Speech Tracking (sometimes called Connected or Continuous Discourse Tracking) is a procedure developed by De Filippo and Scott "for training and evaluating the reception of ongoing speech'' (De Filippo and Scott, 1978, p. 1186). In the Speech Tracking procedure, a talker (usually a therapist or experimenter) reads from a prepared text, phrase by phrase, for a predetermined time period, usually five or ten minutes. The task of the receiver (the person with hearing loss) is to repeat exactly what the talker said. If the receiver does not give a verbatim response, the talker applies a strategy to overcome the breakdown. This can take the form of simple repetition, the use of clue words, or a paraphrase of the original segment. The therapist goes on to the next phrase only when the receiver is able to repeat correctly every word of the original segment.
At the end of the time period, the number of words repeated correctly is counted and divided by the time elapsed to derive the receiver's Tracking Rate, expressed in words per minute (wpm). For example, if a receiver was able to repeat 150 words in five minutes, his Tracking Rate would be 30 wpm. The Tracking Rate represents the time taken for the text to be presented and repeated. Estimates of Tracking Rate for people with normal hearing vary, but it is generally recognized that a rate of around 100 wpm (De Filippo, 1988) is obtained if the same presentation and response rules are followed.
Since its introduction, Speech Tracking has also been used extensively in investigations evaluating the effectiveness of sensory aids for people with hearing loss. These have included studies of cochlear implants (Rob-bins et al., 1985; Levitt et al., 1986), tactile aids (Brooks et al., 1986; Cowan et al., 1991; Plant, 1998), and direct contact tactile approaches such as Tadoma (Reed et al., 1992) and Tactiling (Plant and Spens, 1986). These studies usually compare a receiver's Tracking Rate in two or more presentation conditions such as aided and unaided lip reading. For example, Plant (1998) looked at a subject's Speech Tracking performance with materials presented via lip reading alone and lip reading supple mented by the Tactaid 7 vibrotactile aid. After about 30 hours of testing and training with Speech Tracking, the subject's mean Tracking Rates in the two presentation conditions were 33.7 wpm for lip reading alone and 46.1 wpm for lip reading supplemented by the Tactaid 7.
Despite the widespread acceptance of Speech Tracking in research projects, its use as an evaluative tool has been severely criticized by Tye-Murray and Tyler (1988). These researchers cited a number of extraneous variables that they believed were extremely difficult to control. These included the characteristics of the speaker (segment selection, ability to use cues to overcome blockages, speaking style, articulatory patterns, etc.), the receiver (degree of assertiveness, language proficiency, motivation, etc.), and the text (degree of syntactic complexity, vocabulary, etc.). Hochberg, Rosen, and Ball (1989), for example, found that text complexity could greatly influence Tracking Rate. In a study using the same talker/receiver pairs, they found that Tracking Rates varied from 62.9 wpm for "easy" materials (controlled vocabulary readers designed for English-as-a-second-language learners) to 29.5 wpm for "difficult" materials (popular adult fiction).
Tye-Murray and Tyler (1998) believed that these factors made Speech Tracking unsuitable for across-subject test designs. They did, however, feel that with some modifications the procedure could be used for within-subject test designs. These recommendations included an insistence on a verbatim response, the use of only one speaker, training of the speaker/receiver pairs, the use of appropriate texts, and limiting the repair strategies used to repetition and writing down a blocked word after three repeats.
A number of groups (for example, Boothroyd, 1987; Pichora-Fuller and Benguerel, 1991; Dempsey et al., 1992) have attempted to make the technique more suitable for evaluative purposes through the use of computer-controlled recorded materials. These approaches ensure that the problems created by sender differences are controlled and minimized. Although promising, none of these systems have become widely accepted and used.
The KTH Speech Tracking Procedure (Gnosspelius and Spens, 1992; Spens, 1992) represented another attempt to more closely control the approach. This computer-controlled modification used live-voice presentations, but the segment length was predetermined and only one repair strategy—repetition—was allowed. The written form of any word repeated three times was automatically presented to the receiver via an LED display. At the end of a Speech Tracking session the program automatically calculated a number of measures including Tracking Rate, Ceiling Rate (time taken when all words in a segment were correctly repeated after only one presentation), and the Proportion of Blocked Words (the total number of words that had to be repeated divided by the total number of words in the session). This approach has been used in a number of studies (for example, Plant, 1998; Ronnberg et al., 1998) but has not gained widespread acceptance.
Speech Tracking has also been widely used as a training technique for use with adults (Plant, 1996) and children (Tye-Murray, 1998) with profound hearing loss. When used for training, modifications can be made to provide receivers with practice in the use of repair strategies. Owens and Raggio (1987), for example, provided receivers with a list of directives they could use when they did not correctly repeat a segment. The receiver could ask the sender to:
1. Say that again.
2. Say it another way.
3. Spell that word.
4. Write that word.
5. Spell an important word.
6. Write an important word.
Lunato and Weisenberger (1994) looked at the comparative effectiveness of four repair strategies in Speech Tracking. The repair strategies used were:
1. Verbatim repetition of a word or phrase.
2. The use of antonyms or synonyms as clues to the identity of blocked words.
3. Providing the receiver with phoneme-by-phoneme correction of blocked words.
4. Providing context by moving forward or backward in the text.
These researchers reported that strategy 1 yielded the highest Tracking Rates and strategy 2 the lowest.
When Speech Tracking is being used for training, compromises can also be made in the receiver's response patterns. Owens and Raggio (1987), for example, argued for the use of nonverbatim responses in training sessions using Speech Tracking. While acknowledging the importance of a verbatim response for test purposes, they felt that in training it may be better to provide the receiver with practice in picking up the gist of the message rather than expecting absolute identification at all times.
Although Speech Tracking has become a widely used training procedure, there are some people with hearing loss for whom it is unsuitable. These include people with very poor speech reception skills, resulting in Tracking Rates of less than 20 wpm. At these levels receivers find the task extremely difficult and stressful. Others for whom the technique may be unsuitable include people with poor speech production skills. In such cases the sender may be unable to determine reliably whether the receiver gave the correct response. This necessitates the use of a written or a signed response, which serves to greatly reduce the Tracking Rate.
Plant (1996, 1989) developed a modified version of Speech Tracking designed to be used with such cases. Simple stories are divided into parts, each consisting of 200 words. Each part is in turn divided into short segments ranging in length from 4 to 12 words. The segments are presented for identification, and the receiver is asked to repeat as many words as he or she can. The receiver can use repair strategies to obtain additional information if he or she experiences difficulties. The re ceiver is then scored for the number of words correctly identified and shown the written form of the segment. At the end of each part, the percentage of correct responses is calculated, based on the number of words correctly identified. This approach provides the receiver with immediate feedback on the correctness of her or his response and ensures that she or he is able to benefit from ongoing contextual information.
Speech Tracking is an innovative technique that can be used for both testing and training the speech perception skills of people with hearing loss. When used for testing, however, a precise protocol must be followed to minimize the effects of sender, receiver, and text variables. In training a less rigid approach can be used, and the approach may be modified to include practice in the use of repair strategies.
Was this article helpful?
Have you recently experienced hearing loss? Most probably you need hearing aids, but don't know much about them. To learn everything you need to know about hearing aids, read the eBook, Hearing Aids Inside Out. The book comprises 113 pages of excellent content utterly free of technical jargon, written in simple language, and in a flowing style that can easily be read and understood by all.