During a conference call between the US Secretary of Education and reporters as published in an article in Chronicle Journal of Higher Education on November 5, 2013 (please click here to read the whole article). Basically there are three points raised by the reporters on the application of CAR.
- Data aren’t good enough to calculate the rating.
- Ratings will penalize colleges that serve the neediest students; and
- Ratings will punish colleges and students for factors they cannot control.
AAEA would like to comment each of the issues which have been raised. In this page we will comments on “Data aren’t good enough to calculate the rating”
Let us critically analyze the first comment on the data. The skeptics argue that the ratings are calculated based on flawed data. The Secretary of Education has admitted that the data may not be perfect. He further said that “imperfect data should not be an excuse for inaction”.
AAEA’s comments:
- Imperfection of data will always be there. However, there are statistical techniques and approaches which can be used to minimize such an imperfection. For example, if the benchmark is calculated based on the state’s average (statistical mean) number; one can apply the standard deviation to overcome such random errors.
- Having said that in point 1, perhaps using the statistical average (mean) has advantages than of using the median.
- However, the median is better if the sample size has a lot of missing values and if observations from different institutions are unequal in size.
Please let us know your comments below. Should or should not data imperfection be used as a reason for not applying the CAR?