Samenvatting Inleiding econometrie Economie en bedrijfseconomie

Inleiding econometrie – Economie en bedrijfseconomie
College 1:
Learning via tradition  little effort
Learning via experts  little effort
Learning via own experience  causal chain
Problems  no accurate observations, overgeneralization, illogical reasoning

Scientific learning: extend existing knowledge, learning via scientific methods, theory/data/analysis

Association: useful to describe facts, niet zinvol als we willen beslissen wat we willen doen aan iets
Causal effects: understanding relationships between variables and effectiveness of policy
interventions

Omitted variable bias: het vergeten van een belangrijke variabele en daardoor een verkeerde
conclusie trekken

Types of data:
- Experimental = designed to estimate causal effects
- Observational= not designed for causal effects estimation
- Time dimension = a. time series, b. cross-section  samen leiden tot c. panel data

Ecological fallacy = erroneously drawing conclusions about lower levels of aggregration from higher
levels of aggregation

Conceptualization = the process through which we specify what we mean when we use particular
terms in research
Operationalization = the development of specific research procedures that will result in empirical
observations representing those concepts in the real world  about measurement of a theoretical
concept

Ordinal: ordening
Nominal: geen ordening

Reliability = the quality of measurement method that suggests the same data would have been
collected each time in repeated observations of the same phenomenom
Validity = a term describing a measure that accurately reflects the concept it is intended to measure

College 2:
Regression models: any relationship between two variables
Linear regression = continuous dependent Y as function of any kind of variable X

Scatterplot : x and Y variables
Positive relationship = als het rechts omhoog gaat
Negative relationsihp = als het rechts naar beneden gaat

,Covariance=




- Tells us if X and Y tend to move in the same (+) or opposite (-) directions
- Units = Units of X x Units of Y

Correlation:




Always between -1 and 1
- Strength of linear relationship between X and Y
- 0 = uncorrelated
- 1 = perfect positive correlated
- -1 = perfect negative correlated

Linear regression model:

Ui = errorterm = vertical distance between regression line and observation point

What if we don’t know B1 and B0
 OLS estimator




Bij predicted komt er dus zo’n dakje bovenop
Residuals = u = Yi – Ydakjei

, Linear regression = more explicitly tries to assess a cause and effect relationship, and quantify such
causal effect

Goodness of fit measures:
-


= proportion of sample variance of Yi that is explained by Xi
= always between 0 and 1

- Standard error of the regression




= large SER  predictions very different from actual values In
STATA  SER = Root MSE



College 3:
OLS simple linear regression models assumptions:
1. Zero conditional mean: Xi is uncorrelated with Ui en with other factors that influence Yi



 Most of the times difficult to know if ZCM holds
 ZCM also does not hold when there is simultaneous causality (if Y also affects X)

2. Observations are independent and identically distributed (i.d.d)
 Holds if sample is drawn from simple random sampling
 NOT when = time series data, panel data (more times)

3. Large outliers in X and Y are unlikely



 If it happens, drop it
 Otherwise you get an plausible assumption