by CDO Review Staff | Mar 4, 2019 | Chief Data Officer, Issues & Insights
Marie Kondo’s advice about keeping only things that “spark joy” has a great deal of relevance for today’s data managers. Not all data is equally important and the costs of hoarding unproductive data can be quite substantial. Ian Matthews elaborates on why more data is...
by CDO Review Staff | Mar 4, 2019 | Issues & Insights, Statistics
Deborah G. Mayo, a philosopher of error statistics, expounds on the “myth of objectivity” in this excerpt from long article in errorstatistics.com: We invariably sully methods of inquiry by the entry of background beliefs and personal judgments in their specification...
by CDO Review Staff | Mar 4, 2019 | Issues & Insights, Statistics
Almost all researchers use statistical methods but only few researchers receive intensive training in such methods. This has resulted in a wave of statistical malpractice and butchered numbers that has coursed through the ranks of even the most prestigious scientific...
by CDO Review Staff | Mar 4, 2019 | Issues & Insights, Statistics
The traditional approach is to formulate a hypothesis and then look for data to validate that hypothesis. Yet many data scientists consider it a virtue to come to a trove of data with unbiased eyes and simply find patterns already present in it. The idea is not to try...
by CDO Review Staff | Mar 4, 2019 | Issues & Insights
Zombie statistics are false facts that just never seem to die. Examples include such whoppers as “People use only 10% of their brains” and “You need to drink eight glasses of water a day”. Less obvious zombie statistics haunt researchers in every field and steps must...
by CDO Review Staff | Mar 4, 2019 | Issues & Insights, Statistics
Highly-accurate statistical models may not work in real world contexts as has been proven by the large amount of published research that has resisted replication. Stephen Chen blasts the sacrifice of model accuracy for predictive power in this report from Predictive...
