I have read this amazing book about a year ago; recently I read another book about randomized trial : The why axis by Uri Gneezy and John List and decided that I can write about both together. As a researcher, planning experiment is our almost second nature. The ideal experiment would be set inside a complete controlled environment with multiple parallel replicates, so that the results observed are unbiased and reproducible. That is far far away from real-world-problems; to study these problems in a complex system, full of biased people with complicated behaviours, we embrace the noise in a form of randomized field trials. The general idea is that we collect as much data from as many different individuals, each with their own ideosyncracies, as possible to find the general tendency amid all the complicated individual biases.
Then why randomized trials? It has become more openly accepted that certain disciplines are dealing with subjects that are hard to predict (i.e. psychology), so collecting empirical data is a more effective way or the only way to study the underlying principles or validating hypotheses.
The randomistas gave a plethora of field trial examples from informing clinical decisions to making social policies. Randomised trails used to be either inhibited or ignored in these disciplines. One example is addressing the impact of knee replacement surgery, the researcher found that prescribing a placebo (incising a cut to convince the patient that the knee is replaced) or a real surgery had the same impact on the health of the recipients. This had led to a policy change to surgery prescription as well as evaluation of the effectiveness of certain treatment.
Similar to the The why axis, both books also gave a lot of examples on randomized trails in social science. These randomized trails have shown to beat traditional wisdoms, sometimes even rationale and logic of the researchers. They gave unexpected results and opened up a new way to study human, yielding fruitful public policies to save resources and improve life. Another example is how to make more people turn up in the poll. Randomized trails showed that by telephone reminder combined with asking people to think about how to get to the polling station is the most effective stategy.
Randomistas also showed how easy it is to set up randomized trials with today’s internet world. The authors used Google ad to show people three different names of the book and chose the official name by going through the data of how many people clicked into the pre-order link. As the book mentioned, google had actively subject its huge user populations to optimize searching, advertisments, and all the other google products in development.
On the other hands, The Why axis has probed into the issues of discriminations and showed that by gaining the information of the randomized trials, we as individual have a chance to correct the bias and restore justice. It also emphasied the “economic discrimination”, the act of poor treatment to a group of people based on economic motives, rather than animus from conventional beliefs.
Both books have showed that sometimes we gained a lot of insights from the empirical data; we shall not shy away from setting up experiments as it is easy in the nowadays big-data era. These data may not fit well with textbook-theory or conventional wisdoms; they are messy and failed many assumptions. But these data serve as a starting point to dive into the complexity of reality, which is essential in improving the rigor and power of our understandings about the world. The books are also about courage, the academic courage to accept ignorance and treat real-life observation as the main subject of studies not annoying outliners we have to tweak to fit the “theory”.