You are on page 1of 12

Quantitative Methods

Introduction

Experimental Data

Non-Experimental Data & Inference

Probabilistic versus Deterministic Models

Political Methodology
Introduction

Experimental Data

Experiments are a set of observations performed to support or


falsify a hypothesis. In order to demonstrate causality, one
generally must show that a phenomenon occurs only in the
presence of a particular causal factor—and that the
phenomenon does not occur in the absence of that causal
factor.
Introduction

Experimental Data

An controlled experiment involves the comparison of results


obtained from an experimental group to those obtained under
the control group. The control group is exactly like the
experimental group except for the manipulation of one
method.
Introduction

Observational Data

A natural experiment or quasi-experiment does not involve


manipulation or a controlled environment.
Introduction

Observational Data

In observational studies, data are gathered and the association


between predictors (independent variables) and the response
phenomenon (dependent variable) are assessed.
Introduction

Observational Data

Descriptive statistics involve summarizing a collection of data.

In inferential statistics, we are generally using a sample. We model patterns


in the data in such a way to account for randomness and uncertainty in the
observations, and then draw inferences about the process or population
being studied.
Introduction

In Inferential Statistics…

In inferential statistics, we may be interested in predicting Y with X, or with


the casual effect of X on Y.

We call the population measure (in these examples, either the mean or the
effect of X on Y) the “parameter”, and the sample measure the “parameter
estimate”.
Introduction

What is a model?

A model is a representation or an abstraction of reality.


Introduction

Deterministic & Probabilistic Models


In deterministic models, if certain conditions are met, the outcome is certain
to happen. There is no “error”.

In probabilistic or stochastic models, if certain conditions are met, the


outcome is more or less likely to happen.

When we are modeling, we are essentially fitting a deterministic model to


actual data.

Click here for a paper by Gelman et al on the two types of models.


Introduction

A Few Other Items…

As noted, in descriptive statistics, we may be interested in


presenting information about the data—such as measures of central
tendency (i.e., means, etc.)

We may also want to take a sample and estimate the effect of one
variable (or a set of variables) on another. In this case, we are
generally using inferential statistics (but contemplate the difference
between a population and a sample, and the meaning of “inference”)
Introduction

A Few Other Items…

Explanatory variables (or independent variables, or “left hand side”


or “LHS” variables, or “covariates” ) are often signified by X.

Dependent variables (or “outcome” or “right hand side” or “RHS”


variables) are generally signified by Y. Yi is a random variable (that
is, we don’t know the value); we know the particular value for lower
case yi.

You might also like