The Student Loan Problem is Worse than I Thought

Power Stats does allow querying on "all those with loans," both in that particular academic year and cumulatively. One of the main benefits, however, is you have the option to incorporate all the students & parents who did not borrow into the analysis. You also get to look at completers vs. non-completers. Many analysts, including College Board, use these raw data sets, but anyone can go in and run their own queries, for example, with or without parent PLUS, and removing the spin that organizations such as CB are accused of including in their narratives. You can also easily separate out undergrad vs. graduate/professional debt. If FICO looked at them separately, the results would likely be similar to NCES', as you will find hardly any undergrad-only students incurring debts greater than $100K. http://nces.ed.gov/pubs2013/2013156.pdf http://trends.collegeboard.org/student-aid/figures-tables/distribution-cumulative-debt-among-2009-bachelors-degree-completers-and-non-completers http://trends.collegeboard.org/student-aid/figures-tables/distribution-total-undergraduate-debt-sector-and-type-degree-or-certificate-2007-08 One person's "timely" is another person's "not peer-reviewed." Regarding timeliness, the reality that the govt's data focus has been on K-12 ever since NCLB may certainly be misguided and debatable. The fact that higher ed has been losing the battle for resources to K-12 is not something to be celebrated, although some higher ed lobbyists may be just as happy that the data is only available every four years. While I haven't done a recent Nexis search, a news release with a figure of 0.1% would be far more of a "myth-buster" of current conventional wisdom than 4.25% greater than $100,000. It sounds like you are not engaged "in the scientific pursuit of understanding" but rather hopping from one hot story to another. I would expect this of the general news outlets but would have hoped to expect more of Education Sector. In fact much about student loan performance is counter-intuitive, particularly to conventional wisdom, such as the gradual/slight decline in delinquency rates from during a recent 10yr period, 2001-2011, that included two economic downturns. I had wondered whether Ed Sector would be focusing on independent data analysis leading up to reauthorization, but maybe I am misunderstanding the purpose of these types of organizations. While I do not have a solution to all of these dilemmas, I also don't believe that pursuing the perfect should result in discarding the good or the very good. While it is opinion that 2008 data set x is still better to use overall than 2012 paper y, it is not opinion but fact that it is risky to pull from analyses using different data universes, different segmentations, different definitions, different purposes, and so on. Your policy prescription, should you choose to recommend one, would depend on the precision of these analyses. For example, if the FICO report leads to changes in federal education loan requirements, then this prescription would be largely misdirected, if it turned out, for example, that only 15% of FICO's greater than $100K group had federal loans. "formed in ignorance of this data" -- Is this to say that a company or a person can publish an analysis and not provide details about methodology, while at the same time anyone who questions it, even based on informed knowledge of similar studies, is ignorant? That sounds like a Catch 22.

Why is there an automatic assumption that any of this is remotely correct? The FRBNY analysis is inaccurate, due to data set issues and misunderstandings of the major differences between federal student loan products, private educational loan products, and other consumer loans and mortgages. Yet at least FRBNY largely discloses the details of its data sample and methodology. FICO does none of that. What FICO does do is state that it uses the same problematic type of source as FRBNY, a tiny sample from one of the national credit bureaus but then stops there. Was this meant to be a \"Monday humor file\"? Apparently what you are saying is that you are letting someone who is not a specialist in the field completely change your long-held beliefs? With your null hypothesis as the percentage being under 1%, you haven\'t disproved the null hypothesis. You may want to use PowerStats, which would allow a variety of analyses, such as (1) top 1% of all borrowers, whether dropped or or completed, (2) top 1% of borrowers who completed a degree or certificate, or (3) top 1% of borrowers who completed a bachelor\'s degree.