Step 1 (Due 3/7) Find a published article that estimates a linear model. Each student must report on a different article. (I’ll keep an up-to-date list of claimed articles on Canvas. First come, first served). You will be analyzing the estimation of only one model from the paper. Send an email to the Professor with the name and author(s) of the article, a link to the article (or a PDF) and the page number(s) for where the estimate you will discuss is presented. Please be sure that all three of these things are included in the email. You may want to choose papers from journals outside the field of economics as they are more likely to use simpler techniques. Please consider what you are going to have to write about the paper in choosing it.
Step 2 (Due 3/14) Write an introduction to your paper that is up to two single spaced pages in 12pt type with 1” margins. The introduction should describe the motivation for computing the estimate(s) of interest. The motivation should explain what question is being asked of the data and should be specific about whether the author is trying to develop a model for prediction, to estimate a causal effect, or simply to show correlation (or partial correlation). The introduction should also briefly describe the data and methods being used. Turn this in on canvas.
Step 3 (Due 3/21) Turn in a completed paper with an introduction, an analysis section, and a conclusion. In the analysis section consider each of the assumptions we need to prove the Gauss-Markov theorem. In discussing them ask the following questions
Does the writer consider that the assumption might not hold?
If so, what does the author propose doing about it or how does the author qualify conclusions of the analysis due to the possible failure of the assumption?
If not, can you see any possibility that the assumption would not hold? Give a specific example. How would that affect the conclusion the writer draws from the analysis.
In doing this cite the page (and if appropriate the line number on the page) where the issue is discussed or where you think it should have been discussed if it was not.
In your conclusion explain why you think the analysis was or was not convincing.
The paper can be up to 8 pages long.