Master Methods, quantitive

Alla Inga Spara

• The exogeneity assumption no reverse causality; no OVB
• Choice of control variables affects both x and y; are uneffected by x and y; not too strongly correlated with x; data quality
• OLS assumptions The exogeneity assumption; linear relationship; no systematic measurement errors; well-behaved residuals; technical issues
• Outliers discrepancy; leverage; influential observations
• Assumptions for experiments random assignment; non-interference, excludability
• types of random assignment simple; complete; clustered; blocked
• Common problems in experiments non-compliance; attrition; spillover
• 3 types of spillover displacement; communication; social comparison
• Evaluation of natural experiments plausability of as-if random assignment; credibility of model; relevance of the treatment
• Assumptions for RDD as-if-randomized; no manipulation; continous function around the threshold on both sides; nothing else changes at the threshold
• Assumption DD parallel trends

Alla Inga Spara