Saturday, August 1, 2009

Reaching a different conclusion

Teaching is a learning experience, done properly. I've been doing something one way for a while, with what I think is good cause, but more information has come forward recently and I think I have to change my ways for the good of the students.


College is expensive, and community college, while it is supposed to be the most reasonably priced option, keeps jerking the students around with steep tuition hikes, especially when we have an idiot Republican governor who thinks any tax on anyone is bad while any fee increase is good. (More on that in the next few days.)

At Laney, I've been teaching statistics without a required text for some time now. I do require my students buy a minimal calculator, and recommend Texas Instruments models because I have the most experience with them, so I can show students how to use them. While the TI-83 or 84 are the best choices for students taking stats, the TI-30XIIS is the most reasonably priced model that does what is needed. The difference in price is significant, $100 to $120 for the more expensive compared to $15 to $20 for the cheaper. There are things the more expensive model does that the cheaper does not do, of course, but that also works the other way around in a few cases.

My feeling is that I should switch things around and have a recommended text. There are a lot of students who really like having more examples than I give in class or on homework or the practice problems I put on my class blog, and it makes sense to accommodate them.


I also think I should get rid of the required calculator option and switch over to teaching the students how to use a spreadsheet, specifically Microsoft Excel.

As regular readers already know, I'm an Old School Programmer who worked professionally on computers in the 1980s, so I can lose standing in the Old Programmers Club if I do not hate Microsoft, reject all their works and mock their products mercilessly.

I linked to my posts mocking Microsoft now because I have to admit, as of 2009, Excel looks like a very good option to teach students.

Lemme 'splain.

Years ago, I had students turn in statistics homework with wrong answers, only to be told by the students that the answers they got were what Excel told them. The concept being used was simple, the median. For those who don't recall, you find the median by putting the data in order from highest to lowest, count the number of entries and find the middle one. If there are an even number of entries, the middle one is the average of the two middle values. For example, if there are ten things on an ordered list, the median is the average of thing5 and thing6.

Somehow, Microsoft spreadsheet programmers got it into their heads that median, and the related concept of quartiles, were much more complex things, and gave answers I am not sure how they computed. I didn't have Excel on a computer at home, and this first experience convinced me that Microsoft continued to write sucky software in much the same way they did when I was in the business.

But I now do have Excel on my computer at home, and it works fine on median and quartile, and actually helps a person who has some knowledge of how spreadsheets work while that person is typing. More than that, I am not forcing students to buy the Microsoft product, because they can work on computers on campus that have Excel installed, and there are labs available when I give a test where students can use the software for free.

More than the reasonable price, teaching students to use Excel more effectively is a very useful skill in the marketplace, in nearly any professional setting. For people who will go on to do more serious statistical analysis, standalone packages like MINITAB and SPSS are the industry standard, but for stats used in presentation in most businesses, Excel is the most used and useful option.

There are some in the Laney math department who are resistant to using technology of any kind in the teaching of statistics, even calculators. While I don't get paid to do so, I consider part of my job now to change their minds, even those with tenure.

Actually, especially those with tenure.

Teaching is a learning experience, done properly.

10 comments:

Margaret Benbow said...

Matty Boy, I respect that you've modified your ideas about Microsoft and are honest enough to say so. But (sigh), your posts when you were "hating on" Microsoft and Gates had a truly sumptuous flair. I sort of miss them.

Matty Boy said...

Don't worry, Margaret. Hatin' on Microsoft is an old habit, and old habits die hard. If I run into trouble with Excel, I'm sure I'll be back to whining about it.

And products like Bing still suck eggs. Big time.

47th Problem of Euclid said...

Matt: I would have taken you for a HP (RPN) over TI (Algebraic) calculator person.

Matty Boy said...

No, I've been using TI ever since a friend gave me an 89 as a present.

Are the new HP models still RPN?

Karlacita! said...

Yay Professore! I learned a smattering of Excel, and I love that program. And it is necessary in the job market, so kudos to you for helping your students get ahead.

Now when they ask When Will I Use This?, you'll have Less 'Splainin' to Do!

bobmando said...

On my Desktop PC I still use Office 97 and therefore Excel 97.... It's so "last century" and going on 13 years old (copyright 96). I have found NO compelling reason to upgrade and yet I wonder... Does this version do the incorrect median calculation? Is there a quick test to tell? and Thanks... I'm an HP RPN guy but finally gave it up when my HP died (HP 25C) Great to Google image and bring back fond memories...

Matty Boy said...

Hey, Bob. It probably does get it wrong. Here's a quick test.

Input the following data set in a1 through a10:

0
1
2
3
4
5
6
7
8
9

put in the following formulas in cells a11, a12 and a13.

=quartile(a1:a10,1)
=quartile(a1:a10,2)
=quartile(a1:a10,3)

The answers should be 2, 4.5 and 7.

if quartile isn't available on older versions of Excel, you can do these instead.

=median(a1:a5)
=median(a1:a10)
=median(a6:a10)

I don't know if Excel always got median wrong for all data sets or if they had to be fancier than this, but it definitely screwed up the median from time to time.

bobmando said...

Thanks Matty Boy... no quartile on the 97XL... and the median answers are correct. So, it must be more complicated numbers that mess up XL. I guess, for now, there is no need to upgrade! YES!

sfmike said...

I hate Microsoft but it's good to have a basic knowledge of their products simply because they are so ubiquitous.

@bobmando: Microsoft Office programs 1997 through 2003 are essentially the same except for minor deviations and upgrades. Though none of the individual programs work together very well, even though they are supposed to, Excel and PowerPoint are not badly designed and can be quite powerful.

The newest version of Office is something of a disaster, just like their Vista operating system, and I wouldn't recommend upgrading right now because it really doesn't give you any new capabilities and half the people you know won't be able to read the files. When somebody's sending you a file created in the newest version, have them save the file in a "97-03" downgradable format or you won't be able to use them.

bobmando said...

Thans @sfmike. Matty Boy, I just tried Google Docs spreadsheet and got the correct answers 2, 4.5,7 with median but got two "wrong" answers with quartile: 2.25 4.5 6.75. IF this works, Google may be the long term answer (cloud computing) to replace ms office.