Tiêu đề 9.7. Matching Process.pdf ✅

pacmann.io © 2022 – Pacmann AI 59 Matching Process
pacmann.io © 2022 – Pacmann AI 60 Matching Process ● The matching process ○ Selecting matches or ○ Constructing a matched weighted sample. ● Neither approach is the default, and both have their advantages and disadvantages. ● If a true treatment effect indeed exists and is significant, the results from both matching processes would not be really different.
pacmann.io © 2022 – Pacmann AI 61 1. Selecting Matches Approach ● Control observations are chosen to be either included or excluded from the matched control sample. ● A good match is included, while a poor match is excluded. Good Poor Good Poor Good Raw Control Group Control Group After Selecting Matches Good Good Good
pacmann.io © 2022 – Pacmann AI 62 1. Selecting Matches Approach All observations in the matched sample receive equal weights, although some methods may assign different weights based on certain criteria. For example, ● If treatment observation 1 has one matched control, and treatment observation 2 has two matched controls Good Good Ob 2 Good Ob 1 Treatment Group Control Group