This preview shows page 1. Sign up to view the full content.
Unformatted text preview: 10/13/10 SAT Scores Hypothesis Tes0ng Par0al Review 200 300 400 500 400 500 600
3
2
1 0 1 2 3 Students in general SAT Scores of Facebook Users Reality A • The distribu0on of scores for Facebook users is the same as the distribu0on of scores for students in general Adopt parameters of the general popula0on for the Facebook scores 200 300 400 500 400 500 600
3
2
1 0 1 2 3 NULL HYPOTHESIS H0: µFacebook = 500 1 10/13/10 Reality B • The distribu0on of scores for Facebook users is diﬀerent than the distribu0on of scores for students in general Reality B • The distribu0on of scores for Facebook users is diﬀerent than the distribu0on of scores for students in general 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 Reality B Reality B • The distribu0on of scores for Facebook users is diﬀerent than the distribu0on of scores for students in general • The distribu0on of scores for Facebook users is diﬀerent than the distribu0on of scores for students in general 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 2 10/13/10 Reality B • The distribu0on of scores for Facebook users is diﬀerent than the distribu0on of scores for students in general Reality B • The distribu0on of scores for Facebook users is diﬀerent than the distribu0on of scores for students in general 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 ALTERNATIVE HYPOTHESIS H1: µFacebook ≠ 500 Below Only Above Only • The distribu0on of scores for Facebook users is diﬀerent than the distribu0on of scores for students in general • The distribu0on of scores for Facebook users is diﬀerent than the distribu0on of scores for students in general 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 H0: µFacebook ≥ 500 H1: µFacebook < 500 H0: µFacebook ≤ 500 H1: µFacebook > 500 3 10/13/10 We know the general popula0on But we don’t know the Facebook popula0on — That’s the ques0on Hypothesis Tes0ng 200 300 400 500 400 500 600
3
2
1 0 1 2 3 So we take a sample And use it as a best guess for the Facebook popula0on • We use the sample to decide between H0 and H1 Does this sample support H0 or H1? So we take a sample And use it as a best guess for the Facebook popula0on • We use the sample to decide between H0 and H1 Does this sample support H0 or H1? 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 4 10/13/10 So we take a sample And use it as a best guess for the Facebook popula0on • We use the sample to decide between H0 and H1 Does this sample support H0 or H1? So we take a sample And use it as a best guess for the Facebook popula0on • We use the sample to decide between H0 and H1 Does this sample support H0 or H1? 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 So we take a sample So we take a sample And use it as a best guess for the Facebook popula0on • We use the sample to decide between H0 and H1 Does this sample support H0 or H1? And use it as a best guess for the Facebook popula0on • We use the sample to decide between H0 and H1 Does this sample support H0 or H1? 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 5 10/13/10 So we take a sample And use it as a best guess for the Facebook popula0on • We use the sample to decide between H0 and H1 So we take a sample And use it as a best guess for the Facebook popula0on • We use the sample to decide between H0 and H1 Does this sample support H0 or H1? Does this sample support H0 or H1? 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 Deciding We use the null hypothesis as a star0ng point because we have a good understanding of it (of scores in the regular popula0on) Deciding If the sample has a high probability of coming from a distribu0on just like the regular popula0on, then we say it probably does come from such a popula0on H0 is supported 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 6 10/13/10 Deciding Of course If the sample has a low probability of coming from a distribu0on just like the regular popula0on, then we say it probably does not come from such a popula0on H0 is not supported If we’re talking about the probability of drawing samples means, we use sampling distribu0ons of the mean (for a speciﬁc n) to compute probabili0es. 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 Of course Of course If we’re talking about the probability of drawing samples means, we use sampling distribu0ons of the mean (for a speciﬁc n) to compute probabili0es. If we’re talking about the probability of drawing samples means, we use sampling distribu0ons of the mean (for a speciﬁc n) to compute probabili0es. n n 200 300 400 500 400 500 600 200 300 400 500 400 500 600
3
2
1 0 1 2 3
3
2
1 0 1 2 3 7 ...
View
Full
Document
This note was uploaded on 08/22/2011 for the course PSY 207 taught by Professor Pfordesher during the Fall '07 term at SUNY Buffalo.
 Fall '07
 Pfordesher

Click to edit the document details