# EC Thesis: Class Notes

Examination of methods of analysis commonly used in economics and business. Emphasis on non-experimental and quasi-experimental designs necessitating the use of models. and large sample methods, case studies, surveys, regression and forecasting.

## Hw do you know?

### How do you know something?

• tenacity
• authority
• intution
• science

Scientific Method - observe, develop a theory to explain, test - use new data to cross-check

## Key terms

### Variables

1. Random (stochastic, non-deterministic) - value not known at start
2. Non-random (non-stochastic, deterministic) - value is known

Probability - the chance that a random variable will take a certain value
- (distribution must add to 100%)
Probalilty Distribution - set of all possible values of a random variables with the probabilities for each

Dependent, independent, discrete, continuous and limited in a range or not

### Distribution

Continuous distribution (points on a continuous scale)
vs.
Discrete distribution
(clumps, not all values are possible)

Subjective Distribution - have someidea of probalities of outcomes
Objective Distribution - based on # of times an outcome occurs divided by the total number of outcomes

Correlation- when two+ variables show a systematic patrern of movement
Causation - when 1 variable actually causes the other variable to change

* Correlation DOES NOT imply causation.
* Canusaion implies correlation.

Regression vs. Causaiton -  a significant sign on a regression coefficient DOES NOT imply causation.

## Manipulation

### 3 Types of Data

1. Time Series Data - same set of variables in different time periods
2. Cross Sectional Data - same set of variables at a point of time over different cross-sections
3. Pooled Data - cross-section dataset over different time periods

### Some Sample Statistics

Mean - add all, divide by total #

Median - sum average/middle value of mean and mode

Mode - max probality, most frequent value of variable

Standard deviation - average deviation from the mean

CV (coefficient of variation) - in relation tp mean it ttells how signigicant/how much variance)

Measure of Associaion - relationship between variables

1. positive, negative, none

2. covariance (cov) and correlation (p) - makes relative to variance
*Sometimes it is heapful to figure the prooduct of deviations and the sum of squared deviations

Regression - process of finding line or curve that "best" fits a given set of data points (F. Galton, 1886)

OLS linear regression - smalles sumof sqaure error for all points from the line/curve

- good for estimating what comes next
- good at average

Multivariate - plane of best fit (rather than a line)

## Models

1. structural or reduced forms
2. robustness or specification tests
3. intuition or hedonics