AI-ordlista
Den kompletta ordlistan över AI
Differences in Differences
Quasi-experimental method estimating the causal effect of a treatment by comparing outcome changes between treated and untreated groups before and after the intervention.
Parallel Trends Assumption
Fundamental postulate of DiD stating that in the absence of treatment, treated and control groups would have evolved parallel over time.
Average Treatment Effect (ATE)
Average causal effect of the treatment on the entire population, estimated by the difference in differences between groups and periods.
Treatment Group
Subpopulation exposed to the intervention or treatment of interest in the Diff-in-Diff analysis.
Control Group
Subpopulation not exposed to the treatment, serving as a counterfactual to estimate the evolution of the treated group without intervention.
Fixed Effect
Time-invariant characteristics for entities (individual fixed effects) or periods (time fixed effects) controlled for in the DiD model.
Interaction Specification
Multiplicative term between treatment and period variables capturing the differential causal effect in DiD regression models.
Pre-trends Test
Empirical validation of the parallel trends assumption by comparing pre-treatment trends between groups.
Heterogeneous effects
Variations in treatment effect based on individual characteristics or exposure time in DiD analysis.
Diff-in-Diff with matching (PSM-DiD)
Combination of propensity score matching and DiD to improve group comparability by balancing observable characteristics.
Synthetic Diff-in-Diff (SDID)
Hybrid approach combining synthetic control and DiD to create an optimal weighted control group.
Time effects
Factors invariant between entities but varying over time, controlled to isolate the causal effect of treatment.
Group effects
Characteristics invariant over time but varying between entities, controlled to eliminate selection biases.
Treatment indicator variable
Binary variable identifying treated units (1) and untreated units (0) in the econometric DiD specification.
Pre-treatment period
Time interval before the intervention used to establish parallel trends and validate causal identification.
Post-treatment period
Time interval following the intervention where the causal effect is measured by comparison with the pre-treatment trend.
Two-way estimator
DiD estimation method simultaneously including individual and time fixed effects to control for confounding factors.
DiD validity criterion
Set of conditions (exclusion, monotonicity, parallel trends) ensuring causal identification in DiD models.
Robust DiD estimator
Modified version of DiD resistant to violations of the parallel trends assumption through weight adjustment.
Dynamic treatment effects
Evolution of the causal effect over multiple post-treatment periods to capture impact delays and adaptation.