WeBWorK Main Forum

Private WH Freeman problems for Rogawski 2E ET

Private WH Freeman problems for Rogawski 2E ET

by Louis Zulli -
Number of replies: 10
"W. H. Freeman offers a comprehensive WeBWorK problem library for the Second Edition of Calculus by Jon Rogawski.  Coded by experts, quality assured, and with complete solutions, the Rogawski Calculus library provides instructors with a large pool of questions for homework, practice, and assessment." ----from http://www.whfreeman.com/catalog/static/whf/rogawski2epreview/media_webwork.html

I have been very disappointed with the general quality of the private WeBWorK problems provided by W. H. Freeman to adopters of Rogawski's text. In just the first week of our semester, we've already found about ten problems with significant errors. In addition, other problems are badly formulated. For example,

Find the volume of the solid obtained by rotating the region under the graph of the function f(x)=x^2-4 x about the x-axis over the interval [0,4].

I want members of this forum to be aware of this situation.
In reply to Louis Zulli

Re: Private WH Freeman problems for Rogawski 2E ET

by Louis Zulli -
I'm patching problems at a rate of about two per day.

My suggestion: Avoid this collection of problems until W. H. Freeman completely reviews/revises it.

What a disappointment.
In reply to Louis Zulli

Re: Private WH Freeman problems for Rogawski 2E ET

by Alex Jordan -
Oh no!

I was one of the people who coded those problems. I believe that they contracted a bunch of members from this forum to do as many problems as they could do. I hope I'm not responsible for too many of the mistakes you are encountering.

They have a system in place for error reporting. If you notify them when you find errors, the original author gets a message requesting a correction. Since I submitted 189 problems to them over a year and a half ago, I've had five problems sent back to me for correction. I'd say either I have a fair record or the errors aren't being reported. The latter is likely, since two of those five errors were reported to me just last week. It's a huge task, but this collection could become the most vetted collection out there if users continue to report errors to them. Anyway, I assume that even the two problems I just corrected last week will be corrected in their database immediately. Is that how they are making the problems available - through a dynamically changing database? Or have they sent you hard copies that you are stuck with until you request updated copies?

Regarding the wording of problems, are you sure that the textbook problem is not worded in the exact same way? I tried not to take any liberties in rewording any problems I had been assigned. If they had assigned me problem 7.55, say, then I tried to word the problem exactly as the textbook author had. All of these problems are supposed to be randomized copies of problems from the printed version.

Alex


In reply to Alex Jordan

Re: Private WH Freeman problems for Rogawski 2E ET

by Louis Zulli -
Hi Alex,

I asked about the "system" for reporting errors. I was told that I should email/contact the Associate Media Editor for Mathematics and Statistics who sent us the collection. So that's what I've done. Repeatedly.

If there's a "dynamically-changing database" then that's news to me. I receive a return email with an apology and a corrected .pg file attached. In some cases, the corrected file appeared to be nothing more than the patched file I sent when I reported the error---and my patches are not always the "best" way to fix the error, just quick.

In any case, we can't wait days for fixes to buggy problems. Students and instructors lose faith in the entire system.

About the wording of the problem that I included in my first post. The analog in the textbook seems to be: "Find the volume of revolution about the x-axis for the given function and interval." Still not as good as it should be, but at least it doesn't refer to an unbounded region.

Calling these problems "quality assured" is, in my opinion, false advertising. I've spent precious hours responding to puzzled students, fixing bad problems and trying to reassure colleagues that WeBWorK and Rogawski's text were good choices for our department.

Thanks for your reply.

Louis
In reply to Louis Zulli

Re: Private WH Freeman problems for Rogawski 2E ET

by Alex Jordan -
Hi Louis,

Yep, I agree that the level of quality assurance is inadequate. Not that it matters, but I think this probably is the fault of whomever they contracted to do that. I was asked to be one of the problem testers, but I wasn't available at that time. My point is that they did hire people to test the problems. And yet one of my own problems that I corrected last week had an error in it that clearly illustrates no one ever tried to enter the correct answer. The answer was an expression in t; the displayed correct answer was an expression in t; yet entering an expression in t yielded an error message that t was not defined in this context, because I had re-Computed the answer in a new context in the ANS command.

I'm not affiliated with WHFreeman other than a one-time gig coding problems for them. I still fix ones they send back to me out of personal pride; I'm not sure that 1.5 years later if I am obligated to. We still use Stewart for calculus here. So I'm not familiar with the end-user experience of these problems. It's disappointing that they aren't using a database of problems like the NPL, and that there isn't a more streamlined bug submission system.

Alex
In reply to Louis Zulli

Re: Private WH Freeman problems for Rogawski 2E ET

by Louis Zulli -
I and many of my colleagues are giving up on this collection of problems. In my opinion, it is not close to publication quality. The final straw was this problem. Take a look at the "correct" answer I entered. Sure I can fix the problem, but ...

final_straw.png

In reply to Louis Zulli

Re: Private WH Freeman problems for Rogawski 2E ET

by Arnold Pizer -
Hi,

I guess different instructors (and authors) would allow different answers.  You probably want 1-11/e^10 which is approximately .9995.  WeBWorK usually requires  answers to be correct to .1% so 1 is a correct answer and x/x is equal to 1 pretty much everywhere (and so WeBWorK treats x/x the same as 1). Would you accept  x+1-11/e^10-x as an answer? I certainly would.  WeBWorK by default pretty much takes the view that any correct answer is correct. If you want to restrict what the student can get credit for, you have to do that explicitly and the author of this problem (consciously or unconsciously) chose not to.

Arnie


In reply to Arnold Pizer

Re: Private WH Freeman problems for Rogawski 2E ET

by Louis Zulli -
What is "x"? Should we accept "y/y"? Or "garbage/garbage"? It's a definite integral.

I like WeBWorK very much, and am quite aware of what it does by default. I know how to fix this problem. But with all due respect, the author of this problem is pretty much clueless---about WeBWorK, about TeX and especially about the process of reviewing a problem for quality. I've seen the .pg file.

"Coded by experts, quality assured" --- what a crock.

Louis
In reply to Louis Zulli

Re: Private WH Freeman problems for Rogawski 2E ET

by Davide Cervone -
I have to say I'm with Louis on this; I would not want to see students enter answers that include the variable x, since I would expect a definite integral to be a number. Had the problem been coded with an answer that is a Real(), the student would have gotten a message about his answer not being a number (but looking like a formula), which I think would have been a reasonable response. Instead, it appears that the problem is coded with the answer as a Formula() object instead, and I do consider that to be a poor choice. Perhaps they were trying to force the correct answer to look a certain way, but that is better handled via Compute(). Or perhaps they are not using MathObjects at all; I haven't looked at the problem's source code.

Davide
In reply to Davide Cervone

Re: Private WH Freeman problems for Rogawski 2E ET

by Louis Zulli -
Hi Davide,

FYI:

ANS(fun_cmp("(1)-($a1/(e^($a)))" , 'var'=>['x','C']));

Louis

PS---I recently added a button in the Library Browser to give my colleagues quick access to the Union College library. If only the W.H. Freeman problems were as well constructed as those from Union. smile
In reply to Louis Zulli

Re: Private WH Freeman problems for Rogawski 2E ET

by Davide Cervone -
Yes, that is what I thought: they have forced a formula where a number is the correct answer. This was a conscious choice of the problem author, and they have gone to pains to include the variables x and C, presumably to avoid the warning messages about entering a formula when a number is expected. So they have intentionally disabled the type warnings, and allowed the kinds of unusual answers like you have given to be marked correct.

Not only that, they have set things up so that wrong warning messages will now be issued. For example, if you typed 1,2 into the answer, you will get the message "Your answer isn't a formula that returns a number (it looks like a list of numbers)", which incorrectly suggests to the students that they should be tying a formula rather than a number.

Such efforts to bypass the standard warnings are ill-advised and usually cause worse problems than the ones they were trying to avoid. It is rarely a good idea to intentionally use the wrong type of answer from what the students are supposed to type.

Had they used MathObject directly, they could have disabled the type warnings with a flag, and left the answer a real number, so that a formula like your x/x would have been marked incorrect, but without any warnings about the wrong type of answer.

I agree with you, this is poor coding. Unfortunately, it is not restricted to the Rogawski problems.

Davide

PS, thanks for the kind words about the Union problems.