Digital Workshop

Am interested in learning if OPUS can calculate the following from test data:
reliability co-efficient
validity co-efficient
Standard error of measurement

If your question is "does Opus have built-in statistics functions? and specifically the measures mentioned?"
--- the answer is 'no', and 'no'.

However, I think YOU can calculate those measures from test data. You would probably have to learn and use script functions. OpusPro has math commands, which can be made to operate on data sets of varying size (through iteration, looping, etc). For example, there is not a built-in function for returning the Sum of Squares, but you can get there.

A lot depends on your purpose and your vision --- what you want to accomplish, where the data resides, how dynamic the dataset is, and how the data gets into Opus... how you want to display the results (coefficients, etc). And do you need Users to have interaction with the data or control the execution of the calculations. If you are trying to teach or illustrate/animate these principles, Opus gives you a lot to work with. If you are trying to build a business or research application, then you will likely be building some business rules or logic and then spawning some output to other systems.

Back you your original question, try googling
---> javascript calculate "reliability coefficient"
If you see something interesting there, you may find a way to implement in Opus.

Lastly, Opus can generate Pubs built as .EXE executables on Windows PCs --- and those Pubs can access Excel spreadsheet data as well as Access DB records. You are not limited to excel/access, but can use other databases and SQL methods, so perhaps utilize their internal statistical functions.

Could the following website be useful?

http://www.jstat.org

Hopefully yes

Graham

?? No reply from Leticia? Okay, it has not been that long.

Anyway, I came across this tidbit of what I'd call middleware... may be useful for the OP, or for some others looking to expand functionality (punintended). (of course, be conscious of any licensing factors when drawing on others' works/tools).

http://commons.codeforamerica.org/apps/datacouch

I am desigining tests in OPUS in support of e-learning modules and wanted to gather data on the validity and reliabilty of the tests so that we could refine them as needed and or post our ratings as evidence to the reliabilty and validity of our assessments for customers and managment.

Graham - thank you for the jstat lead. It is a great resource. It did not however have the scripts I was hoping to find.

Lar_123 - I also tried googling javascript calculate, but to no avail. What I am looking for is probably best handled in an external application specifically designed for that purpose. OPUS is not set up to do this just yet, maybe sometime in the not too distant future.

Much thanks to you both.

Leticia,
That sounds like a really refined project. There are some folks in the Opus community, 'Mallette' come to mind, who have used Opus in corporate/administrative/?? training. Have not seen him here for weeks, but I suspect he would know what you are attempting.
That's his by-line.

On the search, I used Ask-advanced. Google has similar.
http://www.ask.com/webadvanced
I would try putting..... reliability co-efficient .....in the 2nd line for 'phrase',
...
and the other keywords.... javascript calculate ....in top line.
If you've tried that and the results didn't fulfill, then we'll have to dig deeper.
(as an aside, sometimes one can find an intern who can distill this challenge by integrating with external DB/SQL or Excel, and with Macros + clipboard bring the results into Opus. Server-side coding is another possibility, depending on how you need to handle the datasets and iterative changes.)

Just to mention, and I may have this not-quite-right, Javascript in its traditional syntax can be included/embedded within the HTML of a webpage. So that is one way to reuse JS code you may find. Similarly, Actionscript may be embedded inside Opus' Opusscript, so again possibility to apply code you may find (although with AS, I think the Pub Type may need to be Flex output???). Just some food for thought.

Good luck.

Okay, this topic officially brings back bad memories of analysing the distribution of pebbles on Budleigh Salterton beach. No wonder students drink too much

Anyway, off topic - if you are looking for external software, then we were supposed to use SPSS for this type of statistical work but I recall that we discovered Microsoft also offer an analysis tool kit free of change for Excel which was a lot more friendly - IMHO.

It may (after a few drinks), prove useful!

Mack

Its funny that my query would remind you of Dave Mallete; we work together!
As far as his quote's concerned - he's right