what is the meaning of " treated observation"? [closed]

what are the meanings of "treated observation" and "non-treated observation" in this context ?

Further, propensity score matching is a more robust method of matching that quantitative accounting researchers frequently use (Rosenbaum & Rubin, 1983). With this technique, the researcher first estimates a logit or probit regression with the treatment effect as the dependent variable (i.e., S&P 500 membership, continuing with our earlier example) and independent variables that are known determinants of the treatment (possibly economic performance, firm size, age, industry, etc. in our example). From this first-stage regression, the researcher obtains propensity scores.

6 The researcher then matches each treated observation to a non-treated observation based on the distance between their propensity scores (i.e., those observations with the closest propensity scores are matched).

7 Hoi et al. (2013) utilize propensity score matching in their study that links irresponsible CSR activities with tax aggressiveness. They first generate predicted propensity scores for each firm-year using a model of the determinants of a firm engaging in a high level of negative, irresponsible CSR activity. Then, they match a treatment firm (one with a high level of negative CSR activities) to a non-treatment firm (one with a low level of, or no, negative CSR activities) by matching observations with the closest propensity scores. Hoi et al. (2013) then re- test the relation between negative CSR activities and tax aggressiveness using this matched sample. These matching techniques help mitigate the researcher’s potential failure to consider certain firm characteristics that differ between treatment and non-treatment firms (Dhaliwal et al., 2012).

Non-treated and treated observations are concepts from medical research. It means observations before and after some intervention (e.g. before and after a patient is given a drug)