[Rasch] Questions about interpreting wright/item map for rating scale model

Rebecca L Freund rlfreund at berkeley.edu
Tue Jul 21 23:30:13 AEST 2020


Grace,

The lines are connecting different parameters from the same item. The dots
indicate the parameter locations. It looks like your first item has step
parameters around 1 and 4, your second item’s parameters are around -4 and
-1, etc. The line and shapes are simply to help distinguish the items. The
last two items have extra dots because those items have more score
categories than the first 17 items.

David Torres Irribarra and I have written a package for R to create this
type of map, with a lot of customization options. It is available on CRAN
as WrightMap, or you can download the most recent development build at
https://github.com/david-ti/wrightmap. More information is available at our
website, WrightMap.org, or feel free to reach out if you have any questions.

Rebecca


On July 21, 2020 at 4:33:52 AM, Yingchen (2596011329 at qq.com) wrote:

Dear Rasch lovers:
    Can someone tell me how to interprete the wright/item map for the
rating scale model? I attached the graph  from jMetrik.
    I read user's manual for jMetrik, some publications and tutorials on
item map and checked the estimated item values. I could understand the map
with histogram on the left and dots on the right, but not this graph. I
checked and thought the graph was produced correctly.
   Would appreciate if anyone can help with the following questions.
  *Q1:* What do the lines mean? They must represent the item parameter
estimates. However, I cannot align them with the item file.
  *Q2: *What do the different shapes on the lines represent? There are
squares, diamonds,circles, rectangles, arrows, arrows pointing to different
directions. My guess is that the software developers used different shapes
to enable readers to tell the different items.
 * Q3: *What do the square and the circle in the middle of the last two
lines represent?  I know that they should have something to do with the
item parameter estimates.
 * Q4:* is the length of each line related to confidence interval of item
parameter estimate?
  *Q5*. Can anyone recommend non-commerical softwares with the capacity to
produce the the item map?

  BTW, I did send a message to jMetrik help 2 days ago, but there was no
reply.
  Any help is greatly appreciated.
  Thank you for your time and attention.

Grace
________________________________________
Rasch mailing list
email: Rasch at acer.edu.au
web:
https://mailinglist.acer.edu.au/mailman/options/rasch/rlfreund%40berkeley.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20200721/1cffd658/attachment.html>


More information about the Rasch mailing list