The unit-specific quantity can take different forms:

• Descriptive: The factual value that some variable actually would take for unit $i$ in the absence of intervention (e.g., $Y_{i}$: whether person $i$ is employed).
• Causal: The value some variable would take if a treatment variable $D$ were set to a particular value $d$, producing a potential outcome (e.g., $Y_{i}(d)$: whether person $i$ would be employed if they received job training).
• Causal Difference: A difference in potential outcomes (e.g., the effect of job training on the employment of person $i$).
• Caused by Multiple Variables: An outcome realized under intervention of more than one variable. This includes:
  • Mediation: A claim about the outcome $Y_{i}(d, m)$ that unit $i$ would realize if the treatment were set to $D = d$ and some mediator were set to $M = m$.
  • Interaction: A claim of whether person $i$ would be employed if they received job training and childcare.