… after I had a suspicious growth on my forearm examined. Luckily it was basal cell, a carcinoma that doesn’t spread and only needs to be excised – problem solved. Unfortunately, the tissue sample came back positive for squamous cell, a similar type that doesn’t spread. So I had to undergo Moh’s surgery to remove more tissue, a lot more tissue: That area in the...
Sampling theory in American fiction
From Never Go Back: He said, “No one is here to work with you, boy. We’re here to kick your butt and take our car and our money back.” “OK,” Reacher said. “We can go down that road, if you like. But there’s no reason why all of you should go to the hospital. You ever heard of Gallup?” “Who?” “It’s a polling organization. Like at election time. They tell you this guy is going to get fifty-one...
Good quote about RCTs
“In other words, discovering program improvements that really work is extremely hard. We labor in the dark — scratching and clawing for tiny scraps of causal insight.”
On Megan McArdle’s blog.
Short guide to graphics and data
Schwabish, Jonathan A. 2014. “An Economist’s Guide to Visualizing Data.” Journal of Economic Perspectives, 28(1): 209-34. Abstract: Once upon a time, a picture was worth a thousand words. But with online news, blogs, and social media, a good picture can now be worth so much more. Economists who want to disseminate their research, both inside and outside the seminar room, should...
The debate over RCTs in health policy research
Interesting discussion of problems with RCTs
And it’s not the standard, “too difficult, unethical, etc.” arguments you always hear:
Why Cohen’s effect size rule should be avoided
From Durlak (2009), How to select, calculate, and interpret effect sizes: Many authors still routinely refer to Cohen’s (1988) comments made in reference to power analysis that SMDs of 0.20 are “small” in magnitude, those around 0.50 are “medium” and those around or above 0.80 are “large.” In terms of r, Cohen suggested corresponding figures of 0.10, 0.30, and 0.50. What many researchers do...
Exporting Stata output
This is the most comprehensive list I have seen on all the different ways of exporting Stata output into Excel:
Have we reached peak NSSE?
More trend data from their annual reports: There is a clear decline in the past five years. While some of this is probably due to institutions using the NSSE on a less than annual basis, I think there is a real decline here (the fact that schools may be using it less frequently is also an indicator of how useful they are finding it). My prediction is that this decline will continue as...
College student survey response rates continue to decline
Depressingly, average institutional response rates for the NSSE have dropped over 10 percentage points during the last decade, to 30% in their most recent annual report. While the content of the NSSE leaves much to be desired, their response rates are a good indicator of what is happening across the country on college campuses. Some of this decline is probably due to survey fatigue; some may be...