Geez, where the heck is the summer going? I can't believe it's August already! I had planned to have the data course completely prepped by now so that I could spend August working on other things but of course, things never turn out as planned. Thanks to several trips in July, I'm not nearly as prepped as I'd like but I thought I'd share something about what where I'm at. As I mentioned earlier in the summer, this course is not a typical statistics course - it really is a data analysis course. And I'll be teaching it using team-based learning. The TBL website and listserv have been incredibly helpful with that part.
When I say that this is more of a data analysis course, I mean that we will talk about things like choosing what to count (e.g., if you want to know how income is distributed across households, how do you define 'income'? Once you settle on a variable, do you look at mean, median, different percentiles?) and how to measure those variables (e.g. the Current Population Survey is a household survey so what issues might arise when people self-report income? how do we compare income across geographic areas and time?). Then we'll do a whole section just on reporting data (choosing what type of graphic is most appropriate for your analysis, why the scale of the axes matter, how to read/create tables, etc), and finally, discuss the type of comparative analysis that economists do (e.g., correlation vs. causation, confounding variables). Along the way, they will have to go get data and do stuff with it (calculate descriptive stats, simple regressions) but that is all stuff that technically, they should already know how to do so I'm not going to take any class time to talk about those concepts (though I'll spend a little time talking about how to compute them in Excel).
That last part is tricky - I'm going to be strict about enforcing the prerequisites but just because students have taken a stats class already, that doesn't mean they actually learned (or have retained) much. They certainly need the basic stats concepts in order to talk intelligently about the data, but although they should already have seen those basic concepts, my understanding is that I shouldn't expect them to really have retained much. On the one hand, I don't want to spend too much time on stuff that they should already know; on the other hand, we really can't do much with the data if they don't really know the basic stats concepts. Right now, I'm planning to give students a knowledge survey at the beginning of the semester that will include all the statistics concepts that I will expect them to have at least a basic understanding of; I will tell them that they should use that to gauge how much extra review they will need to do. Then, for each of the modules (I've divided the course into five, maybe six, main topics), I will give them a reading guide that will include the specific stats concepts that they will need to review. Even though they don't have a stats textbook for this class, I'm going to provide some online sources and put a few stats texts on reserve at the library so they have the resources if they need them (if I have time in the future, I may develop some notes of my own but I am pretty sure that's not happening this summer). In general, the students need to have a conceptual understanding of the statistics, more than the technical ability - they need to know what a 95% confidence interval REPRESENTS but I'm less concerned with whether they can CALCULATE it.
In TBL, each module begins with a Readiness Assurance Test which is based on the readings and, in this case, the students' own review of the stats concepts. Then we'll spend class time doing application problems. I'll still in the process of writing all those questions, plus fully fleshing out the individual assignments that students will do on their own outside of class. But it's all definitely coming along...
When I say that this is more of a data analysis course, I mean that we will talk about things like choosing what to count (e.g., if you want to know how income is distributed across households, how do you define 'income'? Once you settle on a variable, do you look at mean, median, different percentiles?) and how to measure those variables (e.g. the Current Population Survey is a household survey so what issues might arise when people self-report income? how do we compare income across geographic areas and time?). Then we'll do a whole section just on reporting data (choosing what type of graphic is most appropriate for your analysis, why the scale of the axes matter, how to read/create tables, etc), and finally, discuss the type of comparative analysis that economists do (e.g., correlation vs. causation, confounding variables). Along the way, they will have to go get data and do stuff with it (calculate descriptive stats, simple regressions) but that is all stuff that technically, they should already know how to do so I'm not going to take any class time to talk about those concepts (though I'll spend a little time talking about how to compute them in Excel).
That last part is tricky - I'm going to be strict about enforcing the prerequisites but just because students have taken a stats class already, that doesn't mean they actually learned (or have retained) much. They certainly need the basic stats concepts in order to talk intelligently about the data, but although they should already have seen those basic concepts, my understanding is that I shouldn't expect them to really have retained much. On the one hand, I don't want to spend too much time on stuff that they should already know; on the other hand, we really can't do much with the data if they don't really know the basic stats concepts. Right now, I'm planning to give students a knowledge survey at the beginning of the semester that will include all the statistics concepts that I will expect them to have at least a basic understanding of; I will tell them that they should use that to gauge how much extra review they will need to do. Then, for each of the modules (I've divided the course into five, maybe six, main topics), I will give them a reading guide that will include the specific stats concepts that they will need to review. Even though they don't have a stats textbook for this class, I'm going to provide some online sources and put a few stats texts on reserve at the library so they have the resources if they need them (if I have time in the future, I may develop some notes of my own but I am pretty sure that's not happening this summer). In general, the students need to have a conceptual understanding of the statistics, more than the technical ability - they need to know what a 95% confidence interval REPRESENTS but I'm less concerned with whether they can CALCULATE it.
In TBL, each module begins with a Readiness Assurance Test which is based on the readings and, in this case, the students' own review of the stats concepts. Then we'll spend class time doing application problems. I'll still in the process of writing all those questions, plus fully fleshing out the individual assignments that students will do on their own outside of class. But it's all definitely coming along...
Will you share your knowledge survey? We've been having something of a discussion about what we expect (or should expect) our undergrad business majors to know about stats when they graduate, and using something like this to assess what they actually know would be useful. At least I think it would, although my colleagues may not agree.
ReplyDelete