In the past, I have used power calculators like GPower for straightforward research designs, but I have never really found a good, practical resource on executing the mechanics of power simulations, especially when research designs and analytical frameworks are more complex. This great article provides a framework (with R code) to execute power simulations and may be useful to students in Research Seminar (AEM 6991). https://cameronraymond.me/blog/power-simulations-in-r/ . There is also a GitHub posting at https://github.com/cameron-raymond/power-simulation-tutorial.
Here are a few resources that students may find useful for research seminar projects that involve data scraping. This list of resources may be updated from time to time:
- Practical Introduction to Data Scraping Using Python – https://realpython.com/python-web-scraping-practical-introduction/
- Python, Data Scraping, and Beautiful Soup Crash Course – https://www.youtube.com/watch?v=XVv6mJpFOb0
- Facebook Data Scraping and Sentiment Analysis Using Python – https://www.danielherediamejias.com/facebook-scraping-and-sentiment-analysis-with-python/
- Twitter Data Scraping and Python – https://www.natasshaselvaraj.com/how-to-scrape-twitter/
Here are some overview papers that may be of interest to people that are learning about ties between machine learning and economics. Admittedly, it helps to have some formal background on causal inference to read these papers.
- Mullainathan, S., & Spiess, J. (2017). Machine learning: an applied econometric approach. Journal of Economic Perspectives, 31(2), 87-106.
- Athey, S. (2018). The impact of machine learning on economics. In The economics of artificial intelligence: An agenda (pp. 507-547). University of Chicago Press.
For more background on causal inference, I highly recommend the book by Scott Cunningham, Causal Inference.
I was posed on Quora a question similar to the above question, and below was my response. I highlight in bold, a key insight drawn from behavioral science.
One the one hand, case competitions leverage a lot of skills that career consultants end up using, such as problem solving skills under pressure, teamwork, communications, creativity, and persuasiveness. But there are also many other skills to master in consulting which are not part of case competitions such as street smarts, empathy, functional excellence, engagement management, leadership, industry knowledge, and advisory orientation. In consulting, you serve your client, not case competition evaluation boards.
Recognize that at the core of consuting is mentorship, apprenticeship, teamwork, and learning from experience. Namely, people learn how to become better and better consultants by starting the profession with enough core skills and appetities to get started. After being involved with a number of projects, one learns from others in the firm how to develop more and more skills and one’s professional identity. You may start in one place and find yourself in a very different place at a later point. Of course, the flavor of the firm and how you experience this may vary a lot from firm to firm.
I wouldn’t overindex and project too far into your future how much you would like consulting based on a single indicator. People are not that great at forecasting in complex situations where they have little prior experience and little feedback. On the other hand, if you are getting multiple warning signs, such as you dislike teamwork, dislike 99% of the people in consulting, have gotten feedback from other consultants that you are the wrong for the profession, have gotten feedback from people that know you well and know the profession that you are not a fit, dislike business, dislike companies, dislike people, then I might reflect a bit more. Remember, you might just start consulting with a strong skills and proclivities in a few areas and flourish over time. My keys to success in consulting (as evidenced solely by my ability to stay in the profession for decades) were probably my appetite for variety and my ability to adapt. I also had supportive managers at key pivot points in my career.
In one of my prior Applied Behavioral Economics lectures, I mentioned the notion of not only looking at individual studies but also finding meta studies (essentially studies of studies) to help inform behavioral perspectives. This article covers a meta study of behavioral architecture interventions: https://www.pnas.org/content/119/1/e2107346118
In this article, there were two observations that really stuck out to me:
1) Considering a range of domains (health, food, environment, finance, prosocial) where behavioral architecture is applied, there is the highest effect on food choices and lowest effects in the financial domain; effects are potentially moderated by domain because of lower behavioral costs and lower perceived consequences in the former versus higher behavioral costs and higher perceived consequences in the latter.
2) Decision structure changes (choice architecture) outperforms decision information (information architecture) and decision assistance approaches, potentially because choice architecture approaches require less demand on cognitive information processing, and there is low susceptibility to individual differences and goals. (But remember that we will start to address personalization and individual differences in upcoming classes).
If you are interested in learning more about meta studies and how to do them, I highly recommend the book: Borenstein, Michael, et al. Introduction to meta-analysis. https://www.amazon.com/gp/product/0470057246/ref=dbs_a_def_rwt_bibl_vppi_i2
This post is based on a question I was posed on Quora.
Based on this question, my gut reaction is that they are looking for a sense of a) how deep do you go with similar clients in similar situations, and b) what’s the outside norm in terms of project length for a similar clients in similar situations.
Key things to be aware of though are what is the client thinking and what you are thinking. With respect to client thinking, are they really talking about any client? Or are they more likely to be talking about a similar client in a similar situation? Where on the spectrum? In terms of what you are thinking, do you think the client has the right perception about any client versus similar client?
How you should respond depends on where you think the client is anchored and where you think they should be.
If they are anchored on similar clients in similar situations, then you could qualify to what extent you see them as similar and then give a range of how long you’ve worked with such clients depending on whether X was involved or X, Y, and Z were involved. You can use this opportunity to answer their questions while also giving them a sense of your depth of involvement.
If they are anchored on any client, then you probably have more work to close any deal. They may be just kicking the tires. If you want to close a deal, you’ll probably need to figure out how to shift perceptions so that they see themselves as similar to clients you’ve worked with, work you’ve done, and/or processes you’ve used in the past. The client prospect could very well be a fit for what you do, but there is a psychological gap that should be addressed.
Welcome to Applied Behavioral Economics in Finance and Marketing (AEM 6150)!
This new course is geared toward developing knowledge, ability, and professional skills to apply behavioral economics in business settings. The course is especially geared toward students who may consider future professions in consumer finance, marketing, product development, data science, or consulting/advisory. Because this is a 7-week course, I assume that you have some exposure to behavioral science concepts (e.g., AEM 6140 – Behavioral and Managerial Decision Making or equivalent background). If you do not, I have identified some supplemental readings in the syllabus that you should consider reviewing; you should also contact me (email@example.com) so that we can better assess whether the course is right for you.
In terms of personal background, I am a new, incoming faculty member at Cornell. I have a non-traditional background and have spent more than 30 years in the corporate world with more than a decade of those years setting up behavioral science units and initiatives in the real world (e.g., nudge units). I continue to work full-time as a consultant and researcher in the behavioral economics space. I especially bring my experience in setting up some of the first nudge units in the world to the design of this course (which includes some cases from companies I have worked with). I am also a Cornell alum who has had a very fortunate career, and so I have a deep attachment to the community here. I would like you to be challenged in this course and succeed in life beyond Cornell. If you would like to learn more about my background, please refer to my bio.
Thank you for considering taking this course. Please feel free to reach out to me via email if you have any questions (firstname.lastname@example.org).