Political Science Methodology Equations


 

Additive Law

 
 
 

Bayes' Theorem

 
 
 

Beta Distribution

 
•Values between 0 and 1 •Percentages are often beta distributed
 
 
 

Binomial Distribution

 
•Discrete Random Variable •Fixed number of trials •2 outcomes in each trial •p=Probability of success •q=Probability of failure •trials are independent (p+q don’t change from trial to trial) •random variable y = number of successes example is p of getting 5 questions on 8 question exam by guessing- 5 choices y=5 n=8 p=1/5 q=4/5
 
 
 
 

cdf

 
 
 

Chi Squared Distribution

 
Special type of gamma distribution
 
 

Conditional Probability

 
 
 

Covariance

 
 
 

Derivatives

 
rate of change of a function = slope
 
 

Distributive Laws of Probability

 
 
 

Gamma Distribution

 
•Continuous Random Distribution •Used for non-negative events skewed to the right •Parameters are alpha and beta •Alpha is the scale beta is the shape
 
 
 
 
 

Geometric Distribution

 
•Discrete Random Variable •Independent Trials •Asks when the first success (or failure) will occur •Example- guy has 1% chance of winning election p=.01 q=.99
 
 
 
 
 

Hypergeometric Distribution

 
•Discrete Random Variable •proportion changes as we sample •example- 20 phds – pick 10 what is p of picking 5 best N=20 n=10 r=5 y is subset of r that you want – in this case 5
 
 

Integration

 
Inverse of derivative
 
 

Linear Regression

 
 
 
 
 

Multiplicative Law

 
 
 
 

nCr

 
 
 
 

Negative Binomial Distribution

 
•Discrete Random Variable •Know xth success happens on nth trial •r = event of interest •example – p of stringing oil on 3rd oil strike on 5th try (p of oil is .2) y=5 r=3 p=.2 q=.8
 
 
 
 

nPr

 
 
 

pdf

 
 
 

Pearson's r

 
 
 
 
 
 

Poisson Distribution

 
•Discrete Random Variable •interested in counts- number of times something happens •The number of occurances in any two subintervals must be independent •Probability of an occurance in any short time interval must be approximately proportional to the length of time interval •Subintervals where only one event can happen •Lambda = number of trials, x probability of success •Example number of bills reported in a session (25 is the historical average) reported by a committee in congress- what is probability of them reporting only 10? y=10 lambda=25
 
 
 
 

Rules for Independence

 
 
 

Spearman Rho

 
 
 

Standard Deviation

 
 
 
 
 
 

Standard Error of OLS Reg. Estimate

 
 
 
 

Uniform Distribution

 
•Continuous Random Distribution •For any point, the probability distribution is the same •All points are equally likely •It is flat – uninformative •theta 1 and 2 are parameters for distribution •Example- bills reported sometime over 10 hours – what is p of between hours 9 and 10? Theta1 = 0 theta2 = 10… integrate between 9 and 10 dx=(1/10)y so (y/10)-(y/9)=1/10 chance of bills reported in between 9 and 10
 
 

Variance

 
 
 
 
 

Variance of OLS Regression Error

 
 
 
 

Z-Score