ECONOMETRIC

METHODS

Generalized linear models  

Logistics regression (or probit)

• For binary Y variables

• Odds ratio of female voting for Obama

• For count data (integers)

• For categorical data

• For ordinal data (ranks)

Panel Data Analysis Methods

Entity FE

(when time constant)

 If you need to control for omitted variables that differ among panels, but are constant over time

Here, earning (wage) obviously is influenced by factors other than experience (tenure), such as personality of the person, which can be assumed to stay constant over time

Time FE

(when panel constant)

Random Effects Model

(random variability
among time and panels)

Similar to taking the mean of each variable in the model for each panel across time and running a regression on the collapsed data set of means

What is the effect of serving in the Vietnam War (D) on subsequent civilian mortality (Y)?  Choose the Instrumental Variable (Z) to be draft status.

Difference-in-Differences

Miscellaneous methods

Three Types of data

1. Cross sectional data
• One dimensional data: (Obs v Variables)
• Random sample; if not random then selection bias
• Each observation is a new individual, firm, etc, with information at a point in time
• Ex: household survey from year 2012
2. Panel data
• Multi-dimensional data: (Obs v Variables v Time)
• Can pool random cross sections, and then treat similar to normal cross-section
• Need to account for the time difference
• Ex: compiled household surveys from 2000 - 2012
• In Biostats - often called Longitudinal data (here, a patient is considered a panel)
3. Time series data
• Each observation is from a time (Time=Obs v Variable)
• Has a separate observation for each time period (eg. stock prices)
• Not random sample
• Trends and seasonality become important 

Types of measurements

• Transaction data (price, interest rates, exchange rates)
• Aggregate data (GDP, employment, unemployment, price index, money, etc.)
• Survey data (qualitative, quantitative)
• Data revisions (real time v historical measurements)

Data Mining Process

1. Anomaly detection
• (Outlier/change/deviation detection)
• The identification of unusual data records, that might be interesting or data errors and require further investigation.
2. Association rule learning
• (Dependency modeling) –
• Searches for relationships between variables.
• Ex: Market basket analysis: For example a supermarket might gather data on customer purchasing habits. Using association rule learning, the supermarket can determine which products are frequently bought together and use this information for marketing purposes.
3. Clustering
• Method of discovering groups and structures in the data that are in some way or another "similar", without using known structures in the data.
4. Classification
• is the task of generalizing known structure to apply to new data
• For example, an e-mail program might attempt to classify an e-mail as "legitimate" or as "spam"
5. Regression
• Attempts to find a function which models the data with the least error.
6. Summarization
• Providing a more compact representation of the data set, including visualization and report generation.

Glossary

Dummy Variables

• A categorical variable can be broken down into multiple variables, each with only 2 values 0 and 1 called dummy variables

• STATA command:    Xi

xi expands terms containing categorical variables into indicator (also called dummy) variable sets by creating new variables and, in the second syntax (xi: any_stata_command), executes the specified command with the expanded terms. The dummy variables created are

i.varname creates dummies for categorical variable varname

i.varname1*i.varname2 creates dummies for categorical variables varname1 and varname2: all interactions and main effects

i.varname1*varname3 creates dummies for categorical variable varname1 and continuous variable varname3: all interactions and main effects

i.varname1|varname3 creates dummies for categorical variable varname1 and continuous variable varname3: all interactions and main effect of varname3, but no main effect of varname1