Friday, December 23, 2011

Elementary Statistics

Elementary Statistics

By Bobby Neal Winters

The Learning Phase

Mathematics and Statistics are separate disciplines. This is something that both mathematicians and statisticians are insistent upon. Statisticians use mathematical tools much in the same way physicists and chemists do, and many mathematicians take statistic for the same reasons they take physics: as an application of their mathematics and to make themselves more marketable when it comes time to look for a job.
My undergraduate degree included a course in Probability and Statistics, but it was very theoretical in nature without much indication of how one might actually apply this in the real world. This is common among undergraduate math degrees. Also commonly offered is a class in elementary statistics that is rather short on theory but long on practical applications.
Fifteen or twenty years ago, as a part of my responsibilities for teaching service courses in the math department, I was assigned elementary statistics. It was a course I’d never had. The first time I taught the course, being a team player at heart, I used the text that had been adopted by the department and, as I taught, staying a few sections ahead of the students, following the syllabus that had been determined by the department.
It was in this way I learned statistics. As I learned it from the book the department was using, that book must have done something correctly, but--and you saw this coming--I hated the book.
You may be familiar with the French post-Impressionist painter Georges Pierre Seurat. If you are, then you probably know more about him than I do. What I know is that he painted pictures using dots. If you stand with your eye just an inch or two away all you see are dots; if you stand back there is a park or dogs playing poker or something else.
This text was written like that. When I was done teaching the course, I took a step back and saw the whole picture at once, but I was not convinced that any of the students would do that. For them, it was all dots. Or, if I my phrase it another way, it was just one damn thing after another without any connection between them. Nevertheless, being a team player, I continued in this way.

The Synthesizing Phase

In thinking about the course, it occurred to me there were four parts: descriptive statistics, basic probability theory, applications of the Central Limit Theorem, and advanced applications.
The descriptive statistics consisted of the topics that one, for the most part, could cover in a high school or even middle school class: Make a frequency distribution; list numbers from largest to smallest; find the average; find the standard deviation; draw a bar graph; draw a pie chart. Students, even the ones who were ultimately going to fail the class, rocked on this. Among those who showed up for the exam, it was rare to see anything less that a B. This made the second part of the class even more tragic.
Students find what I refer to as basic probability theory to be difficult. Topics in this section can be anything from if you have a can that has 300 red beads and 700 blue beads, then what is the probability of drawing a single blue bead (Yes, it’s 0.7.) to the hypergeometric distribution. This section cleaned the students clocks, which is to say they uniformly found it to be quite difficult.
The portion of the course I describe as “applications of the Central Limit Theorem” is actually the meat of the course. This is the part of the course that our clients in the university, i.e. those who use statistics in their classes, want their students to know. This includes estimation and hypothesis testing.
The final portion of the course consisted of the sample proportion, which is used to estimate the percentage of the vote a politician will get, and the chi-square test, where the question of whether particular models fit is discussed.

Teaching as Literature

After I’d taught the course a number of times, it was clear that it all did fit together like one of Seurat’s paintings, but that the students might be better served if it were taught like a classic work of literature. By this I mean, I thought I should take advantage of certain opportunities to foreshadow concepts.
For example, there is something called Chebyshev’s Rule. It is the most difficult topic covered in the first section. To use layman’s terms, it says that only a small proportion of the data can be very far away from the average. This opens up the opportunity to mention proportion, which has already come up with relative frequency, and foreshadows the central means of hypothesis testing. In terms of literature, it helps to tie the course together.
Spending more time on Chebyshev’s Rule also serves to make that first section a bit more challenging. If it is de-emphasized, as I tried a few times, there is little to give the students that they are now in a college course because so much of the material in that first part of the course is at a middle school level. Chebyshelv’s Rule, by way of contrast, can be challenging in a number of ways.
There are other opportunities to create ties they run the length of the course. I use the jar of colored beads as a recurring example. It’s used when I introduce probability, when I talk about Bernoulli Trials, and when I talk about the sample mean. By this means, and others, the student has a better opportunity to see that the course is unified and that it’s not just one damn thing after another.

Embedding in Time

Shortly after I’d separated the Elementary Statistics course into four parts, I took a year of sabbatical a Brigham Young University. They had a testing center that was very popular with the students. Teachers left copies of the exam with the testing center for an period of time--I think as much as a week--and during that interval of time, students could come and take the exam when they felt as if they were ready. Staff at the center proctored the exam, so pressure to be a policeman was removed from the professor. As I said, the students loved it and I grew to love it, but there was a catch. In order to use it, you had to schedule your exams ahead of time.
Up until that point in my career, I had not. I had exams when I felt we’d covered enough material, and the idea that this could be a predictable thing was foreign to me. In order to use the testing center, I scheduled my exams and discovered that this was not such a difficult thing to do.
When I can back from sabbatical, I scheduled my exams for Elementary Statistics and discovered a number of things. One of them was that the students didn’t complain. Indeed, while my students have never said so directly to me, I’ve read studies that support the idea that students like structure.
More important to me, however, was the discovery of how easy it made everything. Instead of having to weigh the decision of whether to have a test on a particular day, there it was on my calendar! I could make out the damn test and not procrastinate. Once those dates are set, they serve a similar function for the course as the arbor does for a grape vine: the course just grows around them.
Quite frankly, in some very real sense, when the dates are set for the exams and other assignments there is nothing much else for me to do.

Assessing and Modifying

As you’ve seen, course design is a dynamic process. I’ve modified my course as I’ve learned new things. I have changed as I’ve observed that worked and what didn’t. I was doing course design before I new the name for it. I was doing assessment before I heard the word.
That having been said, those words and phrases have power. When we teach a course, there are certain things that we want students to learn or we wouldn’t be teaching the course. Those items are called student learning outcomes. When we’ve taught those things we want to determine whether the student has learned them. This is called assessment.
When we assess, we are assessing a communication channel. The channel has two ends, the transmitter and the receiver. We give a lot of attention to the receiving end of the channel in that we assign a grade of A, B, C, D, or F to the student, and this is right in that it is the student who is paying money to learn and those who hire that student will want some measure of how well the learning has taken place.
But there are two ends of that channel. For my own sake as a teacher, I need to know how well I am doing and I need to change what I do in order to make it better. When I learn what I need to change, I should carry it out.


Over the last twenty-odd years there have been numerous changes in technology. When I began teaching, I had to walk two blocks across campus to check my e-mail. I did it once a week, and frequently I didn’t have any. Now, God help me if I skip a day of clearing out the junk mail.
Regardless of the downside, technology has provided more tools to reach out to the student. I would like to mention three of these: PowerPoint, lecture-capture, and the learning management system.
Much has been made of PowerPoint and the effect it has had on bullet-pointing the learning process. There is a danger than the medium will affect the message. That is a legitimate concern. PowerPoint is not a panacea. I still use the chalkboard in many classes as I believe that the students need to see the practitioner at his (or her!) craft so that they may be empowered themselves.
If I’ve gotten my hands dirty, then they know they might have to get their hands dirty.
Yet there is so much material that simply needs to be put in front of the student and talked about. This can be done badly with PowerPoint, yes, but it can also be done badly with acetate slides and with chalk on the board. What I put on PowerPoint, I used to put on sheets from a yellow-pad and just transfer it to the board. There is no loss of empowerment and there is a great gain in efficiency.
As far as the bullet-point-ization goes, this is taken care by talking and adding meaning while you lecture over the slides. The audio can be recorded and synchronized with the slides on your computer by what is called--straightforwardly enough--lecture capture. It can be done in an almost effortless fashion and uploaded to the Internet.
On the Internet, we has access to our university’s learning management system (LMS). We can upload the PowerPoints ahead of time for the students to print-off and take notes on. We can put the captured lectures there by them for the students to listen to.
I have my whole course online, organized in four parts. In each part there is a schedule of what I plan to do everyday. All of the deadlines are there. I’ve interspersed quizzes for the students to take online, where they are automatically graded and recorded.
For the students who want to learn, there is unparalleled opportunity. The students who don’t want to learn are forced to be more creative in their excuses. It’s win-win in other words.

The Human Touch

Yet with all of this, we cannot forget the human touch. If we don’t care, the students won’t care. If we aren’t excited, the students won’t be. If we don’t think the material is important, then the students won’t.
In other words, you still have to teach.


Prof Emeritus said...

Yes, students do not see the big picture in statistics. I think that it can taught that way, but I have retired and our text (ie. co-authored text) didn't sell. So, I will not try to write it. But I loved to teach it.

saranya magesh said...

interesting blog. It would be great if you can provide more details about it. Thanks you

Mathematics and Statistics Jobs