This preview shows pages 1–4. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 6.2 Tests of Significance Example #1 borrowers at private 4year college: mean debt ( a )= $21,200 (survey result) borrowers at public 4year college: mean debt ( b )= $17,100 (survey result) the difference $4100 ( a b ) is fairly large/ but these numbers are estimates of the true means Can we conclude from these data that the private 4year students have greater debt than public fouryear borrowers (that the two populations means are different)? Could ask: What is the probability of obtaining a difference as large or larger than the observed $4100 assuming, in fact, there is no difference in the true means? probability big→ our result is not rare when the population means are the same→ do not conclude that the population means are different probability small→ our result is rare when the population means are the same→ do conclude that the population means are different computed prob is 0.17 (do later in book)/ this prob is not particularly small (0.05 standard) Conclude: The data do not provide evidence for us to conclude that the mean debts for private fouryear and public fouryear borrowers are different. There is no evidence to question the possibility that the true difference is zero. Example #2 1997 mean debt for undergraduate study ( c ) was $11,400 (survey result) 2002 mean debt for undergraduate study ( d ) was $18,900 (survey result) the difference $7500 is fairly large ( d c )/ but these numbers are estimates of the true means Can we conclude from these data that there is an increase in borrowing over this period (that the two population means are different)? Could ask: What is the probability of obtaining a difference as large or larger than the observed $7500 assuming, in fact, there is no difference in the true means? computed probability is 0.00004 (do later in book)/ this probability is very small (0.05 standard) probability big→ our result is not rare when the population means are the same→ do not conclude that the population means are different probability small→ our result is rare when the population means are the same→ do conclude that the population means are different Conclude: The data do provide evidence for us to conclude that there is increased borrowing over this period. The evidence suggests that the assumption that underlies the calculation (no difference in mean debt) is not true. 2 Stating hypotheses Null hypothesis H 0 The statement being tested in a test of significance is called the null hypothesis . The test of significance is designed to assess the strength of the evidence against the null hypothesis. Usually the null hypothesis is a statement of “no effect” or “no difference.” Example #1 H : there is no difference in the true means we try to find evidence against the null hypothesis Alternative hypothesis H a The alternative hypothesis is the statement we hope or suspect is true instead of H ....
View
Full
Document
This note was uploaded on 02/22/2012 for the course PAM 2100 taught by Professor Abdus,s. during the Fall '08 term at Cornell.
 Fall '08
 ABDUS,S.

Click to edit the document details