This web page is for error corrections and other changes to the slides and homeworks.
- (posted May 13). Homework 9, problem 9-6 solution was obviously wrong in that the calculation use n = 20 rather than the n = 10 the problem stated. It has been fixed on the homework solutions.
- (posted May 6). In the solution of question 8 of the 2012 final exam the second to last displayed equation should say g'(μ)² σ² = (− λ² ⁄ α)² ⋅ α ⁄ λ² = … (the parentheses and exponent on the first factor were missing).
- (posted May 6). In the solution of question 7 of the 2012 final exam the second displayed equation should say var(X) = σ2 var(Z) = … (not var(X) in both places). In the third displayed equation the middle term should be 5 σ2 ⁄ 3 n (the n was missing).
- (posted May 6). In the last line of the last displayed equation of the solution of question 4 of the 2012 final exam the expression should be exp(− n n θ − n n ⁄ θ) (one of the n's was missing).
-
(posted May 6). In the third displayed equation
of the solution of question 1 of the 2012 final exam
the PDF should be θ (θ + 1) (1 - x)
xθ − 1 (not
xθ − 2).
Later on in this same question
it should say
Case I (5101, deck 6, slide 9)
(not deck 4). Later on in this same question it should sayCase II (5101, deck 6, slide 24)
(not deck 4). -
(posted May 6). At the end of the solution of question 1 of the
2010 final exam it should say
Case I (5101, deck 6, slide 9)
(not deck 4). - (posted May 6). The comment about question 2 of the 2010 final exam has been moved from before the question to after it.
- (posted May 6). The bottom equation on slide 83, deck 6 was wrong (not even an equation). The right-hand side was correct, but the left hand side should have been r c − 1 − (r − 1) − (c − 1) and the middle term should be changed accordingly.
- (posted Apr 28). In linear models and dummy variables computer examples web page a new section about comparing possible models for the different intercepts but same slope example has been added.
-
(posted Apr 28). In the text on on slide 22, deck 6,
changed from
solid dots the MLE predicted values
tosolid dots the MLE mean values μi
with a hat over the μ. -
(posted Apr 28). In the text between the last two displayed equations
on slide 19, deck 6,
added
for β
afterFisher information
. -
(posted Apr 28). In the first paragraph on slide 13, deck 5
a
hat
was put on the μ in the phraseμ is the projection of y on V
. - (posted Apr 5). In the displayed equation on slide 89, deck 4 In(θ) should be I1(θ). Fixed.
- (posted Apr 3). In first displayed equation on slide 33, deck 4 the term exp(−λ n) should be exp(−λn n). Fixed.
-
(posted Apr 3). In the last line of slide 32, deck 4
log likelihood
should be justlikelihood
. Fixed. - (posted Mar 28). Multiple changes to slide 32, deck 4. Before, both sample sizes were the same (n in the second displayed equation and the right-hand side was wrong — the exponent of λ should have been 2 n rather than n). Changed to having different sample sizes (m and n).
- (posted Mar 25). In homework problem 7-5 the greater than or equal sign in the alternative hypothesis should be strictly greater than.
-
(posted Mar 12).
Slide 93, deck 3, deleted
of
. Slide 98, deck 3, deleted first occurrence ofsquare roots of
insquare roots of inverses of square roots of
. Slide 104, deck 3, deleted firstthe
in first line. - (posted Mar 11). On deck 3, slide 61 the reference to slide 45 has been changed to slide 43 (which is more informative).
- (posted Mar 5). The following addendum has been added to homework problem 5-2. (In order for this to make sense, we must assume both populations have means and the distribution of differences is symmetric. For the t test we also have to assume both populations are normal.)
- (posted Feb 18). Slide 109, deck 2, while not strictly wrong, was confusing and has been changed: the χ21 − α ⁄ 2 and the χ2α ⁄ 2 now occur in that order so they match the following Rweb. That is, χ21 − α ⁄ 2 = 2.700389 and χ2α ⁄ 2 = 19.02277.
-
(posted Jan 30). Slide 69, deck 1 had a typo; the last sentence
said
orthornormal matrices
instead oforthogonal matrices
. Now fixed.