Study Guide: Statistics

Columbia College

Study Guide: Statistics

Material which is critical to your understanding

===========================================

The 4 Types of Data:

Nominal (Categorical) Data:

No greater/lesser than -just categories (no parameters)

Descriptive:

- do a bar graph

- the mode is the central tendency

Inferential:

- test of significance: Chi Square

- test of association: coefficient of contingency

Ordinal Data:

has greater than/lesser than, but still lacks known parameters

Descriptive:

- do a bar graph

- the mode is the central tendency

Inferential:

- test of significance: Chi square

- test of association: coefficient of contingency

Interval Data: has parameters

has greater than/lesser than and there is known distances between

Descriptive:

-do a histogram

- use mean and median on central tendency

-variability: Standard Deviation or Interquartile Range

-test of significance: t test or z test

Inferential:

- test of association: Pearson's Correlation

- test of casuality: Regression

Ratio Data: has known parameters

in addition to interval data, has a known zero point

Descriptive:

-by it's very nature, it compares frequencies of a variable

Inferential:

- test of association: Pearson's Correlation

- test of casuality: Regression

==================================

Greek Letters (Some are Used in Statistics)

NOTE: sometimes we use the lower case of the letter (example: mu), sometimes the upper case (example: Gamma) , sometimes both (example: Sigma)

Infiniti Symbol we will use
--------------------------

When you are dealing with Nominal and Ordinal Data:

Chi Square is the bivariate statistic to use with nonparametric data. It is a test of significance. Once a statistical significance is found, then the Lambda or Gamma statistic proceedure is used to measure the power of the association.

LAMBDA - A measure of association for nominal data

A measure of association for nominal data which indicates the degree to which we can reduce the error in predicting values of one variable from values of another.

---------------------------

When doing a regression:

The main difference between the inferential statistic proceedures of correlation and regression is that you know which variable is the independent variable in a regression but you do not known which variable (if any) is the independent in a correlation. This is a question of causality.

The Independent Variable (the one that comes first in time) will always be placed on the X axis while the Dependent Variable is placed on the Y axis.

----------------------------------------------------------------

Thank You for Visiting!