The salary data from College Scoreboard in all from students who receive federal loans. According to the College scoreboard, 21% of Brown undergrads receive federal loans. About 260 CS majors graduate from Brown a year, so we’re talking about maybe 52 students a year. Because of all the different uncertainties and data issues, the CS data sets from Brown or Harvard aren’t very informative. So their high ranking by College Scoreboard in these fields is unreliable.
It is an interesting human characteristic that humans either ignore data, especially when it contradicts one’s dearly held belief, while, on the other hand, people will often accept data analysis conclusions as dogma, without any critical review of the data quality or the analysis, or of whether the conclusions are justified, because they believe that it’s “scientific”.
@CU123 While the rule of Data Analysis is GIGO, often the data is good, and the analysis is good, however, the conclusions are still unjustified. If the data which the College Scoreboard was non-representative, you can have these types of results. Data samples which are too small tend to be non representative, and therefore often produce results like Brown having the highest median salary for CS, or that the median salary for Criminal justice majors who graduated from Union Institute & University was $109,800.
The problem is not that the data or the analysis were garbage. It is that they should not be generalizing from their results, because of the sample size and other issues.
On the other hand, I would accept that median salary they calculate for CS majors from CMU, MIT, or any of the public schools with very large undergraduate CS programs is a reliable indicator of whether these programs are a good choice, financially.