One more PS7 question. Has anyone figured out how to get WhatIf to work for Section 3?
I followed the instructions in various sites on the web and put this in R
install.packages ("WhatIf", dependencies = TRUE)
install.packages("lpSolve")
It seems to load both packages, but when I try a command like
whatif.result <- whatif (data= bioChem1, cfact=bioChem2)
summary (whatif.result)
It says it doesn't recognize a function called whatif even though that seems to be how one is supposed to call this function.
Any luck? What else do I have to do to load whatif.
Many thanks.
Best.
Tom
Hi,
Please read:
(1) Matching for Causal Inference Without Balance Checking (
http://gking.harvard.edu/files/cem.pdf)
(2) Misunderstandings among Experimentalists and Observationalists about
Causal Inference (http://gking.harvard.edu/files/matchse.pdf)
Thanks,
Miya
---
Miya Woolfalk
Ph.D. Student
Harvard University
Government and Social Policy
Hi Abigail,
I am forwarding your question to the class list.
See my comments below.
Best,
Miya
On Thu, Apr 2, 2009 at 9:58 AM, Allen, Abigail <aallen at hbs.edu> wrote:
> Hi Miya/Patrick-
>
>
>
> Can you provide some additional guidance on what our deliverable for this
> replication of others results is.
>
The purpose of the assignment is to help the other group with their
project. The basic guide for the content of your memo should be that you
evaluate the quality of the replication (e.g. how successful was the group
at replicating the original results? does the "replication package" the
group provided enable you to easily replicate the article's/group's
results?) and provide feedback on the kinds of things that you hope the
reviewers of your project will be giving you. So, yes, this includes
comments and suggestions related to both the r-code (e.g. is it clear and
easy to follow? is it correct?) and extension ideas. Comments and
suggestions should certainly cover methodological issues and can obviously
include substantive issues related to the theories/arguments/etc advanced in
the original article.
> Specifically are you looking for us to provide a feedback report with
> comments/suggestions related both to the r-code and extension ideas?
>
Yes.
> Or, are we actually supposed to turn in a report which has a reproduction
> of replicated result (i.e. are we supposed to write our own rcode which
> replicates the results they produced) + our suggestions/comments.
>
It is not necessary to do this.
>
>
> I assumed the former but just wanted to be sure. J
>
>
>
> *Abigail Allen*
>
> *Harvard Business School*
>
> Doctoral Student
>
> Accounting and Control
>
> aallen at hbs.edu
>
>
>
--
Miya Woolfalk
Ph.D. Student
Harvard University
Government and Social Policy
hey, when we are replicating other people's study, do we have to include the
tables and charts and everything. Or, is it enough to say we ran the R code,
we replicated all their results, now here are our suggestions...
thanks!