Speech Recognition - General questions about Dragon Naturally Speaking

[schleel]

I have some general questions about using Dragon Naturally Speaking:
+ Have any of you found assessment tools (surveys, questionnaires, etc.) that help in making decisions about voice recognition software as a tool for students with writing challenges?
+ Is anyone using it with students in elementary school or are most users in middle and high school?
+ Have any of you found really useful training materials (tutorials, videos, etc.)
+ What microphone are people having good success with?

[Julie]

I don't know of any specific assessment tools to determine usefulness of Dragon for a particular student, but I can share what I do.
First off, it's definitely not 'least restrictive' so I try everything else first- keyboarding, Co:writer (if spelling is a big problem), talking word processor, printing, cursive, etc.- and I do timed one minute samples to have a baseline for comparing speed, legibility, etc.  Sometimes I do a second timing that includes the correcting (e.g. correct as they go in Word with Right-click, correct Dragon dictation each phrase) to get a really accurate picture of efficiency.

A sample dictation to a scribe can be very useful (student composes aloud, you write or type exactly what they say). If a student can't compose, Dragon will not benefit them!

I find the most important requirements for a student to do well with Dragon are:
Good frustration Tolerance!
Ability to read what they've written and edit it
Ability to think through and organize what they want to say before speaking
Knowledge of how to get around the computer in general (e.g. where to go on menus, etc.)
Clear articulation

I have used Dragon with elementary school kids who have painfully slow and/or illegible writing AND are cognitively able. I've never tried SpeakQ but have heard it is easier to use for the younger set.

Lastly, before I would ever purchase Dragon for a student, I would do a trial for a month or two with a loaner to make sure they are using it and that it is actually increasing their efficiency.

Training? I model the program first (student watches me use it), then I always have the student go through some/most of the training that comes with the program. I always have them actually try each training page as they go through it. Then I start with having them dictate from text  (so it removes the figuring out what you want to say piece) before having them actually compose.

Hope this helps. Let me know if you have more specific questions.

[schleel]

I was recently browsing the GPAT (Georgia Project for Assistive Technology) website and discovered that they have a checklist for evaluating the use of voice recognition software in the 'Reading and Writing' protocol in their AT Evaluation materials. This could be useful for documenting a student's success (or lack thereof) when trying a voice recognition program. This is how you get to the form:
- Go to the GPAT home page at www.gpat.org
- Click on 'Resources' at the top of the page
- Click on 'AT Evaluation' on the left side of the page
- Click on 'Protocols' on the left
- Click on 'Writing and Spelling' in the center of the page
- The 'Voice Recognition' section is on page 9 of this PDF document

[Julie]

I just wanted to add that I'm a firm believer in trial periods to determine the usefulness of AT for any student. So once a student has been introduced to Dragon and has run through the tutorial and practiced, I'd suggest a 6 week period (at least) of ongoing supervision. For related service staff, this may be just once a week directly. Have the student dictate from text for a few sentences, correcting as they go, every phrase or so. Make sure they are selecting rather than saying 'scratch that' to everything so that the program will continue to get better at recognizing their speech.

The student also needs time to practice on their own, and should ideally be using Dragon regularly. During the trial period, check in with teacher(s) to make sure the student has easy access when needed and that they are actually using Dragon for any writing assignments longer than a few sentences. It may sound simplistic, but the question 'Are they using it regularly?' is a key one. If not you need to look at why. Some students just hate to write and/or have trouble composing. Some need more training with Dragon.

For many students, speech recognition does not speed up their writing! It can be laborious and frustrating. Even those who do well with it and learn the process quickly do not always choose it as their preferred writing method.

Last but not least, at the end of their 'trial' period, do a timed comparison. Try dictating a few sentences and time how long it takes the student to write AND EDIT the sentences. Compare this time (with the same dictated sentences) for handwriting, keyboarding with right-click, keyboarding with Co:Writer, or whatever the other best options are for that student. And finally, ask the student what they prefer and what they're most likely to use.

Regarding recommendations in the IEP, if you're not sure the student needs or will use speech recognition, I usually states something to this effect... A trial of speech recognition software is recommended to determine it's efficiency (speed plus accuracy) for (student) for producing written language.