LDL Babbler 
                       An Occasional Newsletter of the
                      Language Development & Disorders Lab
                      Communicative Sciences & Disorders
                      Michigan State University
                 East Lansing, MI
               

LDL Sponsors McNair/SROP Scholars
This summer semester the Language Development & Disorders Laboratory (LDL) of the Department Audiology & Speech Sciences served as a sponsor site for two McNair/SROP student scholars.  Dr. Michael W. Casby, Professor of Audiology & Speech Sciences, served as research mentor to the following two scholars:  Nora Sailor and Tamika Lucas.  Ms. Sailor is an Audiology & Speech Sciences student at Michigan State University, and Ms. Lucas is a Speech-Language Pathology student at Southern University in Louisiana.

McNair/SROP Program
The McNair/SROP Scholars program is a special program of research and advanced study for achieving students.  The program is named for Dr. Ronald E. McNair.  Dr. McNair was a prestigious physicist who met an untimely death while serving aboard the USS Challenger Space Shuttle, which had a fatal explosion.  SROP stands for summer research opportunity program. 

Scholars work with a faculty mentor in conducting research over a selected time period.  The Scholars then present their research at local and national McNair/SROP meetings. 

The program is funded by the U.S. Department of Education, and is under the local coordination of Nettavia Curry, Michigan State University, Office of Supportive Services.

The Project
The scholars worked with Dr. Casby on a project entitled -- "The Acoustic Substance of Morphophonemic Forms."  Its purpose was to measure and examine the acoustic substance, e.g., duration, of the morphophonemic form of is in various linguistic contexts.  For example, the morphophonemic form of is may function as a copula, or linking, verb as in the utterance -- "The dock is painted."  Or it may function as an auxiliary, or helping, verb as in the utterance -- "The Doc is painting."  The same form also may be part of a noun as in the utterance -- Mr. Dockiz painted the fence."

It should be noted that the phonemic context of the form is in the example utterances is always the same for experimental control purposes -- the context only differs by the various linguistic/ morphologic functions.

The project is of particular interest given that young children learn the copula and auxiliary forms later than they might learn a noun form, and further, children with language impairments have marked difficulty learning such grammatic forms as the copula and auxiliary is, without demonstrating similar difficulty learning nouns. 

One hypothesis of this research was that the acoustic substance of the form is may be less, or briefer, when produced as a grammatical verb form than when produced as part of a noun.  This shorter duration may account for the verb forms' later development by children, as well as account for the difficulty in learning the grammatical forms by children with language impairments.

Preliminary results of the research have indeed showed that the morphophonemic form is was of less acoustic duration, and briefer, when produced in either the copula or auxiliary verb context, than when produced in the noun context.

Scholars' Activities
As McNair/SROP scholars, Sailor and Lucas ran subjects by audio taping them producing experimentally controlled stimuli.  They then transferred the speech samples to computer for acoustical analysis. Using computer-generated acoustic waveforms of the speech stimuli, Sailor and Lucas then made durational measurements of the is form in the various linguistic contexts.  Such durational measurements were made at the level of milliseconds.  The scholars then entered these data into a database for later statistical analyses. 
 


 


FFW -- FFW
Fundamental Flaws With -- FastForWord*

In this edition of the LDL Babbler, we discuss the roundly hyped and heavily marketed central auditory processing computer-based language intervention (cautionary practice) program of FastForWord.  We will take a strong, objective, and critical look at some of the pertinent issues regarding this program.

Flawed Research Control

  • No control group -- The first of the two peer-reviewed articles on FFW reporting significant improvements did not have a control group.  There is no evidence of fundamental research controls in the often cited "field research" project with 100s of subjects across various developmental disabilities for whom significant improvements are reported.
  • Control group improved -- The second of the two peer-reviewed articles did have a control group.  However, this control group demonstrated significant improvements in language performance at post-testing as well.  Furthermore, the group exposed to the FFW program also had other language intervention activities.
  • Order Effects -- Yet another research design flaw of the FFW research is the lack of control for order effects on a number of fronts.  The research makes use of a number of different activities or "games."  Some of these components consist of non-linguistic auditory processing activities, some consist of non-meaningful linguistic auditory processing tasks, and yet others consist of language comprehension activities.  Once again, control for order effects is lacking across these program components.  Nor is there a concern for an examination of which component or combination of components, and what order of exposure may be effecting any improvement in language.  Additionally, FFW makes use of five different levels of duration- and intensity-altered acoustic stimuli.  Research control over these different levels of acoustic modification is also notably lacking. 
Flawed Measurement
  • The use of AEs -- The FFW research has routinely used age-equivalence scores (AEs) when reporting improvements.  A major difficulty with AEs is that for many tests an increase of a very few points may make an apparent 1 to 2 year gain vis-à-vis age equivalence.  Such small point changes however will not result in large gains when evaluated against psychometrically-sounder evaluation schemes such as utilization of standard scores or means and standard deviations.
  • SEMs -- In reporting increases on standardized tests the FFW research often fails to report or only partially reports standard error of measurement (SEM) data from their repeated testing of subjects.  SEM data are of great importance in the evaluation of the effects of repeated testing and potential regression to the mean.
Flawed Reporting of Improvement in Language Usage
The research on the the effectiveness of FFW has not routinely examined the participants' actual language for communication use in pre- or post-testing. There are no controlled data to demonstrated gains in the functional-pragmatic usage of language of children after they have completed the intervention program.

Teaching to the Test
One of the most disconcerting aspects of the FFW research is that many of the activities of the program are "teaching the tests", or most certainly "teaching to the tests", that are used in post-testing to document behavior change.  For example the "Block Commander" activity closely resembles the Token Test for Children, and the "Language Comprehension Builder" training activity is closely based on the Curtiss and Yamada language comprehension test, as well as resembling numerous other tests of receptive language.  Both of these instruments are often used in the post-testing of the participants. The training activity of "Circus Sequence" closely resembles the auditory temporal resolution test used in post-testing.

Similar Levels of Improvement are Found Across Other Intervention Strategies
In general the FFW research reports an average gain of approximately 1 standard deviation improvement in participants' assessed language performance at the conclusion of the training.  When we look at the broader field of speech-language pathology and language intervention with young children, this level of improvement is not that uniquely remarkable.   That level of improvement is a fairly standard effect of most reported research on the effectiveness of many different language intervention approaches.  Meta-analyses of the efficacy of language intervention for young children have reported an effect size of approximately 1 standard deviation improvement.

Children Make Short-Term Gains on Auditory Temporal Resolution Tasks Without Intervention
Other research has demonstrated that young children make significant gains in their performance on auditory temporal resolution tasks -- so critical and core to the FFW approach -- within a limited exposure and familiarity time.  This calls into question the significance of the core auditory temporal resolution impairment.  If such deficits were so significant and core to children's language learning, then it is doubtful that such limited exposure and familiarity would lead to significant improvements in performance.  This other research also is demonstrative of the lack of a need for a long and intensive period of training to gain improved performance on auditory temporal resolution tasks.

FFW Acoustic Modification of Natural Speech Makes it Unintelligible
Recent research in the Language Development & Disorders Lab, of the Department of Audiology & Speech Sciences, Michigan State University has demonstrated that the early levels (i.e., 1-3) of the 5 levels of acoustic modification of speech found in FFW are essentially unintelligible to listeners.  It is counter intuitive to think that distorted and unintelligible speech stimuli could lead to language gains in young children.

Flawed View of Psycholinguistic/Language Processing
Numerous other psycholinguistic and speech-language pathology researchers have seriously called into question the basic core of the model that FFW is based upon.  One that claims that children with language impairments have a specific deficit in the processing of rapidly presented stimuli.  Or alternatively, that young children who have language impairments have great difficulty in the temporal resolution/auditory processing of rapid acoustic-phonetic stimuli.  Many have challenged these views, arguing for a language-phonemic, or auditory linguistic information processing perspective, versus the primacy of an acoustic-temporal perspective.

In evaluating the demonstrated effectiveness of FastForWord across the following points, it is found to be notably lacking as are other central auditory processing approaches:  a) lacking a sound deductive/rational/logical basis, b) lacking a sound theoretical basis, c) lacking sound inductive/empirical support, d) lacking demonstrations of potential for or actual improvement in communication performance, and e) viable and sound alternative explanations, interpretations, etc. exist.

* (Editor's note:  The phrase FastForWord and FFW are registered trade names/marks of Scientific Learning Corporation of Berkeley CA.  Their use here, or any other reference to them or other products or rights of Scientific Learning Corporation, is for educational purposes.  No other purposes are expressed or implied.  References for cited research are available upon request.)