Saturday, October 11, 2014

Betty McCollum and Irreconcilable Statistics

Betty McCollum shows us once again the perils of naive thinking and use of poorly handled statistics.

Arguably she does not do the analysis herself, she simply reports what her favorite organizations (agenda driven, not truth driven groups) put out, without adequate investigation. This is an excellent example of the use of "happenstance data".  In this case from the:

Women’s Foundation of Minnesota.
Below is a sampling of data from the Status of Women & Girls in Minnesota, an ongoing collaborative research project of the Women’s Foundation of Minnesota and the University of MN Humphrey School’s Center on Women & Public Policy. Annually, data specific to Minnesota women and girls is gathered and analyzed in economics, safety, health, and leadership.
As is standard for these egregious studies in the misuse of statistical data, it is an "apples to anvils" comparison taken simply by naively saying "compare all men to all women", irrespective of life choices that put them in completely different economic paths and positions. It is a statistical strategy guaranteeing flawed results that will never give a picture other than what they want, not a search for "truth". They are simply looking for "victimization" rubrics. It is a way to continue to demagogue issues, that they feel promotes their political agenda.  It is an "issue" that can never be "solved" if the data is viewed naively. Thus providing perpetual outrage and growth of pet programs.

When the data is more correctly handled, the analysis gives a much different picture. These results were published at the Huffington Post. It is a very thoughtful article with this take-away:
That's not a comparison between people who do the same work." With more realistic categories and definitions, the remaining 6.6 gap [Note: they had explained away most of the purported difference earlier in the text] would certainly narrow to just a few cents at most.
Could the gender wage gap turn out to be zero? Probably not. The AAUW correctly notes that there is still evidence of residual bias against women in the workplace. However, with the gap approaching a few cents, there is not a lot of room for discrimination. And as economists frequently remind us, if it were really true that an employer could get away with paying Jill less than Jack for the same work, clever entrepreneurs would fire all their male employees, replace them with females, and enjoy a huge market advantage.

Read the original, statistically honest/well handled, study here.

Illogical analyses like this study must never be used for determining public policy. That is how we get so many truly bad policy laws. Thoughtful analysis and prudent probative evaluation is required to have the information to make good policy decisions. Decisions that benefit our country, not undermine it.