| Back |
- Chi-squared measures whether difference between expected and observed results is signifficantly different from chance variation alone, tests null hypothesis (no sig diff between expected/observed results, so down to chance)
- make table with phenotypes|observed|expected, state phenotypes of offspring in column, write NUMBER of offspring observed of each phenotype next to it, then carry out genetic cross of parents and state ratio of phenotypes expected, then multiply each ratio part as fraction by total number of offspring for expected column (next to each phenotype), then use x2=sum of (O-E)2/E (O is observed and E is expected, degree of freddom (n-1) used and n is no of phenotypes present, if x2<CV (in table) at p=0.05, accept H0 (no sig diff between O and E), and if x2>CV at p=0.05, reject H0 (sig diff)
|