The Research Design Associates Blog

(statistical sampling)

Probability is hard to understand and does not really get easier with use

While jumping to conclusions can allow fast, heuristic decision that are often correct and expedient; it can also lead to false conclusions when there is insufficient information, and little reality basis for the conclusion beyond immediate experience. Research that satisfies the premises of the statistical model force logical consistency in the math and offer an opportunity to improve information quality. Careful and thoughtful probability research can provide insights and answers well beyond verbal logic. It is a second set of information available to those who will slow the decision making process and grapple with the novel way of looking at the world.

Read more …

10.14.2012

We’re not so good at understanding probability

We routinely work with brilliant scientists, business leaders, legal adjudicators and policy decision makers of all ages and experiences. Yet these very competent folks rarely use the results of probability statistics as primary components in their decision making. They mostly substitute some form of intuitive math, “experience” or “judgment”. During political campaigns it is always frustrating to hear distorted or down right fabricated numbers being used to bolster one candidate or another. It just makes sense to disregard numbers jumble in situations where there is little capacity to easily verify or grasp truth about what the numbers mean. The tendency to disregard all politician’s’ numbers as lies throws out the baby with the bath, it is completely understandable.  Yet scientists, economists and business decision makers ought to be able to use probability to sort out fact from fiction in the fields where they are expert. But there is little evidence that these probability based judgments are routinely used in any field.

Read more …

10.8.2012

How big a sample do I need?

At least once a week a client asks what is the appropriate sample size. The right sample size depends on what you know, how you are measuring it, and how accurate your findings need to be. Your investment of time and thought in experimental design and defining meaningful measures will do more than anything to insure meaningful and accurate research outcomes.

Read more …

4.29.2012

RAC Attacks: The Current Landscape of Medicare Audits

Improper payments for health care are estimated to range between 3% and 10% of total health care expenditures nationally. The Medicare program has introduced “recovery audit contracting” (RAC) as a way of assuring that proper payments are being made for services, identifying fraud and abuse. The goal is to reduce fraud during 2009 to 2012 by 50%. With intense pressure to investigate fraud and audit contractors incentivized with contingency fees to find improper overpayments, the audit system itself is ripe for error and abuse.

Read more …

1.10.2012