Poor survey use

It is regrettable that your articles on the recent survey of faculty attitudes did not bother to find out how questions on the survey were phrased nor did they bother to intelligently ask what those answers might mean.

A particularly egregious example came up in the Oct. 24 article on job satisfaction. Kendall Baker asserts that the faculty is satisfied with the job environment at NIU.

The survey results do not show any such thing. A question on the survey asked faculty to rank their overall job satisfaction, to which 63.2 percent responded that they were satisfied or very satisfied.

Nothing in the results indicate whether these faculty are satisfied with being in their profession or with being at NIU.

The article also noted that 67 percent of the women and 61.9 percent of the men responding to the survey “had overall job satisfaction.”

Professor Lois Self spins a theory of oppression based on this difference, without even bothering to ask whether any difference exists.

The Northern Star might have taken the trouble to talk to someone on the faculty who knows that this difference could arise from sampling errors, perhaps because the faculty responding to the sample are not representative of the faculty as a whole.

Such a person would know there is a standard test, called the chi-squared test, to find out the probability that the difference arises wholly from sampling error.

The result of using the test is that there is a 35 percent probability that the difference is entirely caused by sampling error.

Statisticians quarrel over whether a 1 percent probability is too high and routinely ignore any probability over 5 percent. Professor Self has leapt to conclusions without bothering to ask whether she is talking about something that even exists. the Star should try harder next time.

William Sjostrom

Assistant Economics Professor