height_-_regression.xlsx |

Sociology 211, Elementary Social Statistics

Today we looked at data from RateMyProfessors.com about William Paterson faculty. The names were removed from the data file to protect the innocent (and the guilty).
We looked at the distributions of each of the variables. The distribution of the hotness ratings were skewed to the right. The distribution of overall ratings was skewed to the left. Then we looked at the data in order to answer this question: Do easier professors get higher overall ratings? We looked at a crosstab of overall ratings by easiness. We found that the answer appears to be yes. An overwhelming majority (over 70%) of professors who were rated as very easy also got very high overall ratings. Meanwhile, professors who were rated as very difficult were very unlikely to get a very high rating. OK - if people rate their professors partly based on how easy the class was, how do we know who is really a "good teacher"? How do we separate out "good teaching" from simple easiness? Next week, we will look at a method that allows you to do this. For homework, please read pp. 292-297 and pp. 300-305 about regression. This is the technique we will be using with the RateMyProfessor data. You should know the following for the test:
* percentiles * frequencies * cross-tab (how to obtain it and how to interpret it * chi-square test statistic (how to obtain it and how to interpret it.) Complete the following survey to generate data for a demonstration we will do today: Click here to take the survey!
Survey results Quarter quiz 2 coming up next week..As you may know, we have a test coming up next week, but it won't be very difficult. The test will cover: * Percentiles (What does it mean to say that someone's income is at the 50th percentile or at the 90th percentile? What does it mean to say that someone is at the 50th percentile in terms of height? The 90th percentile in terms of height?) We talked about this last week. * Frequency distributions. Just a table telling you the percentage of cases that are in a particular category of a variable. We talked about these in the past as well but we can go over them further. * Looking for associations between variables with crosstabs. We did this with the assignment looking at gender and college at William Paterson. We will do more of these today. What does it mean to say that one variable is associated with another? * Using the chi-square test with the crosstab. We haven't done this but we will get practice obtaining the chi-square test statistic and interpreting it today. So the test will cover four topics -- but the emphasis will be on the last two topics (associations with crosstabs, and chi-square tests.) In order to demonstrate statistical inference, we are going to take a brief, anonymous survey as a class. Click here to take the survey!
Up until now, we have mainly focused on only ONE variable at a time. Occasionally, a second variable crept in, but we have mainly focused on one.
Now, we're going to start looking at two variables at one time. In particular, we're going to look at associations between variables. What does it mean to say that one variable is associated with another? It doesn't mean that the two variables are similar. Instead, a variable X is associated with another variable Y if knowledge of X enables you to predict Y with better than chance accu |