• If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • Files spread between Dropbox, Google Drive, Gmail, Slack, and more? Dokkio, a new product from the PBworks team, integrates and organizes them for you. Try it for free today.


Questions and Comments here:

Page history last edited by bob pruzek 9 years, 1 month ago


Comments (21)

Samira said

at 8:59 pm on Jan 26, 2011

What is meant by "pooled standard deviation" (as in the denominator of standardized effect size)?

bob pruzek said

at 9:09 pm on Jan 26, 2011

This refers to the average of the variances in the two independent groups. This is a simple average if the groups have
the same sizes and a weighted average (wts n1/(n1+n2) and n2/(n1+n2) if the groups have different sizes. More tomorrow. b

bob pruzek said

at 9:15 pm on Jan 26, 2011

Another student asked a question "How to load the package 'psych'?" Ans: Go to the top of the page, w/ R open; click on Menu item that
refers to Load Packages. Once you click, you will see that you have to choose a 'mirror' site (usually best, either eastern US or Canada); do that
and then you'll see list of all packages, alphabetically arranged. Find psych, click on it, and then ask for it to be installed (but be sure to choose
the box 'with dependencies' at the bottom of the list first. I'll be putting up a list soon of Most packages I use regularly, and that will suffice for
this class. I cannot load them all until next week on the SPH computer, howwever, for technical reasons -- but you can, and it should go smoothly
if you have R ver 2.12.1 installed on your machine. b

Jamie Kammer said

at 9:25 pm on Jan 26, 2011

What do you mean when you say in the second assigned reading that you 'jittered' the data?

bob pruzek said

at 9:29 pm on Jan 26, 2011

I'll discuss that in detail tomorrow, but in short, it means to add a small amount of random error FOR PLOTTING points, to help separate values that are on top
of one another w/out jittering. b

Jamie Kammer said

at 10:26 pm on Jan 26, 2011

I'm having a hard time getting the following tow work:
Lowry Gps compared after data ranking; NP method",cex=1.1)
I think I'm running into an issue with the ranking? Any help would be much appreciated!

bob pruzek said

at 1:12 am on Jan 27, 2011

I suspect your problem has to do w/ how you've set up your input data. Try this (replace your data name w/mine).
Then run granova.1w(Lowry.rnk,rep(1:2,ea=15),dosq=F) #the other arguments not essential
Let me know IN CLASS how it worked. b

bob pruzek said

at 1:13 am on Jan 27, 2011

ps. I'm assuming Lowry.A.B.dt is a dataframe that contains the data as two columns.

Chuck Yang said

at 11:37 pm on Jan 28, 2011

Did any one tried the granova.1w function introduced on the second page of the Comparing2IndepGroups.pdf? I replicate the exact code in the document and submit it after loading the library HSAUR, however it keeps generating an error message: "Error: could not find function 'granova.1w'"
Here is the codes I was trying to submit:
granova.1w(weightgain[1:20,3],weightgain[1:20,2],dosq=F,h.rng=1.5,main='Comparing two independent groups, weightgain data, n=10 ea. grp [library:HSAUR]')
Any ideas why?

bob pruzek said

at 11:45 pm on Jan 28, 2011

Sounds like you have not loaded the library granova. Even if you installed it, you will not have loaded it unless you use the library command. See first page of
this wiki for details....always look for details. If you do search( ) you should see granova on the list of what packages are loaded; so check that, and load
if you don't see it. Let us know what happened. bob

Chuck Yang said

at 11:54 pm on Jan 28, 2011

You were right Dr. Pruzek, I did install the granova package but did not successfully load it. I used library(granova) to load the package and now the granova.1w function operates smoothly.

knelder1 said

at 6:24 pm on Jan 30, 2011

Is there a particular folder or page that you would like the homework uploaded to? It does not say on the homework handout. Thank you.

bob pruzek said

at 6:28 pm on Jan 30, 2011

Just put it on the upload page and be sure to use my guidelines for naming the file. I'll put these in folders.
Best, Bob

Richard said

at 8:53 pm on Jan 30, 2011

Is it possible to use the granova.1w function when the two groups being compared are of different sizes? Also, how do you calculate effect sizes?

Jamie Kammer said

at 10:13 pm on Jan 30, 2011

To calculate effect size, use the syntax from the paper:
es.twogps <- function(x, y)
{#generates standardized ES for input
# x,y (indep.) sample data
mx <- mean(x)
my <- mean(y)
nx <- length(x)
#nb: nx will generally NOT equal ny
ny <- length(y)
md <- mx - my
vpoold <-((nx-1)*var(x) +(ny-1)*
var(y))/ (nx + ny - 2)
sdpoold <- sqrt(vpoold)
ES <- md/sdpoold
list(md=md,ES=ES) }
#but in place of lowry.a.b.dt above put the name of you data fram

bob pruzek said

at 11:46 pm on Jan 30, 2011

Reply to Richard: Yes, in fact group sizes can be different for either the 1w or 2w functions (2w soon).
Read the help file for the function and you will see that if the groups have different sizes, then the response data
must be entered in a vector, with the grouping variable (indices for groups) defined as gp.
Note to all: If you are to try the ranking idea, it is the vector of responses that is ranked (even if groups have equal sizes);
don't rank within each group, that will give you nonsensical results. ES = (mean difference)/sqrt(average of variances), except
you need a weighted average of the two variances if their sizes are different (see the syntax).

bob pruzek said

at 11:53 pm on Jan 30, 2011

ps. I placed a new file, AnovaDocumentaiton-1w.pdf, in the Upload Files section that some of you may find helpful. There is
little there that cannot be found in the 'Elemental Graphics' paper (see first item under Most Recent, page 1 of wiki), and that
paper carries far more information about granova generally, but both of these might be of some help.

Manchun Chang said

at 9:19 am on Jan 31, 2011

Can we end each class on time at 2:20pm? I need to use the restroom before next class starts.

Richard said

at 4:45 pm on Jan 31, 2011

I am unable to upload files to the wiki. When I go to the files page, the 'upload files' button is gray and cannot be clicked. Also, I have already requested and received access to the workspace and am logged in, but there is still a message that says 'to edit this page, request access to the workspace'.

bob pruzek said

at 8:54 pm on Jan 31, 2011

Richard, There's something strange going on here re: your email address. As I said to you in class the other day, the system says you have never
accessed the wiki, but I know you have tried; moreover, I have received the requests for access you've sent (I think three) even AFTER I'd put your email
in the system. I'll see if there's anything I can do tonight, but in the meantime email me your homework and don't worry about that at least. More
if I get further information. I'm going to email this to you using the wiki copy of your email address just to see if there's a strange bit in that that is
messing things up... so let me know if by 9:15 or so you don't have that email. Thanks, bob

bob pruzek said

at 9:31 pm on Jan 31, 2011

Richard, There was indeed a problem w/ your email address, and while I cannot say why what I just did seemed to have fixed it, the
fact is that my information says the email I entered to replace your old one now is fine. See if you cannot now access the Upload Files page.
(I am putting this here just in case others might encounter problems of this sort... get in touch w/ me if you do). BP

You don't have permission to comment on this page.