Informed Consumption of Quantitative Research

20-22 February 2012

The type of social outcomes that are studied in public and non-profit administration are usually determined by a complex set of factors. It is consequently difficult to understand the effect of a given policy or change in a specific factor on outcomes. Experimental scientists solve this problem with careful laboratory controls that let only a single factor vary. Most social scientists and policy analysts do not have the luxury of a controlled laboratory setting and must analyse pre-existing observational data in which multiple variables change. Statistical tools such as multiple regression are well suited to this type of data but when misapplied, they can give misleading results. This course trains participants in recognising the proper use and abuse of methods in quantitative research.

Methods such as multiple regression are popular in all types of social analysis from business to government because they can isolate and estimate the magnitude of a single effect on an outcome when several potentially causal variables are at play. This precision cannot be offered by alternative methods such as case studies or uncontrolled comparison. Also unlike many alternative methods, regression can quantify the degree of uncertainty in its estimates.

Effective managers increasingly need to understand multiple regression results in social analyses and research reports. Most critically, they must be able to discern properly designed and estimated models from flawed work. And when they suspect a flaw, it is of great advantage to understand whether and how it likely biases results. Participants will complete this course with an enhanced ability to know when to believe the numbers.

Convenor