Chi square test

For categorical data

If experiment on sample gives a different distribution, does the experiment conclude (reject null hypothesis)?

Chi square = Sum of ([Observed(sample) value frequency - Expected (population) value frequency] / expected frequency)

If between lower and upper chi square value, accept null hypothesis, experiment brought no difference.

Covariance

For many variables, i.e. x and y, to determine relationship (if x increases as y increases, or if x decreases as y increases or if there is no relationship), use covariance.

• Positive covariance means x increases as y increases.
• Negative cov means x decreases asy increases
• Zero covariance means no relationship.

As cov increases it shows higher relationship.

Pearson's correlation coeffecient

Restricts cov between (-1, 1)

Sigma x and sigma y are standard deviation.