Chapter 6. Models – Foundations for Organizational Foresight

Statistical Surveys and Consulting

When possible, quantification of qualitative information, collected by surveys, will allow you to see if you can replicate your model in independent efforts, a key step to validation. In very data-driven organizations, quantitative approaches may be the only way to build a credible model. Bilal Ayyub’s Elicitation of Expert Opinions for Uncertainty and Risks (2001), is a primer on how to quantitatively aggregate expert opinions about the future in ways that minimize bias. Quantitative approaches help you discover where your expert pool is still ignorant (no better than random at predicting a system) and where it is consistently good at finding patterns. The former result can signal the need for more expert diversity, or expose shortcomings in your model categories or survey techniques.

image88Few foresight professionals today know how to use good statistics platforms like SPSS or SAS, and few clients ask for statistical studies or data-driven models as part of their foresight work, though most should. Given the increasing ease of online surveying, the most common place statistical approaches to foresight are used today are in market research surveys. The better market research platforms like Sawtooth and Marketsight do conjoint analysis, a way of determining what feature combinations of a product or service people value most, over other combinations. Fortunately these platforms get more affordable, powerful, and intuitive every year. Marketsight is just $95/year for academics.

image89

Cabrera & McDougall (2002)

Cabrera and McDougall’s Statistical Consulting (2002) is a practice primer with many case studies on using statistics in consulting work. Ruland’s Guide for the New Statistical Consultant (2014), also offers good practice tips. Andy Field’s Discovering Statistics Using SPSS (2013) is a good guide to both basic statistics and using SPSS, a powerful (but expensive) statistics platform. An introductory course not just in basic statistics, which is already required by all graduate programs, but on using statistics in foresight consulting should be a requirement in modern foresight graduate degrees. Two particularly good intro books are Naked Statistics (2014) and How to Lie with Statistics (1993) on statistical thinking, and OpenIntro Statistics (2012) on basic statistics, probability and regression. The latter is available at OpenIntro.org, a global group dedicated to providing very high quality textbooks free to everyone, disrupting the very-high-priced college textbook market. These are a great pair with free online Coursera courses on Data Analysis, Statistics, and R, the powerful open source data analysis program. In the meantime, in the absence of such courses in our programs, you can look at the books above and work with tutors online. Challenge yourself to get more quantitative in your next project. Start small, and have fun.

Share your Feedback

Better Wording? References? Data? Images? Quotes? Mistakes?

Thanks for helping us make the Guide the best intro to foresight on the web.