Use of speech generating devices can improve perception of qualifications for skilled,verbal,and interactive jobs

Abstract

BACKGROUND: We have previously demonstrated that when speech generating devices (SGD) are used as assistive technologies, they are preferred over the users’ natural voices.

OBJECTIVE: We sought to examine whether using SGDs would affect listener’s perceptions of hirability of people with complex communication needs.

METHODS: In a series of three experiments, participants rated videotaped actors, one using SGD and the other using their natural, mildly dysarthric voice, on (a) a measurement of perceptions of speaker credibility, strength, and informedness and (b) measurements of hirability for jobs coded in terms of skill, verbal ability, and interactivity. Experiment 1 examined hirability for jobs varying in terms of skill and verbal ability. Experiment 2 was a replication that examined hirability for jobs varying in terms of interactivity. Experiment 3 examined jobs in terms of skill and specific mode of interaction (face-to-face, telephone, computer-mediated).

RESULTS AND CONCLUSIONS: Actors were rated more favorably when using SGD than their own voices. Actors using SGD were also rated more favorably for highly skilled and highly verbal jobs. This preference for SGDs over mildly dysarthric voice was also found for jobs entailing computer-mediated-communication, particularly skillful jobs.

Keywords

Augmentative and alternative communication disability discrimination computer synthesized speech assistive technology stigma

1 Introduction

According to the 2012 American Community Survey [1] only 32.7% of 18 to 64 year olds with disabilities in the United States are employed compared to 73.6% of their counterparts without disabilities. Researchers who have focused on people with complex communication needs estimate that the employment rate might be as low as 15% [2].

This very low rate of employment is partly due to the logistical barriers inherent in hiring and employing persons with disabilities [3]. It can also be attributed partly to the well documented stigmatization of [4], prejudicial attitudes toward, and discriminatory practices against persons with disabilities [5 –8]. Also at play are stereotypes that people have regarding whether a person with a disability is a good fit for particular types of employment [9 –11]. Other factors that affect employers’ willingness to hire include prior experience with hiring persons with disabilities, the type of disability, and concerns related to job skills [12].

Persons with complex communication needs often report difficulties in everyday interpersonal interactions. They frequently report perceiving that other people avoid, infantilize, and distance themselves from them [13]. Furthermore, some report feeling stigmatized and isolated, leaving them with fewer social contacts and friendships [14 –16]. Insofar as they often feel misunderstood, they report a tendency toward keeping conversations brief and simple [17]. They also report trouble communicating with new people, difficulty expressing their personalities, and withdrawing from activities that require speaking [17]. All of this considered, it is not surprising that complex communication needs are frequently accompanied by diminished self-worth, depression, loneliness, and social withdrawal [13 , 18].

1.1 Speech generating devices as workplace accommodation

Anti-discriminatory legislation worldwide is designed to provide protection for people with disabilities. In the United States, the Americans with Disabilities Act (ADA) of 1990 aims to protect people with disabilities from discrimination in employment and requires employers to provide reasonable accommodations to make it possible for people to work. One form of accommodation is permitting and providing for the use of assistive technologies.

A speech generating device (SGD) that converts inputted text into outputted synthetic speech is clearly an acceptable form of reasonable accommodation [19]. Although communication with SGDs has potential limitations, including speed of communication and the amount of content able to be communicated [20 –22], people using SGDs as an assistive technology have found employment in many walks in life. This includes stereotypically verbal fields such as teaching, counseling, journalism, contract negotiating, and brokering [23 –25]. It is clear that SGD technology, continuing to develop in its capacities [26, 27], can facilitate full and meaningful employment that is both financially and psychologically rewarding for many of the people with complex communication needs who need or wish to work[24 , 29].

Paradoxically and unfortunately, people do not react as positively to output from SGDs as they do to natural human speech. Laboratory studies have consistently shown that output from SGDs is less liked, less persuasive [30], more mechanical sounding, and less natural [31] than human speech. In research that has examined perceptions of SGDs outside of the context of disability, such as when SGD is used to create an automated message [32], people rate SGDs negatively compared to natural speech. They also rate the people using the SGDs and content of the messages they are conveying less favorably.

When SGDs are used as assistive technologies, they better tolerated by listeners [30]. Research has specifically shown that output from SGDs is favored in comparison to the speech of people with dysarthria; that is motor speech disorders [33]. This has been found even in cases where the dysarthria can be classified as mild [34]. People with previous experience with people with disabilities respond more positively to SGDs than people without experience [35].

While SGDs are considered to be reasonable accommodations for employees with voice disorders [19], there are numerous challenges in placing SGD users into jobs [36]. SGD users frequently need training in communication as well as their specific vocation. A good job fit will also require a committed employer who can provide resources needed for the accommodation to the technology. Any fit will need to tolerate the limitations of the technology which includes slower communication and interactions [20].

All of these challenges considered, the present line of research is designed to better understand the attitudes and reactions that listeners have toward persons who use SGDs, particularly in regards to hirability. How well someone fits into the workplace is immaterial if they are less likely to be considered hirable for the job in the first place. Hence, laboratory research on perceptions of SGD users helps to better inform us about first impressions held by naïve listeners.

1.2 Overview and research objectives

In three controlled laboratory experiments, we examined perceptions of hirability of people with physical disability who were either using an SGD or speaking with their natural albeit dysarthric voice. Unlike most controlled laboratory work comparing reactions to SGDs with reactions to natural speech, the reported studies go beyond examining listeners’ impressions of intelligibility or social perceptions [34]. We studied the more socially robust and complex issue of employment. Hence, the present line of inquiry should complement the more phenomenological qualitative research on the everyday opportunities and challenges faced by people who use SGD [25, 37].

We recognize that in comparison to qualitative research on the use of SGDs, a laboratory paradigm with naïve listeners watching video clips of actors with disabilities has less mundane realism as well as less social validity, the extent to which the participants in the study are from the population of interest [38]. On the balance, however, experimentally controlled research helps us narrow in on specific factors that might otherwise be confounded or overlooked in more naturalistic and qualitative research settings.

The research question guiding our study was whether the preference of SGD output over mildly dysarthric speech would also apply to perceptions of hirability. We expected that certain salient aspects of jobs such as skill and verbal requirements of the job, would affect participants’ perceptions of hirability of people using an SGD versus their own natural speech. Experiment 1 introduces our paradigm and examines whether the preference of SGD output over dysarthric speech would be affected by the skill level and verbal requirements of the job. Experiment 2 examined whether the preference of SGD output would be affected by whether or not the jobs required interaction. Experiment 3, finally, examined whether the preference of SGD output would be affected by the specific mode of interaction required by the job; face-to-face (FTF), computer-mediated-communication (CMC), or telephone.

2 General method

2.1 Overview

In all three experiments, participants viewed videotaped segments of persons with physical disabilities delivering a persuasive appeal. Following the video clips, participants rated their perceptions of the actor and perceptions of the likelihood that the individual would be hired for a series of jobs. The Institutional Review Board of the University of Pittsburgh approved all procedures.

2.2 Stimuli

Our stimuli consisted of video clips of two men with physical disabilities and mild dysarthria (photographs of the disabled actors can be seen in Fig. 1). They were recruited from a local state run vocational rehabilitation facility and were paid for their participation. Both men used wheelchairs, were dressed in casual street clothes, and can be seen in the video clips wearing t shirts with logos. Actor 1 was a 24-year-old male with a spinal cord injury resulting in paraplegia. Actor 2 was a 19 year old with cerebral palsy with spastic quadriparesis. Although neither of these actors actually required an SGD to communicate, they had peers who used SGDs and were able to emulate using an SGD. By having the same actors use both an SGD and their own natural speech, we were able to experimentally control for participants’ perceptions of the individual actors’ appearance and the extent of their physical disability. From the participants’ perspective, they watched one person talk with mildly dysarthric speech and a second person (not necessarily in this order) talk with the use of an SGD.

The intelligibility of the actors’ natural voices was assessed with a 3×3 Latin squares design (raters X sentences). A sample of naïve raters (n = 21) watched a sentence long excerpt from each of the two actors using their own voices and an actor using an SGD [39]. Each rater listened to each actor reading different sentences. They listened to each sentence three times and transcribed it to the best of their ability. For each actor and the SGD, a mean score was calculated by dividing the number of words understood by the number of words spoken. The mean scores for the two actors were as follows: Actor 1 (spinal cord injury) 86.8%; Actor 2 (cerebral palsy) 70.6%. This indicates that they had mild (Actor 1) to moderate (Actor 2) dysarthria [40]. The mean sentence intelligibility for the computer synthetized speech sample was 75.5%, indicating that it was similarly intelligible to the natural samples.

The text that the actors read was a standardized persuasive appeal [41] in favor of comprehensive exams at universities. This text typically takes approximately 5 minutes to read aloud and raise a number of issues, including the high prestige of other school that are using comprehensive exams and the value that prospective employers place on graduates of schools that use persuasive exams.

Each of these actors was videotaped delivering the persuasive appeal under one of two conditions: (a) using their natural voice or (b) using an SGD input by hand on a keyboard. In the second setting, to create the impression that the actors were using an SGD, the persuasive appeal was simultaneously outputted in the default male voice (Perfect Paul) from a commercially available TTS system (DECtalk^TM Express, V2.4C). Hence, for each of the two actors, we had two video clips; one of them delivering the persuasive appeal using an SGD and one of them delivering the persuasive appeal using their own voice.

When the actors used their own voices the video clips of the persuasive appeals were 528 s for actor 1, 525 s for actor 2. The video clips were 420 s for either actor using the SGD. No effort was made to keep the length of clips constant. We inserted a 5 s pause following each sentence in the SGD video clip in order to account for time needed to input text into the SGD. We were not concerned that the differences in intelligibility between the two actors would affect our findings. Previous research using the same stimuli [34] compared perceptions of Actor 1 and Actor 2. Actor 1 was rated significantly more positively on only one of the three factors (informedness) than Actor 2. There were no significant differences between the actors. More importantly, both actors were consistently rated more favorably using SGD than when using their own natural speech.

It may be noted that the persuasive argument is not typical of a hiring situation. It was, however, the lengthiest speech sample that was available to us from our original videotaping of these actors. Furthermore, the persuasive argument has been a successful paradigm in a number of studies examining the social influence of users of SGDs [42]. We also acknowledge that the manner in which our actors emulated SGD use is not perfectly representative of how an SGD is used in the workplace or how people with complex communication needs use SGDs in general [20, 21].

2.3 Measures

In each of the three studies, participants completed two questionnaires. The first measured their perceptions of the actor. The second rated his hirability for a series of specific jobs.

2.3.1 Perceptions of actor

Perceptions of the actor were measured with an instrument developed to examine factors that affect persuasion [43, 44]. The instrument has proven effective in numerous studies examining the social influence of users of SGD [34]. The instrument consisted of 12 items anchored as follows: 1 (unintelligent) to 7 (intelligent), 1 (straightforward) to 7 (evasive), 1 (active) to 7 (inactive), 1 (qualified) to 7 (unqualified), 1 (sincere) to 7 (insincere), 1 (meek) to 7 (forceful), 1(incompetent) to 7 (competent), 1 (honest) to 7 (dishonest), 1 (unassertive) to 7 (assertive), 1 (uninformed) to 7 (informed), 1 (untrustworthy) to 7 (trustworthy), and 1 (timid) to 2 (bold). Based upon factor analysis from prior research using this instrument [34], we aggregated the items into three factors: Speaker Credibility, Speaker Informedness, and Speaker Strength.

2.4 Hirability questionnaire

Each of the three experiments in this paper required us to develop a separate hirability questionnaire. The jobs listed on each of these questionnaires were chosen based upon the job related variables we were examining in that particular experiment (e.g., skill level, verbal ability needed, interactivity). In the pilot studies conducted to develop these questionnaires, samples of undergraduates (ns 23, 19, and 19 in Experiment 1, Experiment 2, and Experiment 3, respectively) rate a larger sample of jobs on the relevant job related variable. These rating facilitated the construction of hirability questionnaires that listed jobs exemplifying the variable that we were examining.

2.5 Design

Each experiment employed a within-groups design. All participants viewed one actor using their natural voice and the other actor using an SGD. The order of speech and whether each of the two actors was viewed using their natural voice or an SGD, however, were both counterbalanced. This design permitted us to compare perceptions of natural (mildly dysarthric) speech to SGD output while counterbalancing which actor used SGD versus their own voice and order of presentation (natural first vs. SGD first).

This design yielded the following four sequences of presentation: (a) actor 1 using an SGD/actor 2 using natural voice, (b) actor 2 using an SGD/actor 1 using natural voice, (c) actor 1 using natural voice/actor 2 using an SGD, and (d) actor 2 using natural voice/actor 1 using an SGD. Each participant was randomly assigned to one of these four orders of presentation (See Fig. 2 for an illustration of the four sequences).

3 Experiment 1

Experiment 1 sought to examine whether preference for an SGD over natural speech would be extended to perceptions of hirability. More specifically, this study examined whether the skill and verbal ability required by the job would affect the extent to which SGD output was preferred over natural speech.

3.1 Participants and design

A total of 41 undergraduates (12 men, 29 women, age 18 to 40, with a mean age of 19.90 years) from the University of Pittsburgh at Johnstown participated in the study in exchange for course credit (points toward their grade in Introductory Psychology). The 41 participants were then randomly assigned to one of the four sequences, creating three groups with 10 participants and one group with 11. Following the procedure, participants were debriefed, allowed time for questions, and thanked for theirparticipation.

3.2 Procedure

Participants were asked to watch a video segment on a desktop PC computer in a laboratory cubicle. They listened through earphones and were permitted to adjust the volume. After viewing the first video clip each participant completed the measures of perceptions of actor and job questionnaire. Participants then viewed the second video clip of the other actor in a different experimental condition and subsequently completed the same measures.

3.2.1 Perceptions of actor

Perceptions of the actor were measured with 12 item instrument described above. The 12 items were collapsed into three scores for speaker credibility, strength, and informedness.

3.2.2 Hirability questionnaire

To develop a hirability questionnaire that examined jobs varying in skill requirement and verbal ability, a pilot questionnaire with 24 jobs was distributed to the 23 students of an undergraduate research methods class. The students rated each job according to skill level and verbal requirements for the job. Based upon these ratings, we selected 12 jobs that most clearly fit into the following four categories as seen in Table 1: (a) high skill/high verbal, (b) high skill/low verbal, (c) low skill/high verbal, and (d) low skill/low verbal. In Experiment 1, each of these 12 items were rated on a scale ranging from 1 (definitely not hire) to 7 (definitely hire). Hirability scores for each of the four categories of jobs were calculated by obtaining the mean ratings for the three jobs in that category. For instance, the mean rating for the items Dishwasher, File clerk, and Meter attendant was calculated to measure hirability for jobs requiring low skill and low verbal ability.

3.3 Results

3.3.1 Evaluations of the actor

A series of paired samples t-tests examined the differences in perception between synthesized and natural speech. When the actors used an SGD they were rated higher in terms of credibility, t(40) = 5.92, p < 0.001, r = 0.68, and speaker strength, t(40) = 6.79, p < 0.001, r = 0.73. There was no significant effect on speaker informedness.

3.3.2 Ratings of hirability

Each of the dependent measures was analyzed with a 2 (speech type: natural vs. SGD)×2 (skill: high skill vs. low skill)×2 (verbal: high verbal vs. low verbal) MANOVA. This revealed a statistically significant interaction between type of speech and job skill, F(1, 40) = 63.40, p < 0.001, η= 0.78. As seen in Fig. 3 actors using an SGD were rated as more hirable for highly skilled jobs, while conversely, actors using their own voices were rated more hirable for lower skilled jobs.

There was also a significant interaction between type of speech and verbal requirements of the job, F(1, 40) = 14.42, p < 0.001, η= 0.51. Actors using an SGD were favored for highly verbal jobs. Although there was not a significant interaction between all three variables, the data suggests that this was particularly the case when the job was highly skilled as well. There was also a significant interaction between job skill and verbal requirements of the job, F(1, 40) = 17.20, p < 0.001, η= 0.55.

4 Experiment 2

Experiment 2 sought to examine whether the extent to which a job was interpersonally interactive would affect the already established preference for SGD output over natural speech. For this purpose, we designed a new hirability questionnaire in which jobs varied in terms of skill and interactivity required. This yielded a 2 (speech type: natural vs. SGD)×2 (skill level: high skill vs. low skill)×3 (interactivity: FTF vs. mediated vs. non-interactive) factorial design.

4.1 Participants and design

A total of 57 undergraduates (21 men, 36 women, age 16 to 38, with a mean age of 19.26 years) from the University of Pittsburgh at Johnstown. Students participated in the study in exchange for course credit. The 57 participants were then randomly assigned to one of the four sequences, creating three groups with 14 participants and one group with 15.

4.2 Procedure

The procedure for Experiment 2 was nearly identical to Experiment 1. The only difference was the use of an 18 item hirability questionnaire designed to measure perceptions of whether the actors were rated as hirable for jobs varying in skill level and interactivity.

To develop the hirability instrument, a pilot questionnaire with 36 jobs was distributed to the 19 students of an undergraduate psychology class. The students rated each job in terms of skill level (1 = Very Low Skill to 5 = Very High Skill) and interactivity (1 = interactive; FTF, 2 = interactive-mediated; telephone or computer, 3 = non-interactive). Based upon the ratings of the 36 jobs presented, we selected 18 jobs that most clearly fit into the following six categories as seen in Table 2: (a) high skill/interactive, (b) high skill/interactive-mediated (c) high skill/non-interactive, (d) low skill/interactive, (e) low skill/interactive-mediated, and (f) low skill/non-interactive. As in Experiment 1, participants rated each of these 18 items were rated on a scale ranging from 1 (definitely not hire) to 7 (definitely hire).

4.3 Results

4.3.1 Evaluations of the actor

A series of paired samples t-tests examined the differences in perception between synthesized and natural speech. When the actors used an SGD they were rated higher in terms of informedness, t(56) = 6.39, p < 0.001, r = 0.65, credibility, t(56) = 2.88, p < 0.01, r = 0.36, and speaker strength, t(56) = 5.77, p < 0.001, r = 0.61.

4.3.2 Ratings of hirability

Each of the dependent measures was analyzed with a 2 (speech type: natural vs. synthesized)×2 (skill: high skill vs. low skill)×3 (interactivity: interactive vs. mediated vs. non-interactive) MANOVA. There was a statistically significant interaction between type of speech and job skill, F(1, 56) = 41.61, p < 0.001, η= 0.65. As seen in Fig. 4, and consistent with the findings from Experiment 1, actors using an SGD were favored for highly skilled jobs while actors using their own voices were favored for low skilledjobs.

There was also a significant interaction between type of speech and job interactivity, F(2, 112) = 28.07, p < 0.001, η= 0.58. Actors using an SGD are endorsed more highly than actors using natural speech for jobs requiring mediation while actors using natural were endorsed more favorably for non-interactive jobs. There was also a three-way interaction between type of speech, skill level, and interactivity. Actors using an SGD are endorsed more highly than actors using natural speech for interactive jobs that were highly skilled. Natural speech was most preferred over SGD output for non-interactive, low skilled jobs, F(2, 112) = 9.67, p < 0.001, η= 0.38.

5 Experiment 3

Experiment 3 sought to examine whether the particular way in which people needed to interact on the job (e.g., FTF versus telephone) would affect the already established preference for SGD output over natural speech. This yielded a 2 (speech type: natural vs. SGD)×2 (skill level: high skill vs. low skill)×3 (mode of interaction: FTF, CMC, and Telephone Communication) factorial design.

5.1 Participants and design

A total of 43 undergraduates (32 men, 10 women, 1 unspecified, age 17 to 22, with a mean age of 18.83 years). Students participated in the study in exchange for course credit. The 43 participants were then randomly assigned to one of the four sequences, creating three groups with 11 participants and one groupwith 10.

5.2 Procedure

The procedure for the Experiment 3 was nearly identical to Experiment 2. The only difference was the use of a 12 item hirability questionnaire designed to measure perceptions of whether the actors were rated as hirable for jobs varying in skill level and mode of interaction.

To develop a hirability instrument, a pilot questionnaire with 36 jobs was distributed to the 19 students of an undergraduate psychology. The students rated each job in terms of skill level and mode of interaction (FTF, CMC, and Telephone Communication). Based upon the ratings of the 36 jobs presented, we selected 12 jobs that most clearly fit into the following six categories as seen in Table 3: (a) high skill/FTF, (b) high skill/CMC, (c) High Skill/Telephone, (d) Low Skill/FTF, (e) Low Skill/CMC, and (f) Low Skill/Telephone.

5.3 Results

5.3.1 Evaluations of the actor

A series of paired samples t-tests examined the differences in perception between synthesized and natural speech. When the actors used an SGD they were rated higher in terms of informedness, t(41) = 7.27, p < 0.001, r = 0.75, and speaker strength, t(40) = 7.59, p < 0.001, r = 0.77. There was no significant effect on speaker credibility.

5.3.2 Ratings of hirability

Each of the dependent measures was analyzed with a 2 (speech type: natural vs. SGD)×2 (skill: high skill vs. low skill)×3 (mode of interactivity: FTF vs CMC vs Telephone) MANOVA. There was a statistically significant interaction between type of speech and job skill, F(1, 82) = 18.43, p < 0.001, η= 0.43. When the actors used their natural voices, they were rated as less hirable for highly skilled jobs. There was also a significant interaction between type of speech and mode of interactivity, F(2, 82) = 21.72, p < 0.001, η= 0.59. Actors using an SGD were seen as more hirable for jobs involving computer-mediated communication than actors using their own voice. In comparison, there were only small differences between speech conditions for perceptions of hirability of jobs involving FTF or telephone communication. There was also an interaction between type of speech, skill level, and mode of interactivity. As seen in Fig. 5, the above effect was most pronounced for high skilled jobs, F (2, 82) = 26.11, p < 0.001, η= 0.62. People using SGDs were seen as more hirable for low skilled FTF jobs but less hirable for high skilled FTF jobs.

6 Conclusion

Across a series of three experiments, when actors used an SGD to deliver a persuasive appeal, they were rated as more hirable for skilled work than when the same actors used their natural speaking voices. In Experiment 1, using an SGD resulted in higher ratings of hirability for jobs requiring verbal ability. In Experiment 2, using an SGD resulted in higher ratings of hirability for jobs requiring mediated interactivity and less hirable for jobs requiring no interactivity. In Experiment 3 using an SGD resulted in higher ratings for jobs requiring CMC.

Some of these findings are intuitive. It does not come as a shock that a person with a speaking disability would be endorsed for work requiring CMC more so than they would be for a job requiring FTF or telephone communication.

6.1 SGD preferable for skilled and verbal jobs

The most potentially surprising result is that when the actors used an SGD, they were rated as more hirable for jobs requiring high levels of skill and verbal ability. This is a curious finding insofar as the use of SGDs is not typically seen as consistent with verbal interaction. While this finding is exploratory we do have some tentative explanations to offer.

One possible explanation is that SGDs foster a better disability-job fit [10]. When used as an assistive technology, an SGD should objectively increase the ability of the user to engage in employment that requires speaking. From this frame of reference, our findings that suggest that naïve listeners saw the SGD users as more fit for the verbal jobs than the same people when they did not use the technology. This finding is encouraging for the field of assistive technology in general.

Another possible explanation for our findings is that mild speech impairment is stigmatizing and leads to a prejudice against the actors when they used their natural voices. Our findings as well as findings from previous comparisons of dysarthric speech to SGD output [33, 34] demonstrate that listeners prefer SGD output to dysarthric speech. The present research extends these findings to perceptions of hirability; they were consistently seen as less qualified for the skilled jobs. Further exploration should better examine the roots of this bias. We cannot tell if the actors using their dysarthric natural speech are seen as less capable, less intelligent, or in some other way less qualified for verbal and skilled employment than the actors using SGDs.

Insofar as our participants were university undergraduates, there might be concern that young people are biased in favor of technology and gadgetry. Our data, however, is consistent with previous research conducted on adult samples recruited in their community [33], and we do not feel that the finding represents a technology bias in our sample.

6.2 Limitations

In assessing the extent to which findings can be generalized beyond the laboratory, we need to consider whether the effects are statistically significant (i.e., unlikely to be due to chance) as well robust enough to be of practical importance [45]. The r based effect sizes (r and η) we reported for our statistically significant results ranged 0.36 to 0.78 with an average effect size of 0.60. These are substantial effects. Nevertheless, the robustness of the effect in the lab does not guarantee practical or clinical significance of the findings. That is, we cannot state with certainty that the effects detected in the lab would translate into effects in everyday life.

In determining if research results are directly applicable to clinical practice, we also consider both mundane and social realism [38]. The persuasive appeals, for example, were not typical of a hiring situation. Furthermore, video clips are not equivalent to typical interpersonal interactions and do not necessarily depict how the technology is used by people with complex communication needs. Our actors, for instance, used the SGD at a faster pace than might be expected in a real life situation [20 –22]. Furthermore, the undergraduate participants had neither the experiences nor motivations that interviewers have in bona fide hiring situations. A far more involved study would need to be designed to consider such factors as an organization’s hiring practices, history of granting accommodations, [46], understanding of the rights of people with disabilities [47] and appreciation of diversity [6]. Future research would benefit from examining research participants who have more job experience, on average, than undergraduatestudents.

In comparison to many of the qualitative reports frequently published in periodicals dedicated to speech disorders and SGDs, controlled experimental research on perceptions of SGDs technology and SGD users [35, 48] has less mundane and social realism. This type of research needs to be recognized as complementary and, while not directly applicable in the clinic, can guide as well as verify qualitative, field based studies that have higher social validity. Laboratory studies on reactions to SGDs can be useful in confirming observations from qualitative studies as well as to driving inquiry by uncovering issues that deserve further study and bear closer examination in the field. Further research needs to be done in order to see if the findings in the present study align with how users of SGDs find themselves to be treated when they are being considered for employment.

Finally, in our effort to standardize participants’ experience in the lab, our videotaped stimuli represented a very narrow range of disability, only one older SGD system, and limited and narrow use of the technology that might not be fully realistic. Further research would be needed to ascertain whether our findings would be the same for people with greater challenges communicating. While our data cannot speak to how participants would respond to people with truly significant speech impairments we do feel safe in assuming that SGD output would bepreferable to the natural speech of people with severely impaired speech.

6.3 Implications and future directions

One of the theoretical implications of this research is in terms of the more general relationship between the job-fit model and the use of assistive technology (of which an SGD is only one type). Our data provides an illustrative case in which by using a technology that compensates for a person’s disability, they are rated as more hirable than they would otherwise have been rated. Using an SGD, in this case, can be seen as ameliorating the job-fit discrepancies [9, 10] that may account to some extent for job discrimination against persons with disabilities. Hence, our data supports the contention that assistive technologies such as SGDs can make it possible for persons with disabilities to be better able to fully participate in the workworld [29].

On a more applied level, our data has implications for the application of anti-discrimination policies such as the Americans with Disabilities Act. Our data provides added support for the full acceptance of SGDs as a form of reasonable accommodation [19]. Our data also dovetails with reports that suggest that some users of SGDs favor the development of SGDs that interface with other communications technologies. Likewise, it is useful information for vocational counselors and other professionals who work with SGDs, as well as users of SGDs, to recognize that the technology, when used by people with disabilities, appears to be favored over not using the technology.

Our findings can be interpreted as suggesting that when seeking employment, people with complex communication needs may be well-advised to choose to use an SGD to communicate, even if they might be able to communicate adequately without it. The exploratory nature of our data precludes a concrete recommendation. This question does, again, raise the issue of social validity [38]. More research needs to be done to examine the extent to which people who use SGDs would feel comfortable and effective at communicating in a hiring situation while using (or not using) their devices.

Finally, it is clear that we can improve the mundane realism in experimental research on hiring and SGDs. Human resources materials used for hiring in actual organizations would provide clearer, more substantial assessments. Recruiting samples of research participants who have had experience hiring and training in the workplace would also enhance thegeneralizability of laboratory based research. Well conducted lab studies in conjunction with robust qualitative work is a well-worn path in the social and behavioral sciences that can build the best understanding of SGDs and other disability related issues.

Conflict of interest

The authors have no conflict of interest to report.

Footnotes

Acknowledgments

The University of Pittsburgh’s Council for Research Development Fund provided financial support for this work. The authors also acknowledge staff and consumers at the Hiram G. Andrews Center in Johnstown, Pennsylvania for their assistance on this project.

References

Bureau U. Census.gov [Internet]. Census.gov. 2015 [cited 3 December 2015]. http://www.census.gov.

Nelson

. Job-related social networks and communication technology. Augment Altern Commun 2006;22(1):1–9.

Bruyère

, Erickson

, VanLooy

. Comparative study of workplace policy and practices contributing to disability nondiscrimination. Rehabil Psychol 2004;49(1):28–38.

Corrigan

. The stigma of disease and disability. Washington, DC: American Psychological Association; 2014.

Gouvier

, Sytsma-Jordan

, Mayville

. Patterns of discrimination in hiring job applicants with disabilities: The role of disability type, job complexity, and public contact. Rehabil Psychol 2003;48(3):175–81.

McLaughlin

, Bell

, Stringer

. Stigma and acceptance of persons with disabilities: Understudied aspects of workforce diversity. Group Organ Manage 2004;29(3):302–33.

Neufeldt

, Mathieson

. Empirical dimensions of discrimination against disabled people. Health Hum Rights 1995;1(2):174.

Ren

, Paetzold

, Colella

. A meta-analysis of experimental studies on the effects of disability on human resource judgments. Hum Resour Manage R 2008;18(3):191–203.

Colella

, DeNisi

, Varma

. The impact of ratee’s disability on performance judgments and choice as partner: The role of disability-job fit stereotypes and interdependence of rewards. J Appl Psychol 1998;83(1):102–11.

10.

Colella

, Varma

. Disability-job fit stereotypes and the evaluation of persons with disabilities at work. J Occup Rehabil 1999;9(2):79–95.

11.

Louvet

. Social judgment toward job applicants with disabilities: Perception of personal qualities and competences. Rehabil Psychol 2007;52(3):297–303.

12.

, Roberts

, Zhang

. Employer attitudes towad workers with disabilities:A review of research in the past decade. J Vocat Rehabil 2013;38:113–23.

13.

Brady

, Clark

, Dickson

, Paton

, Barbour

. The impact of stroke-related dysarthria on social participation and implications for rehabilitation. Disabil Rehabil 2011;33(3):178–86.

14.

Cruice

, Worrall

, Hickson

. Quantifying aphasic people’s social lives in the context of non-aphasic peers. Aphasiology 2006;20(12):1210–25.

15.

Hilari

, Northcott

. Social support in people with chronic aphasia. Aphasiology 2006;20(1):17–36.

16.

Weitzel

. Overcoming loss of voice. In: Braithwaite

, Thompson

, ed. by. Handbook of communication and people with disabilities: Research and application 1st ed. Mahwah, NJ: Erlbaum; 2000. pp. 451–66.

17.

Hartelius

, Elmberg

, Holm

, Lövberg

A-S

, Nikolaidis

. Living with dysarthria: Evaluation of a self-report questionnaire. Folia Phoniatr Logop 2008;60(1):11–19.

18.

Stancliffe

, Larson

, Auerbach

, Engler

, Taub

, Lakin

. Individuals with intellectual disabilities and augmentative and alternative communication: Analysis of survey data on uptake of aided AAC, and loneliness experiences. Augment Altern Commun 2010;26(2):87–96.

19.

Isetti

, Eadie

. The Americans With Disabilities Act and voice disorders: Practical guidelines for voice clinicians. J Voice 2015.

20.

Friginal

, Pearson

, Di

Ferrante L

, Pickering

, Bruce

. Linguistic characteristics of AAC discourse in the workplace. Discourse Stud 2013;15(3):279–98.

21.

McNaughton

, Bryen

. AAC technologies to enhance participation and access to meaningful societal roles for adolescents and adults with developmental disabilities who require AAC. Augment Altern Commun 2007;23(3):217–29.

22.

McNaughton

, Chapple

. AAC and Communication in the Workplace. Perspect Augment Altern Commun 2013;22(1):30.

23.

Bryen

, Potts

, Carey

. So you want to work? What employers say about job skills, recruitment and hiring employees who rely on AAC. Augment Altern Commun 2007;23(2):126–39.

24.

McNaughton

, Light

, Groszyk

. “Don’t give up”: Employment experiences of individuals with amyotrophic lateral sclerosis who use augmentative and alternative communication. Augment Altern Commun 2001;17(3):179–95.

25.

McNaughton

, Symons

, Light

, Parsons

. “My dream was to pay taxes”: The self-employment experiences of individuals who use augmentative and alternative communication. J Vocat Rehabil 2006;25:181–96.

26.

DeRuyter

, McNaughton

, Caves

, Bryen

, Williams

. Enhancing AAC connections with the world. Augment Altern Commun 2007;23(3):258–70.

27.

Higginbotham

, Shane

, Russell

, Caves

. Access to AAC: Present, past, and future. Augment Altern Commun 2007;23(3):243–57.

28.

McNaughton

, Light

, Gulla

. Opening up a’ whole new world’: Employer and co-worker perspectives on working with individuals who use augmentative and alternative communication. Augment Altern Commun 2003;19(4):235–53.

29.

Williams

, Krezman

, McNaughton

. “Reach for the stars”: Five principles for the next 25 years of AAC. Augment Altern Commun 2008;24(3):194–206.

30.

Stern

, Mullennix

, Wilson

. Effects of perceived disability on persuasiveness of computer-synthesized speech. J Appl Psychol 2002;87(2):411–7.

31.

Mayo

, Clark

, King

. Listeners’ weighting of acoustic cues to synthetic speech naturalness: A multidimensional scaling analysis. Speech Commun 2011;53(3):311–26.

32.

Stern

, Dumont

, Mullennix

, Winters

. Positive prejudice toward disabled persons using synthesized speech: Does the effect persist across contexts? J Lang Soc Psychol 2007;26(4):363–80.

33.

Drager

KDR

, Hustad

, Gable

. Telephone communication: Synthetic and dysarthric speech intelligibility and listener preferences. Augment Altern Commun 2004;20(2):103–12.

34.

Stern

, Chobany

, Patel

, Tressler

. Listeners’ preference for computer-synthesized speech over natural speech of people with disabilities. Rehabil Psychol 2014;59(3):289–97.

35.

McCarthy

, Light

. Attitudes toward individuals who use augmentative and alternative communication: Research review. Augment Altern Commun 2005;21(1):41–55.

36.

McNaughton

, Richardson

. Supporting positive employment outcomes for individuals with autism who use AAC. Perspect Augment Altern Commun 2013;22(3):164.

37.

Paterson

, Carpenter

. Using different methods to communicate: How adults with severe acquired communication difficulties make decisions about the communication methods they use and how they experience them. Disabil Rehabil 2015;37(17):1522–30.

38.

Schlosser

. The efficacy of augmentative and alternative communication. Amsterdam: Academic Press; 2003.

39.

Hammen

, Yorkston

, Minife

. Effects of temporal alterations on speech intelligibility in parkinsonian dysarthria. J Speech Hear Res 1994;37:244–53.

40.

Yorkston

, Strand

, Kennedy

MRT

. Comprehensibility of dysarthric speech: Implications for assessment and treatment planning. Am J Speech Lang Pathol 1996;5:55–66.

41.

Petty

, Cacioppo

. Communication and persuasion. New York: Springer Verlag; 1986.

42.

Stern

. Computer-synthesized speech and perceptions of the social influence of disabled users. J Lang Soc Psychol 2008;27(3):254–65.

43.

Lucia

. The effects of speech rate and speaker-listener congruence on persuasion [Masters]. Wayne State University; 1998.

44.

Leathers

. Successful nonverbal communication. Boston: Allyn and Bacon; 1997.

45.

Rosenthal

, Rosnow

. Essentials of behavioral research 3rd Ed. New York: McGraw-Hill; 2008.

46.

Carpenter

, Paetzold

. An examination of factors influencing responses to requests for disability accommodations. Rehabil Psychol 2013;58(1):18–27.

47.

Chism

, Satcher

. Human resource management students’ perceptions of employment and individuals with disabilities. J Appl Rehabil Couns 1997;28:38–42.

48.

McCarthy

, Donofrio-Horwitz

, Smucker

. The effects of reading personal narratives written by an individual who uses AAC on the attitudes of pre-professionals in business. Augment Altern Commun 2010;26(2):61–74.