Hi 2001,
Solutions to problem set 1 are posted to the course website. We'll hand
back your hardcopies in section on Thursday, and for this problem set I'll
email a brief note and grade to those of you who submitted the entire
problem set electronically.
A couple of notes:
Please use the solutions on the website as a guide for future problem sets.
In particular, note how there are two files, one with R code (run-able; it's
a file we can open and run as-is in R) and one with solutions and
explanations. Submitting a separate R file is important so we can easily
run your code and see where things went awry (or well, of course).
If you use R to calculate something (such as a probability), include the
actual value that R estimates in your write-up; don't just include the code
without its output.
Many of you had clever approaches to the birthday problem- good work. In
the solutions we show two ways of estimating the room sizes in problem 2.
Even if you generated all correct answers, take a look at the solution
code. All of you can benefit from making your code more efficient.
Common mistakes on this problem set include 1d, 2a, 2b and 3d. In 1d, a
common error resulted from forgetting which gender corresponded to 1's and
which to 0's. In both parts of problem 2, a common difficulty was figuring
out how to find triplicate birthdays without counting two pairs of
birthdays. (We discuss this problem in the R file to the solutions).
Another mistake on 2 involved testing a few room sizes without resetting the
counter for 'samedays' so the estimate for the number of required people was
too low. In 3d, some struggled with drawing simulated datasets- you can
review the solutions and also the R code from section 2 for a slick way to
sample from a dataset.
And finally, don't worry if your solutions were not exactly those posted on
the solutions- these are all simulated estimates, so some variance is
expected and perfectly fine.
One problem set down...
Jenn