Are you not making an error in the calculation of the RR after the nondifferential misclassification (9:30)? If 20% is misclassified, then 0.8*450 = 360 (correct) but 0.8*900 = 720 and not 620. Think it's just a typing error because you are adding the right amount of subjects (180 = 0.2*900) to the exposed non-cases but the RR should be 1.65 and not 1.49.
i learnd by my teatcher that the recall bias is a collecting infos bias not misclassification and as i have understand from her that misclasification bias is when for example : we say that if a patient have got (a)deseas when his blood sugar get to (x percent) a couple of years after we have identify thise percent we find out that there is a (b) deseas that is identified when the blood sugar is also(x percent). means that the bias is when we missclasify the amount of the desease index. and also the other type of bias is when we missclasify the amount of exposure for instance, if at a single year we identify the exposure to be 1 segarite box and in the future we change it into more then 1 segarite box. then all of the people that only use 1 segarete box wouldn't be exposed and our researshe is biased.
You said that there is recall bias in the cases where they are more likely to remember exposures? You mean that there are recall bias in the controlgroup and not in the casegroup?(4:00)