CHI SQUARE TEST

Null hypothesis

• assumes data will fit given ratio or model

  • no difference between what you observed and what you expected

  • any apparent different between observation and the model is attributed purely to chance

• can never be proven

  • your experimental evidence can only “support” or “reject” the null hypothesis

  • “support” is also sometimes worded “fail to reject”

Alternative hypothesis

assumes there is a difference between observed data and model

Chi square statistical test

• also known as “goodness of fit” test

• evaluates influence of chance on data

• assesses whether your data is consistent with the model/hypothesis

  • model = expected outcome

p-value

statistical measure of significant differences between datasets 

• p value greater than α = null hypothesis is supported (fail to reject)

• p value less than α = null hypothesis is rejected (too much variation)

p-value = 0.05 (5%)

means that if you repeated the same experiment, there would be a 5% chance you would observe the same variation or greater in your data

X²

• o = observed

• e = expected

• ∑ = sum