Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Eveline Gebhardt

Pages: [1] 2 3 ... 7
1
Questions and Answers / Re: Extreme cases and WLE
« on: November 14, 2016, 01:27:46 AM »
Hi Tong

The default for zero and perfect scores is 0.3 and max-0.3 when using maximum likelihood estimates (this is default but can be changed). WLEs are MLEs correct for the bias at the extreme ends of the scales. You can find more about Warm's correction on the Rasch site: http://www.rasch.org/rmt/rmt231d.htm

Cheers
Eveline

2
Questions and Answers / Re: about empirical standard errors
« on: November 03, 2016, 12:40:08 AM »
Hi Tong

I'm sorry, I'm not sure why it happens. I think the problem is in the highest step parameter (the one that is constraint). Your test is short and sample small, which can lead to estimation problems.

I think the best solution is to use constraints=cases. If you subtract the average item difficulty from the the student abilities and from the item abilities, the results will be the same as constraints=items.

Another option is to use Quadrature as estimation method. That seems to work with constraints=items.

I'll let you know if I find out more.

Cheers
Eveline

3
Questions and Answers / Re: separation reliability chi-squared value
« on: November 03, 2016, 12:34:35 AM »
Hi Tong

Yes, sorry, that is a typo indeed.

Regarding the est/SE, Wald statistics are asymptotically normal. It's the standard large sample stats theory.

Cheers
Eveline

4
Questions and Answers / Re: about empirical standard errors
« on: November 02, 2016, 05:56:47 AM »
Hi Tong

I see ConQuest complains about a negative variance and possible model misidentification when using constraints=items during the estimation of the empirical errors. I am not sure yet why, but I am looking into it. The data looks fine, but CQ may struggle with the small number of cases and items. I'll get back to you when I know more.

Eveline


5
Questions and Answers / Re: separation reliability chi-squared value
« on: October 23, 2016, 11:13:53 PM »
Hi Tong

The null hypothesis is that all parameters are equal, i.e. they must be zero.  So est/se is distributed N(0,1)  then (est/se)^2 is chi-square with df=1 and sum of them is chi-square with df=NI-1.

Best wishes
Eveline

6
Questions and Answers / Re: about empirical standard errors
« on: October 23, 2016, 11:11:52 PM »
Here some responses to both your questions:

Empirical standard errors are the default in ConQuest. If CQ chooses quick standard errors, there is probably a problem in your model. One issue could be that ConQuest needs a zero category for rating scale responses. If the data is coded as 1,2,3,4,5, you could add the command "score (1,2,3,4,5)(01,2,3,4);"

Empirical standard errors are not necessarily larger or smaller than quick standard errors.

If the model converges, WLE estimates should be (close to) identical each time you run the same model. Plausible value results differ slightly because of the measurment error. Measurement errors are larger for short tests. If differences in results seem large between runs, your model may not have converged well. For most simple models, adding "keeplastest=yes" to the SET command will fix the problem. Without this option, CQ takes the results from the iteration with the lowest deviance. Usually this is the last iteration, but not always, in which case CQ takes the results from an iteration before the parameters were converged. It is important to export a log file and examine if the results are acceptable.

Hope this helps.

Best wishes
Eveline

7
Questions and Answers / Re: about empirical standard errors
« on: October 20, 2016, 10:24:21 PM »
Hi Shentong

Could you please email me your input files? I'll take a look and reply.

Cheers
Eveline

8
Questions and Answers / Re: DIF report question
« on: October 05, 2016, 08:55:38 AM »
Hello

The second model is correct. The item*year term will give you the DIF estimates.

(I'd use the order: item + item*step - year + item*year)

Best wishes
Eveline

9
P.S. Please send both command and data files to gebhardt@acer.edu.au.

10
Hi Phil

That sounds strange indeed. Could you please send me the command and data file?

Cheers
Eveline

11
Questions and Answers / Re: Negative variance fix with constraint=cases?
« on: October 05, 2016, 08:47:54 AM »
And could you tell me what code you used for missing by design?

12
Questions and Answers / Re: Negative variance fix with constraint=cases?
« on: October 05, 2016, 08:46:00 AM »
Hi Phillip

Could you email me your data and command file? I'll take a look.

Best wishes
Eveline

13
Hello

Yes, small difference are normal and caused by different random starting values and random draws for plausible values.

Best wishes
Eveline

14
Questions and Answers / Re: How to interpret case fits?
« on: July 14, 2016, 03:30:38 AM »
Hi Hye

Sorry for the very slow response.

Does this note help?
https://www.acer.edu.au/files/Conquest-Notes-3-CaseFitAndResiduals.pdf

Cheers
Eveline

15
Yes, these models are nested.

Pages: [1] 2 3 ... 7