Chi - Squared Tests Flashcards
What does chi-squared test compare
How well the observed results match the expected results
What does chi squared test allow us to do
Accept or Reject the null hypothesis and give a percentage of how sure we are that the differences seen are/are not significant.
What does the null hypothesis generally assume
there is no significant difference between E and O, and any differences seen are just down to chance
What is the formula for the chi - squared test
X² = ∑ (O-E)²/E
What does each letter stand for in chi-squared test formula
∑ = Sum (doesn’t need to be typed into calculator)
O = Observed
E = Expected
X² = Chi Squared (Do not need to square root because we are looking for chi squared not chi)
How do you find the expected ratio
Cross the two organisms using a punnet square and get the ratio from the offspring
What are the observed results
The actual results that occurred
How do you calculate the expected results
Add all observed phenotypes to find total number of observed, add both numbers in the ratio (e.g. 3:1 would be 3+1 = 4), divide total observed number by ratio added up, times the outcome by each number in your ratio (e.g. 214 x 3) to find the expected amount for each phenotype
What do you do once you have calculated observed and expected results
Put values into chi squared formula (do a different equation for each phenotype and then add together answers at the end)
What happens once you’ve got your result
A distribution table showing the range of values is used to determine how significant the value is
How do you know which row on the degrees of freedom column, in the distribution table you should be looking across to compare your chi-squared test result to
df = (number of catergories - 1)
In this case we are testing 2 phentoypes so it’s (2-1) = 1. So we are looking across row 1
Where on the distribution table indicates the differences are due to chance and you cannot reject the null hypothesis
A probability of 0.05 or higher indicates differences are due to chance and are not significant
What is the relationship between the x value and the likelihood results are due to chance
The larger the value of x² the more likely the results are not due to chance
