15- Association and Correlation I

Question 1

What is the chi-squared test of independence?

Answer 1

a non-parametric test based upon an x^2 distribution which can be used to find p-values

Question 2

Purpose of the chi-squared test of independence?

Answer 2

Used to determine whether there is a significant association between two categorial variables

Question 3

What is chi-squared distribution?

Answer 3

begins at zero and is positively skewed (no negative values), mean and variance are both equal to the degrees of freedom and then shape of the distribution depends upon the number of degrees of freedom

Question 4

Contingency Table

Answer 4

displays frequency distribution of two or more categorial variables

Question 5

Degrees of Freedom

Answer 5

related to the amount of information in the table and is important for calculating the p-value

Question 6

Assumptions which come with chi-squared test of independence

Answer 6

1) categorical data with observed counts, two or more levels per variable 2) subjects are counted in one and only one cell 3) No category should have an expected frequency less than 1 and not more than one category in five total should have an expected frequency less than 5

Question 7

Properties of the chi-square test of independence?

Answer 7

1) Tells us about associations between two categorical variables 2) no association = null hypothesis 3) test statistic can be used to find p-values 4) shows significance of association between two categorical variables but not the strength of association

Question 8

What is Cramer’s V?

Answer 8

A test which measures the strength and magnitude between two categorical variables that have been found to be dependent using the chi-squared test

Question 9

Result of Cramer’s V?

Answer 9

number tells the size measure for assessing the practical significance of the chi-squared test of association

Question 10

Bivariate Relationships

Answer 10

Can be positive, negative or neutral

Question 11

Association strength in Bivariate Relationships

Answer 11

1) Perfect Association- all points are PERFECTLY linear 2) Strong Association- points are NEAR linear 3) Weak Association- points are somewhat linear 4) No Association- points are randomly distributed

Question 12

Benefits of Bivariate Relationships

Answer 12

1) Visual (scatterplots, bivariate maps) gives us subjective and qualitative insight 2) Correlation provides us with a quantitive, objective measures of the nature of relationships between variables