6.1/6.2 - Goodness of fit/Degrees of freedom and χ^2 family of distributions

define goodness of fit

how well an observed frequency distribution fits to a known distribution.

what would the hypothesis for testing distribution

what is the formula for goodness of fit (X²)?

if (when testing a distribution) a parameter is unknown, how would you find it?

use the data from the observed frequency to estimate it.

when would you combine a cell in a table?

if one of the expected frequencies is <5

in general what is the number of degrees of freedom in a test?

the number of cells (after any combination) - the number of constraints

what are the constraints that a goodness of fit test could have?

• the observed/expected frequencies must add to the same value (this is always true)

• whether you had to estimate a parameter or not

how would you use degrees of freedom

when choosing which of the chi-squared family you want to use.

when would you reject the bull hypothesis in a Chi-Squared test

when the χ² value from the tables is less then the calculated value of χ².