The AEI now has a video of the Lott/Hassett show alleging media bias against Republicans. (My previous post is here.) What is wrong with their study can be explained very briefly. First look at the graph on the left of unemployment rates. Now, here's what Lott and Hassett say:
"In the case of unemployment, 44 percent of the headlines under the Clinton administration were positive while that same number was only 23 percent under Bush II. By comparison, the average unemployment rates were fairly similar, 5.2 percent under Clinton s eight years and 5.5 percent under Bush during the sample. There is also a great deal of overlap (3.9 to 7.1 percent under Clinton to 4.2 to 6.4 percent under Bush II)."
What they fail to mention and what is obvious from the graph is that under Clinton the unemployment rate decreased from 7.1% to 3.9%, while under Bush it increased from 4.2% to 6.4%. Maybe, just maybe, that's why the headlines were more positive under Clinton. In fact, there seems to be evidence of bias against Clinton---why were only 44% of the headlines about unemployment positive when it just kept going down and down to the lowest levels in decades? Oh, and don't expect to see a graph of the unemployment rate anywhere in their paper or presentation.
Now, they claim to have controlled for level and trends in unemployment in their analysis, but of course they have not. The only control they have for trend is the change since the previous quarter and it is obvious that changes over longer terms will affect the reporting. Do Lott and Hassett believe that no-one ever compares the unemployment rate with what it was a year or two before?
Another look at the graph of the unemployment rate will reveal the futility of the whole exercise. Is it not obvious that unemployment did something fundamentally different under Clinton than under either Bush? To get a meaningful comparison you would need to compare Clinton with a Republican administration where unemployment declined for eight years.
Meanwhile, Donald "Stalker" Luskin has leaped to Lott's defence. His defence is more than a little odd. First, he calls DeLong "Jabba the Economist". I don't see how alleging that DeLong looks like a character from Star Wars proves that he is wrong about Lott. And if DeLong is cast as Jabba the Hutt in that scene, it conjures up unfortunate images of Mary Rosh as Princess Leia in a bikini. Second, he asserts that DeLong only attacked Lott because Hassett allegedly showed him up in paper published over ten years ago. Somehow he didn't notice that Lott and Hassett were different people. Third, he claims that DeLong is a hack because he didn't tell the NY Times reporter that he was only criticizing Lott to revenge himself against Hassett.
King at SCSUScholars displays a wilful refusal to consider the possibility that there is anything wrong with Lott's research, dismissing it all as proof of the NY Times liberal bias.
If anyone is willfully refusing to consider possibilities, it is you that has decided to smear several journals, its editors and referees, and any of Lott's co-authors by taking a perceived flaw in one work as the right to dismiss any other work that person does. It's lazy.
If you think the data will support your claim that Hassett and Lott have the wrong lag lengths, get the data and run it yourself. If and when they try to publish this study in a refereed report, your question should be investigated. But you are assuming the conclusion to an empirical question without testing your hypothesis. Again, lazy.
kb - I would bet on if more than when if I were you. As far as Lott's work, did you not read any of the content of the above post? If it is not clear to you the extent of intellectual chicanery and misrepresentation vis a vis Lott and Hassett's characterization of unemployment levels in their latest "study" then I suggest you get your eyes checked, and have another look at the graph that heads this post. BTW, I suggest you read some of the stories in the archive here, as you will find it is considerably more than one "perceived" flaw in more than one "work".
Also, kb, why is Lott such an important case for a defence?
If you insist on testing a hypothesis such as Lott's, or perhaps mine (that the moon is made of green cheese), you should go right ahead of course. But in the meantime - can you see anything wrong with Tim's graph? If something is simple enough, without being "too simple" or simplistic, why not prefer it to a complicated protest from a partisan hack of seriously discredited credentials? ("... and I have to say that he was the best professor that I ever had ... There were a group of us students who would try to take any class that he taught" and blah blah).
Is it that despite his faults Lott is a researcher of such worth that you'll overlook a thousand humbler people's efforts in order the better to spend your time on him? Or is there something you prefer in his particular prejudices? Perhaps you think that all those more modest folk of less thrusting ambition than Lott must in truth have more than Lott to be modest about? Seems doubtful to me.
At this point I'll settle for Tim's little graph, and thank him very much for his efforts.
Umm this looks bleedingly obvious but I have to know. If I go from 90kg to 100kg and then 100kg to 90kg, because in both instances I averaged around 95kg. It was unfair of people to comment in the first instance that I was packing on the weight?
If you don't want to hear that you're packing on the weight in that case, apparently you should hang about with rightwingers.
This is similar to the use of averages in economics claims we hear in the States too -- in things like income or tax breaks, the obviously inappropriate use of averages. Odd that people have trouble seeing, in these examples, the rather obvious misuse of statistics.
That is a GREAT graph. If that doesn't convince people there is a fundemental difference between the Bush years and Clinton, I want what they are smokin!
Actually I don't smoke or do drugs.
What a silly hypothesis! I'd love to see their criteria for defining positive or negative headlines (How many times do you get to define your dependent variable? give me a job like that!) But I suspect if they used the over-the-month change in unemployment as the independent variable, their model would have an R2 of 1, that is, if the inclusion of the Bush, trend and Clinton variables didn't achieve some form of multicollinearity and blow the model up.
Has anyone actually looked at the Lott and Hassett paper rather than simply taking Lambert's word for it? While most of the regressions take the change in economic variables over up to a quarter, L&H do try looking at economic data up to six previous reporting periods on page 15. See their paper at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=588453. What am I missing?
Bob H, you are missing that their lags only go back six months. It seems natural that the nwspaper stories might discuss how things have changed over a year or longer.
If I understand correctly, the issue is picking up trends. My look at your graphs seem to imply that six months into the past plus the current month would give you a pretty good idea of the trend. L&H also have variables for whether there is a recession. Your graph shows trends, but I guess you will have to explain why these variables won't pick up the trend and the general economic conditions enough. It just seems like you are making an assertion that honestly doesn't seem particularly strong. What in your graph indicates that their seven periods plus the recession variable are not enough?
Six months is a short term trend. They have no controls at all for longer term trends.