<p>chi-square:</p>
<p>It’s about drosophila lab right? So we are basically checking weather the outcome result of the lab “fits” well into the expected results.</p>
<p>The chi-square equation is : Sum of (Observed-Expected)^2 / Expected</p>
<p>degree of freedom is K-1, which is column - 1.</p>
<p>So let’s suppose we produced our f2 generation. We crossed Bb x Bb 1000 times, which B is black and b is brown. The actual outcome was black: 800, brown: 200. at the 10% significance level, does this outcome support our assumptions?</p>
<p>H0: chi-square fits well yeah!
Ha: It doesn’t fit well :(</p>
<p>possible genotypes of Black are BB, Bb. If we make monohybrid cross between above two traits, we should get BB, Bb, Bb, bb. Since there are three blacks, (BB, Bb, Bb), the probability of getting black is 3/4 = 0.75, and probability of getting brown is 1/4 = .25</p>
<p>The expected values will be then multiplying those probabilities with the N, which is 1000. Then we should’ve got (according to the probability) 750 blacks and 250 bronws.</p>
<p>Black: observed-800///////expected-750 (800-750)^2/(750)
brown: observed-200//////expected-250 (200-250)^2/(250)</p>
<p>add those two results up, then we get 10+3.3333 = 10.3333 as a chi-squre test statistics.</p>
<p>Degrees of freedom here is (black and brown = 2) - 1, so it should be 1.</p>
<p>AP biology test makers would give you a chart that shows corresponding chi-square values for certain signifcance level. Then, just compare the obtained chi-square result with the one that shows up on the chart. If the obtained result is bigger, that means there is lower probability of rejecting H0, so we say “at the 10% significance level, there is not enough evidence to reject H0, blah blah”</p>
<p>But if the obtained result (13.333) is smaller, then there is higher probability to “reject H0”. Meaning that, we say “At 10% significance level, there is enough evidence to reject Ho, blah blah”</p>
<p>-Thank God I took ap stat course.</p>
<p>If i made any mistake, please correct it :)</p>