Title 9.5. Distance Matching.pdf ✅

pacmann.io © 2022 – Pacmann AI 41 Distance Matching
pacmann.io © 2022 – Pacmann AI 42 Matching Criteria: Distance Matching ● Observations are considered similar if they have similar values of the matching variables. ● The goal is to minimize the distance between treatment and control observations, in term of how far their matching variables values. ● By reducing variation in matching variables, distance matching helps address potential back doors.
pacmann.io © 2022 – Pacmann AI 43 Matching Criteria: Distance Matching Standard Back-door Causal Diagram for Distance Matching ● This causal diagram is the kind of thing that distance matching applies to. ○ We have some back doors ○ Measure at least one variable on each of the back doors ○ Match the observations based on these variables, which helps in achieving identification. The Effect Chapter 14
pacmann.io © 2022 – Pacmann AI 44 Distance Matching Example ● In our credit card debt example ● We have two groups ○ Treated Group: their payment was late in April ○ Control Group: their payment was not late in April The Effect Chapter 14