Misused Statistics at U-Michigan
Original Commentary

He uses statistics as a drunken man uses lamp posts - for support rather than illumination. Andrew Lang.

2005 CLAWSON, MI
Original commentary and analysis.
by Chetly Zarko

With recent news about the petition to require the legislature give schools and universities a 5% increase in total allocations every year(sponsored by the K-16 coalition), and universities complaining about how "their share" of the State budget has declined, its time to revisit an old statistical mispresentation that U-Michigan has engaged a number of times in the last 15 years.

The author has written several times (Gurin here, Kaufmann here) about an alarming pattern of unethical misuse of statistics and research at the University of Michigan (by some of its researchers and administrators). Any ethical attack on an institution should be justified by looking to its leadership and the culture they instill by example. Former University President James J. Duderstadt provides just such an example of leadership in the blatant misuse of statistics for political cover. In continuing research by the author into the U-M Bentley historical archives, yet another clear example of data misuse is uncovered - one which the lay reader should be able to understand quickly.

In 1996, as in almost every year, the University was seeking to get the largest possible appropriation from the Michigan legislature. As is customary, lobbyists lobbied, Duderstadt would write op-eds and letters to congressional staff, and data was compiled and sent to the legislature to persuade them to give more.

According to data obtained from a file in Duderstadt's digital collection, now archived at Bentley, Duderstadt seemingly misused a comparison of the percentage of U-M's budget from state appropriations to suggest state support was actually declining. By ignoring differential growth rates, Duderstadt suggests trickily that state support for U-M fell from 18% to 12%. It did only in a limited way though - as a percentage of U-M's overall budget, which the state has no control over. In reality, state support rose in real dollars an average of 8% each year over the period Duderstadt cites, but U-M budgetary growth grew an an outlandish average of 16% per annum. Over the whole 12 year period, due to compounding, U-M's budget tripled while the state's contribution "only" doubled (comparing the ratio of 3 to 2, one would expect that the state support as a percentage of U-M's budget would fall by the inverse (2/3) percentage - - 12% is exactly 2/3 of 18%). Despite the fact that an average per annum growth rate of 8% in state funding vastly exceeds inflation (during the non-inflationary 80s), Duderstadt referred in an op-ed to this as the "... the eroding public support of higher education ...". In multiple speech manuscripts and publications, Duderstadt complains of a media mischaracterization of the problem of "out-of-control higher education costs," and instead shifts blame to declining government funding. A close look at the data reveals Duderstadt's view for the farce that it is and leads to the conclusion that it is the cost of education, not public support for it, that is to blame.

Below is webpage rendition of the Excel Spreadsheet Duderstadt created to "market" the allegedly "declining state support" for U-M. Everything in pink is the original document, data white and other colors represent data I've added with other colors used for highhight and effect. I'd love to be able to also compare this to actual tuition rates and number of students admitted, however, that data is not available

What the dishonest Duderstadt has accomplished through statistical sleight of hand is blame U-M's outrageous growth in expenses on the State of Michigan.

Below are two bar graphs.. The pink is Duderstadt's comparison of "state support" as a percentage of the total U-M budget, the yellowed "stacked bar" graph is a more appropriate representation of state funding showing that U-M administrative spending is rising at a faster rate than the rise in state appropriations.

This, of course, is just one of many examples of where data and statistics have been abused.