Flashcards
Case selection 10
Flashcards do conjunto (4)
- 1Synthetic control method -> one is interested in the causal effect of rare/unique events that pertain to highly aggregate units such as regions (the effect of a specific policy, civic war)Synthetic control method basic ideas:
->a combination of comparison units (synthetic control) often does a better job of reproducing the characteristics of the case of interest->Use of a weighted average of comparable units, where the weights are determined by the control cases‘ (observed) similarity to the case of interest->Aim is precise quantitative inference in comparative case studies - 2Statistical procedure of the synthetic CM: Required: lengthy time-series data about outcome Y and predictors X for both the treated unit and controls, before and after the event->Synthetic C estimator of T effect is given by the comparison of post-intervention outcomes between the treated unit and the synthetic controlRobustness checks:
->In-time placebos: application of the procedure to dates when the intervention did not occur and In-space placebos: re-assign the intervention to control units - 3Case selection(CS) for mechanism based explanation basic ideas: ->Inferences about the average magnitude of a causal effect are pretty silent about how the effect occurs (CM)->Investigations in to causal mechanisms often require intensive analyses of a small number, or even a single case(process tracing)->CSis crucial, for in choosing cases, one also sets out an agenda for studying those cases->Seawright and Gerring propose to use large-N techniques(mostly regression diagnostics) for case selection in small-N studies
- 4Objectives of CS: Obtaining representative cases, generating useful variation on the dimensions of theoretical interest->These (large-N quantitative) methods require the population be reasonably large, relevant large-N data must be available, No a priori interest in a specific caseDeviant CS (present theoretical anomalies to be analysed in detail and representativeness is an open issue); Diverse CS (choose cases to maximise the variance along relevant dimensions to enhance the sample's representativeness); Influential CS (the choice is motivated by the need to check the assumptions behind some general model)
Informações
- Descrição
- FlashCards sobre Case selection 10, criado por tri_nana em 28-06-2015.
- Criado por
- tri_nana
- Visualizações
- 0
- Avaliação