I have compiled a sample of firms which have joined industry groups/exchanges between 2014 to 2018. For each company (56 in total), I have at least 4 years of data (firm and industry characteristics like size, performance, i.e., continuous in nature) prior to its decision to join industry group. I have set the binary variable equal to one for the year in which firm has joined an industry group, and to zero for previous four years. I want to find which firm or industry level variables significantly predict firm's decision to join industry group. Do I need a matching (control) group of firms which have never participated in an industry group over my study period, i.e., from 2011 to 2018?
Does the construction of binary variable make sense? By definition, a variable in computer terms is represented by Boolean logic or better defined, Boolean algebra. Its a flag from what I can tell in reading this question. True or false semantics, at the very least, should dictate whether or not a zeros and ones type of variable is what you need in order to code the low level things you are working to write. If it is anything to do with reverse engineering a software compiler, I would steer clear of that kind of logic and sit down and draw out a new/revised flowchart and get yourself a clean build going.
- Badges
- Science topic
- Similar topics
- Econ
- Manufacturing
- Industry