mirror of
https://github.com/jackyzha0/quartz.git
synced 2025-12-24 21:34:06 -06:00
22 lines
895 B
Markdown
22 lines
895 B
Markdown
# Econometric Process
|
|
1. Question
|
|
2. Model
|
|
'do police reduce crime?'
|
|
![[Excalidraw/Drawing 2021-09-01 14.07.57.excalidraw.md]]
|
|
When there's a systematic relationship between epsilons and x, data can be misleading.
|
|
E is for expectation/average
|
|
Crucial assumption:
|
|
$E[\epsilon|\textit{condition}] = 0$ for all x
|
|
The Intuition is that other factors influencing y are not systematically related to x.
|
|
This assumption is often suspect.
|
|
![[Conditional Expectations]]
|
|
3. Data
|
|
Data Types:
|
|
- Cross-Sectional: each obs different 'individual' ie person, city, group
|
|
- Time Series: each observation different time period ie nanosecond, minute, day, week, year
|
|
- about a single entitiy
|
|
- Panel: combination of Time Series and Cross-Sectional data types. Each observation corresponds to an individual at some moment in time.
|
|
4. Estimation
|
|
![[Bivariate (simple) Regression Estimation]]
|
|
|