Katya Hill1,3, Barry A. Romich1,2, Jennifer Thiel3,
1 University of Pittsburgh, Pittsburgh, PA 15260
2 Prentke Romich Company, Wooster, OH 44691
3 Edinboro University, Edinboro, PA 16412

Automated data logging for AAC has become a reality with the development of performance monitoring tools such as the Language Activity Monitor (LAM). In order to facilitate widespread application of accumulated LAM performance data, procedures need to be documented and disseminated for editing logfiles to increase the usefulness of raw logfile data. One of the most valued clinical summary measures is communication rate, which can be calculated using logfile data either manually or automatically with software. However, reliability of the reported results is based on adherence to operational procedures established for each method of analysis. This paper presents procedures developed to automatically calculate communication rate.

AAC automated language activity monitoring (LAM) provides the field of AAC with tools needed to collect and analyze language samples in a variety of clinically useful contexts (1). The essential function of the LAM is the recording of each language event and the time that it occurs. Hill and Romich (2) as well as Higginbotham (3) have proposed a standard protocol for automated data logging to address compatibility issues and facilitate the widespread application of actual user-performance data collection. Presently, the LAM function is available as an add-on device or computer monitor for any AAC system with a serial port representation of language events and as an internal function in newer high–end AAC devices. Recorded language samples include content and time stamps. The following is an excerpt of a raw LAM logfile converted to four columns:

16:26:05 "It's ”
16:26:08 "faster”
16:26:14 "than ”
16:26:41 "sp”
16:26:42 "e”
16:26:45 "l”
16:26:45 "l”
16:26:46 "i”
16:26:47 "n”
16:26:48 "g"
16:26:49 " "
16:26:58 "everything "
16:27:02 "out "
16:27:05 "which "
16:27:08 "is "
16:27:11 "what "
16:27:14 "I "
16:27:19 "used "
16:27:22 "to do "

This example illustrates that the raw LAM logfile data requires editing to improve the usefulness and value of the time and content information recorded. The editing process is required to prepare the logfile for analysis. Specific procedures are required based on the software application used for analysis and the summary measures selected for reporting. To date, operationalized procedures have been developed for logfiles to be reliably edited for several analysis programs. The software programs are selected based on the summary measures available for analysis that have been proven to be clinically useful. Since communication rate is a summary measure considered highly valued and frequently requested by AAC consumers and practitioners, the focus of this paper is the procedures for editing logfiles for communication rate analysis and setting up Augmentative Communication Quantitative Analysis (ACQUA) program to provide the desired report.

In order to maximize the usefulness of logfile data, procedures need to be developed for editing the logfiles in preparation for analysis. Specific procedures needed to be identified for editing raw LAM logfiles for calculating communication rate using the ACQUA program(4). In order to develop editing procedures, the following problems were identified: 1) the need for a systematic approach for defining and calculating communication rate, and 2) the need to incorporate this definition into any suitable software application, such as ACQUA.

A standard method for calculating peak and average communication rate in the clinical setting has been proposed by Romich and Hill (5). The method provides a listing of steps to be followed for converting raw LAM data into peak and average words per minute. ACQUA was developed for computing a wide variety of AAC usage statistics based on logfile data. The Romich and Hill method for peak and average communication rate calculation was added to ACQUA version 1.0. Several logfile editing procedures needed to be developed for the program to automatically calculate the same results obtained through manual calculation methods. Since ACQUA does not provide for editing capabilities within the application, editing of the raw logfiles is performed once the data is uploaded into a word processor. The most important editing step involves utterance segmentation and the insertion of an utterance terminator. Frequently, a previously prepared language transcript is used as a model for the utterance segmentation process.

A total of twelve editing rules have been documented for ACQUA to calculate communication rate. The editing process requires the following basic steps. 1) insert utterance terminators at the end of the last word of an utterance, 2) if a terminator exists, move it to the end of the last word of the utterance, 3) take out error words, 4) delete all pre-stored messages.

ACQUA requires the following features be selected to perform the peak and average rate calculations: 1) set Type = Utterances, 2) set Size = Global, 3) set Gap = 1, 4) in tool options select exclude first entry, 5) in tool options check peak value over data windows.

Pilot study data comparing the manual method with ACQUA results indicate that the statistical analyses are consistent using these operational procedures. A prototype LAM report and Communication Rate Worksheet have been designed to record and report the results for clinical application. Tables 1 and 2 are examples of analyzed reported logfile data.

Table 1: Example of results reported on Communicate Rate Worksheet for picture description task

Words after 1st event


She is just washing away, not knowing that the water is about to run over.
The brother is trying to get a cookie from the jar and it looks like it could fall.

Table 2: Communication rates for four augmented communicators’ performance during an interview.

AAC System
Selection Technique
Average Rate (WPM)
Peak Rate (WPM)
Unity/Deltatalker Direct keyboard
Unity/Liberator Direct keyboard
Custon/Vanguard Direct keyboard
Unity/Pathfinder Optical Headpointing

Over 40 logfiles under two sampling conditions (picture description and interview) have been edited following the procedures described in this paper. Early inter-rater reliability in utterance segmentation is 96%, and 100% for word-by-word agreement. The use of these editing procedures is a component of the application of tools for measuring AAC performance. Standardized editing procedures provide clinicians with reporting protocols that are comparable, compatible and have reliable quantitative data for a variety of clinical applications. As improvements are made to available tools clinicians will have access to even more time efficient and accurate methods. These tools will increase the application of evidence-based practice using performance measurement based on automated data logging. This in turn will benefit people who rely on AAC through improved clinical intervention service and more consistent, periodic performance reporting.

The authors wish to acknowledge the support and cooperation of the AAC-RERC. In particular, they would like to express their appreciation for the collaborative efforts with Jeff Higginbotham, State University of New York-Buffalo, and Greg Lesher and Rod Rinkus at Enkidu Research, Inc.

1. Romich, B.A. & Hill, K.J. (1999) A language activity monitor for AAC and writing systems: Clinical intervention, outcomes measurement, and research. in Proceedings of the RESNA '99 Annual Conference, Arlington, VA: RESNA Press. 19-21.

2. Hill, K.J. & Romich, B.A. (1999). A proposed standard for AAC and writing system data logging for clinical intervention, outcomes measurement, and research. in Proceedings of the RESNA '99 Annual Conference, Arlington, VA: RESNA Press. 22-24.

3. Higginbotham, D.J. & Lesher, G.W. (1999). Development of a voluntary standard format for augmentative communication device logfiles. in Proceedings of the RESNA '99 Annual Conference, Arlington, VA: RESNA Press. 25-27.

4. Lesher, G.W., Rinkus, G.J., Moulton, B., & Higginbotham, D.J. (2000) Logging and analysis of augmentative communication. in Proceedings of the RESNA ’00 Annual Conference. Arlington, VA: RESNA Press. 82-84.

5. Romich, B.A. & Hill, K.J. (2000). AAC communication rate measurement: tools and methods for clinical use. in Proceedings of the RESNA '00 Annual Conference, Arlington, VA: RESNA Press. 58-60.

Katya Hill
102 Compton Hall
Edinboro University of Pennsylvania
Edinboro, PA 16444
Tel: (814) 732-2431
Fax: (814) 732-2184
Email: khill@edinboro.edu