median is a better description of the 'typical' value than the mean.)

Highest score was 100%, and two students in different sections managed to achieve this - well done.

Section means were very similar and no section mean was more then 3% away from the overall mean of 80%

The histogram of scores is above and the breakdown by grade is below.

A (ie >=90%) - 31%
B - 31%
C - 17%
D - 12%
F - 9%

Looking at the individual questions I see no problems. Lowest % correct was 53% for Q6 (a calculation), Q37 (a calculation) and Q43 ('what evolved before what' question). These were all meant to be hard questions.
Highest % correct was 97% (two questions).
I have looked closely at all the questions where either less than two thirds of you got the correct answer (5 questions) or more than 20% of you picked a particular incorrect answer (8 questions) and don't see any problems.

John

Calculations

One of the advantages of having a multiple choice exam is that I get abundant feedback on how many people got every question right, what proportion picked each answer etc. I was looking through the results of last years exam as preparation for writing this years and although I had always realized that people found genetic problems difficult I didn't realize the magnitude of the effect until I decided to quantify it.

If I crudely divide the questions up into three categories and calculate the average percentage of people that got questions in that category correct I get something like:

Genetic problems requiring calculations - 54%
Genetic problems requiring just logic or punnet squares etc - 73%
Everything else - 82%

Another way of looking at this is that although only perhaps ten questions fall into the first two categories the average person will lose 1/3 of the total points they lose on these ten questions alone. Hmmm.

Exam Tips and Tricks

Mike Moser and I were talking about what we could do to help some of the people who are clearly interested in the subject matter of Bio1B and work hard, yet are dissapointed by their midterm grades. Others seem to get good grades with less effort. Some people, it seems, are just better at exams than others. So to help level the playing field I thought I'd try to compile a list of exam advice, tips, tricks, whatever you want to call them. I have focussed particularly on multiple choice exams because that's what we have in Bio1B but many of the principles apply to other types of exam.

Brett has posted the list I came up with to the Bio1B website here.

To most of you, I hope, most of these are common sense. My reason for posting here is to request your input. Any good tips and advice I'm missing? You can e-mail me direct or post a comment here. (Bio1B is NOT graded on a curve so you are not decreasing your own fitness by helping your fellow students!)