This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Jamie Julin: ok, so pick the point where the difference is greatest like it says Jamie Julin: even if its just a small difference between the choices roohi: oh so does it make sense to have a difference of .9 roohi: be like the greatest? roohi: does tha sound reasonable Jamie Julin: thats for you to decide roohi: am i on the right track Jamie Julin: if youre using the applet like we just talked about, then yes K. Wei: i have a quick question about q29 ch15 K. Wei: if u have a moment Jamie Julin: go for it K. Wei: i understand that as the number of draws goes up the proximity to the normal curve increases K. Wei: but does percentage of draws/total affect it at all? Jamie Julin: its number of draws that matters, not percent of draws roohi: oh hey jamie si the difference in percentages that we report--or the difference in areas K. Wei: so if you drew 10 draws from 20 possible numbers, it would be less accurate than if u drew 11 draws from 1000 possible numbers? Jamie Julin: percentage is equivalent to area in histograms Jamie Julin: yes, the normal approximation improves based on the # of draws not the percent of draws... even if it seems strange K. Wei: that what i thought at first but when i said that the normal aprox to the distrubtion fo the sample sum of 10 indepdent random draws with replacement from a box of 100 was less accurate than the normal approx to the distrubtion of sample sum of 100 indepdent random draws with replacement from box of 1000 K. Wei: it was marked wrong K. Wei: which is why im confused now Jamie Julin: ok, but is the number of draws the only thing that matters for normal approximations? K. Wei: well skewed ness K. Wei: is what matters as well Jamie Julin: right good Jamie Julin: so do you know anything baout skew here K. Wei: but it doesnt mention skewedness at all K. Wei: so i guess theres no way to determine it Jamie Julin: right K. Wei: so i guess accuracy cannot be determined without knowing if they have same skwewed ness Jamie Julin: yep K. Wei: tricky K. Wei: lol K. Wei: thanks a lot Jamie Julin: no problem James: can i ask a question? janet: hi jamie, for q22, The distribution of the sample sum for a sample of size 10 is (Q22) the distribution of the population janet: oh ok i'll wait Jamie Julin: sure james go ahead Jamie Julin: ill look at q. 22 too though Jamie Julin: ok janet, so did you click the 'take sample' button and can you see the distribution versus the sample janet: oh for me, i'm not sure how to tell whether its less skewed or more skewed James: I have question on Q28,29,30 on problem set 15 janet: does the distribution mean the normal curve Jamie Julin: yes thats what i mean janet: oh ok yeah i did that Jamie Julin: well skewed would be like less normal looking janet: so for q23, when i clicked the take sample button, the bars were all underneath the normal curve, so does that make it more skewed?...
View Full Document