Skip to main content
NCBI home page
Journal of Speech, Language, and Hearing Research : JSLHR logoLink to Journal of Speech, Language, and Hearing Research : JSLHR
. 2022 Apr 20;65(6):2391–2397. doi: 10.1044/2022_JSLHR-21-00557

Speech Recognition in Noise Performance Measured Remotely Versus In-Laboratory From Older and Younger Listeners

Jing Shen a,✉, Jingwei Wu b
PMCID: PMC9567433  PMID: 35442717

Abstract

Purpose:

This study examined the performance difference between remote and in-laboratory test modalities with a speech recognition in noise task in older and younger adults.

Method:

Four groups of participants (younger remote, younger in-laboratory, older remote, and older in-laboratory) were tested on a speech recognition in noise protocol with 72 sentences.

Results:

While the younger remote group performed more poorly than the younger in-laboratory group, older participants' performance was comparable between the two modality groups, particularly in the easy to moderately difficult conditions. These results persisted after controlling for demographic variables (e.g., age, gender, and education).

Conclusion:

While these findings generally support the feasibility of remote data collection with older participants for research on speech perception, they also suggest that technological proficiency is an important factor that affects performance on remote testing in the aging population.

Over the past decade, the use of remote testing protocols in behavioral research has rapidly increased (Archibald et al., 2019; Sassenberg & Ditrich, 2019). Collecting data remotely has many advantages, including time/cost efficiency, larger sample size, better generalizability, access to special populations, and increased sample diversity. Remote protocols, however, have not been widely adopted in hearing research. Since the onset of COVID-19, identifying strategies and limitations in remote data collection has become a pressing concern for hearing researchers due to the urgent need to collect data remotely. While results from auditory research are largely mixed with respect to comparing performances on auditory tasks between remote and in-laboratory testing environments (Acoustical Society of America, Psychological and Physiological Acoustics Task Force, 2020), data comparing participants' performance using the same protocol in-laboratory versus remotely are needed to ensure effective employment of remote protocols in auditory research.

Remote Testing in Speech Perception Research

In the field of speech perception, only a few studies have compared participants' performance between in-laboratory and remote environments (Cooke et al., 2011; Cooke & García Lecumberri, 2021; Mayo et al., 2012; Slote & Strand, 2016; Wolters et al., 2010). While the performance patterns across conditions are largely consistent between in-laboratory and remote protocols, several studies found a penalty in performance measured using remote protocols. In one of the earliest studies showing this performance gap, Wolters et al. (2010) measured recognition accuracy of synthesized sentences from 167 listeners recruited through Amazon Mechanical Turk and 20 in-laboratory listeners with the same testing paradigm. While the remote group was more sensitive to the differences between stimuli conditions, speech intelligibility of synthesized sentences showed a 13% error rate for the in-laboratory group as compared to 20% for the remote group. A few factors have been proposed to potentially explain this performance gap between in-laboratory and remote modalities, including inability to control equipment or environment, uncertainty in participant characteristics, and difficulty monitoring participants' engagement (e.g., Cooke & García Lecumberri, 2021). With respect to equipment and environment, in-laboratory hearing studies use high-quality sound cards and headphones, which provide a tight control of the sound level and quality, and testing protocols are usually implemented in a sound attenuated booth or a quiet room. In contrast, there is limited control over equipment and environment with remote testing, particularly when web-based participant recruitment platforms (e.g., Amazon Mechanical Turk, Prolific.co) are used. It is difficult to verify participants' use of headphones, although headphone screening tests (Milne et al., 2021; Woods et al., 2017) can be helpful. The testing environment also varies substantially in terms of background noise, distractions, and interruptions. Concerning participant characteristics, while utilizing samples recruited remotely does have the advantage of stronger diversity, verification of participants' hearing and language backgrounds can be challenging. Lastly, it is easier to monitor participants' attention to the task when participants come into the lab and work with a tester; it is more difficult to monitor attention in remote protocols.

In-Laboratory Versus Remote Data Collection With Older Participants

While there are fewer studies using remote protocols with older participants, to date, this research is mostly from health care fields and includes remote protocols such as computer-based surveys and questionnaires (Remillard et al., 2014; Thorén et al., 2012), eHealth interventions (Abrams et al., 2015; Nahm et al., 2011; Polsinelli et al., 2020), and cognitive assessments (Binng et al., 2020). However, the challenges of remote data collection with older participants are not trivial. A few identified barriers include low familiarity with the technology, difficulty with testing interface (i.e., difficulty with reading or navigating the software), limited access to Internet or technology, and lack of technical support (Nahm et al., 2011; Remillard et al., 2014).

In hearing research, there are additional challenges associated with the presentation of sounds to older participants. Given the high prevalence of undiagnosed age-related hearing loss (30%–60%) in the 65 years and older cohort (Agrawal et al., 2008; Cruickshanks et al., 1998), ensuring audibility of auditory stimuli becomes an additional challenge when implementing hearing research remotely and having limited control over hardware and/or software. As the field is working on protocols and solutions for telehealth and remote research data collection with older adults, data that provide information about how older adults perform on auditory tasks implemented remotely are urgently needed. While several studies have collected data remotely from older listeners with a variety of auditory tasks (e.g., Potgieter et al., 2018; Shafiro et al., 2020; Smits et al., 2006), fewer speech perception studies have compared performances of older and younger participants using remote protocols (Brown et al., 2021; Luthra et al., 2021). The available work typically used Prolific.co for participant recruitment and Gorilla (Anwyl-Irvine et al., 2020) for building the experiment. Although these studies showed consistent data patterns regarding the effects of interest across younger and older groups, all the data were collected remotely and no comparison was made to data collected in-laboratory. Therefore, a critical question remains: For older participants, is there a performance gap due to test modality that is similar to what has been found with younger participants? To answer this question, the primary goal of this study was to compare speech recognition in noise performance of younger and older participants using a self-paced computer-based protocol implemented in either a laboratory setting or remotely.

Factors Contributing to In-Laboratory Versus Remote Performance Gap

Focusing specifically on research concerning speech recognition in noise, a few studies have compared speech recognition in noise across in-laboratory and remote testing modalities (Cooke et al., 2011; Mayo et al., 2012; Slote & Strand, 2016). Although findings from these studies demonstrated the performance gap due to modality, an important next step is to examine the factors contributing to this gap. A recent study by Cooke and García Lecumberri (2021) started this work by comparing speech recognition performance between data from a known listener cohort tested remotely and in-laboratory. Their data revealed a larger performance gap in the more difficult (i.e., less intelligible) noise conditions, but this interaction between modality and condition disappeared in two of their three experiments after controlling for headphone quality. While their method of collecting information about headphone quality (indicated by price) in a group of students at a local university provides useful information, this method may be less reliable when the participants are recruited through web-based platforms and may be from different countries and more diverse populations. Furthermore, the variability in listener and environment characteristics is more substantial in a cohort of participants recruited through web-based platforms than in a known listener cohort. For example, unlike known listener cohorts, the remote samples are usually more diverse in terms of demographic variables than in-laboratory groups. Those variables, such as language acquisition and education attainment, have been shown to have an impact on speech perception in noise (Hartshorne et al., 2018; Knight & Heinrich, 2019; Mayo et al., 1997). With a cohort of participants recruited through web-based platforms, this variability is anticipated to drive the performance gap between remote and in-laboratory groups, on which headphone quality and task difficulty are likely not the primary influence. As a result, the performance gap was expected to persist without controlling for headphone quality but may be attenuated by controlling for demographic variability in the samples. Its interaction with task difficulty, on the other hand, is likely to be nonsignificant with a participant cohort recruited through web-based platforms. Therefore, the second goal of the current study was to examine whether the performance gap between remote and in-laboratory modalities would persist after controlling for demographic variables and task difficulty (as indicated by signal-to-noise ratios [SNRs]).

Method

Participants

Four groups of participants took part in the study: younger remote, younger in-laboratory, older remote, older in-laboratory. Table 1 provides key characteristics of the participants. All participants were native speakers of General American English, except for two younger participants in the remote group whose native language is British English. Due to the strong impact of hearing loss on speech recognition in noise measures, the ratio of normal hearing was approximately matched across remote and in-laboratory groups within same age group (all below 10%, see Table 1). Hearing status of the in-laboratory groups was determined by clinical audiometric testing, while that of the remote groups was self-reported. The two remote groups were recruited through Prolific.co. The two in-laboratory groups were recruited from Temple University and the surrounding community in Philadelphia, PA. All participants were paid for their time. The study protocol was approved by the institutional review board of Temple University.

Table 1.

Key characteristics of the participants.

Remote data received Younger remote (n = 68) Older remote (n = 32) Younger in-laboratory (n = 16) Older in-laboratory (n = 10)
Removal rate due to incompleteness 10% 0%
Final sample
Younger remote (n = 62)
Older remote (n = 32)
Younger in-laboratory (n = 16)
Older in-laboratory (n = 10)
M age (range)
28.6 (18–35) years
68.8 (65–78) years
19.7 (19–21) years
69.4 (60–76) years
Mean education (range) 15.7 (11–22) years 17.5 (10–29) years 13.9 (13–17) years 15.3 (12–19) years
Sex 29 females, 31 males, 1 other, 1 N/A 14 females, 18 males 15 females, 1 male 7 females, 3 males
Rate of normal hearing 98.4% 93.8% 100% 90%
Pass rate of headphone screening 88.7% 87.5%
Open in a new tab

Note. N/A = not applicable.

Materials

While previous studies have used stimuli from sentence corpora, it is worth noting that these sentences are well controlled in terms of syntax and vocabulary and differ from the language used in real-life communication. In an effort to strengthen the ecological validity of the stimuli (Keidser et al., 2020), sentence materials that were derived from real conversations were included and recorded from multiple talkers to measure speech recognition in noise performance.

Seventy-two sentences were selected from the Diapix task (Hazan & Baker, 2011; Van Engen et al., 2010), a dialogue elicitation paradigm. In this task, two conversation partners must work together to find differences between two highly similar pictures while each talker can only see one picture. These sentences were grouped into four lists based on their psycholinguistic properties, including the number of total words, the number of content words, word length, lexical frequency, and phonological neighborhood density (Shen et al., 2022). Sentences were recorded from four female native speakers of American English, who were instructed to read the sentences clearly with natural prosody. The sentences were root-mean-square normalized before being mixed with steady-state noise that was spectrally shaped to match each individual talker's voice. Four SNR of −4, −2, 0, and 2 dB SNRs were chosen to produce a range of recognition performance.

Procedure

This study involved a transcription task and was developed using Qualtrics with embedded audio files, which were stored on Ensemble Video System. For each trial, participants were instructed to click on a button to start playing the sentence and type what they heard in a textbox on the same page. Afterwards, they clicked on another button to go to the next trial. All participants self-administered the test except for two older in-laboratory participants who could not type due to motor issues or unfamiliarity with technology. These two participants responded verbally, and an experimenter typed in their responses.

The testing included four blocks of 18 trials each, with SNR/list combination and order counterbalanced by Latin Square Design. Each participant only listened to speech from one talker, and the distribution of talker in each group was balanced. A 10-s calibration tone of 1 K Hz was played before speech testing to check audibility. The participants were asked to adjust the volume to make the tone comfortably loud. There were two practice sentences before the testing, made up of 72 sentences, began.

Participants in the remote groups completed consent, a list of basic demographic questions, and a headphone screening (Woods et al., 2017) prior to the test. In this test, participants judged which of three 200-Hz pure tones was quietest, with one of the tones 180 degrees out of phase across the stereo channels. Regardless of headphone screening results, participants continued with the protocol. The demographic questions included the participant characteristics that are reported in Table 1. For the in-laboratory groups, the participants were seated in front of a BenQ Zowie LCD monitor and the protocol was run using Google Chrome on a Mac mini desktop computer. Auditory stimuli were played using a RME Fireface UCX sound card and a pair of Sennheiser HD-25 headphones at 65 dBA. In-laboratory participants did not have to complete the headphone screening, and demographic information was collected by an experimenter prior to starting the protocol.

Participants' responses were cleaned manually to remove typographical errors before being scored. The dependent measure was the proportion of correctly recognized keywords in each sentence.

Results

Analyses were performed by using SAS version 9.4. The recognition accuracy data were expressed and plotted as mean ± standard error (see Figure 1). A generalized linear regression model was built with generalized estimation equation method (GEE) to examine the effects of test modality (in-laboratory vs. remote) and age group (younger vs. older) on speech recognition, adjusted by task difficulty levels (as indicated by SNR), time spent on task, and demographic variables (gender, age, education). GEE is a statistical method frequently used for analyzing longitudinal or repeated-measures data sets (Diggle et al., 2002; Zeger & Liang, 1986). This method was selected because it specifies the correlation among speech recognition scores repeatedly measured from the same participant (across task difficulty conditions) as a working covariance structure, so that standard errors of the regression parameters yield more precise estimates in the test of significance. A p value less than .05 was deemed significant in the analyses.

Figure 1.

Figure 1.

Open in a new tab

Speech recognition accuracy (% keywords correct, mean ± standard error) by noise conditions. SNR = signal-to-noise ratio.

Firstly, the results showed a significant interaction effect between test modality and age group (p < .0001, see Table 2). The performance gap between in-laboratory and remote modalities was much larger in the younger groups as compared to the older groups (15.3% in younger group; 1.2% in older group). In other words, younger listeners did significantly better when they were tested in-laboratory than remotely, while older listeners did not show a significant difference when tested in the two modalities. Secondly, there was no significant interaction effect of task difficulty and test modality effect on speech recognition, in either younger or older groups (p > 0.30), meaning the performance gap due to test modality was not modulated by task difficulty. Further, the results persisted after exclusion of participants with hearing loss (3.3%) or failed headphone screening (9.2%).

Table 2.

Parameter estimates from generalized estimation equation method model.

Variable Beta coefficient SE Chi square p value*
Age group (reference: younger) 0.0259 0.3546 0.66 > .1000
Modality (reference: remote) 0.7434 0.1330 19.66 < .0001
Age group × modality −0.6882 0.1745 14.67 < .0010
SNR (reference −4) 110.81 < .0001
 −2 0.6272 0.0961
 0 0.9842 0.0977
 2 1.2433 0.0978
Age −0.0051 0.0088 0.34 > .1000
Education 0.0042 0.0103 0.16 > .1000
Gender (reference: other) 7.28 > .0500
 Female −0.8215 0.3174
 Male −0.7473 0.3180
 N/A −1.6337 0.4140
Time spent on task (min) −0.0049 0.0027 3.19 > .0500
Open in a new tab

Note. SNR = signal-to-noise ratio; N/A = not applicable.

*

Significant Type III test is based on score statistics.

Discussion

This study compared speech recognition in noise performance of younger and older listeners when data were collected remotely versus in-laboratory. With younger participants, the current findings align with the results from participants recruited via Amazon Mechanical Turk (Cooke et al., 2011; Mayo et al., 2012; Slote & Strand, 2016; Wolters et al., 2010), but contrast with recent data (Cooke & García Lecumberri, 2021) from a known listener group (i.e., college students). The latter study showed comparable performances with remote versus in-laboratory protocols. There are a number of explanations for this contrast. First, it is possible that large variability in participants' background and experience in the remote group (and the differences between remote and in-laboratory groups) could contribute to the performance gap. While the performance gap persisted in this study after statistically controlling for several demographic variables (i.e., age, gender, education), future research should include matched samples of participant characteristics for a direct comparison. Second, regarding headphone use, the results of the current study were not affected by headphone screening; over 90% of remote participants in our study passed headphone screening. Considering the results from Cooke and García Lecumberri (2021), variability in headphone quality may have still contributed to the performance gap; this possibility should be addressed by collecting headphone quality measures in future research. In addition, younger remote participants as a group took significantly longer time to complete the task than the in-laboratory group, remote group: M = 29.26 min, SD = 17.25 min; in-laboratory group: M = 20.44 min, SD = 3.04 min; F(1, 76) = 4.04, p < .05, possibly indicating that remote participants took more breaks and/or were more distracted. While statistically controlling for this variable did not affect the performance gap in this study, this issue should be examined in future research with data from attention checks designed specifically to address this issue and with larger samples that are matched in size. Lastly, our data from younger participants did not show any impact of task difficulty on the performance gap between remote and in-laboratory testing.

Regarding older participants, the present findings suggest a different pattern with respect to performance and participant characteristics as compared to the younger groups. While all of the younger participants were comfortable with the test protocol, based on observations of the older participants tested in-laboratory, using a computer to self-administer the test was a challenging task for most of them. They particularly struggled with typing and using the mouse to click on buttons. Two of the 10 participants could not type and asked the experimenter to help with typing their responses. The older in-laboratory group also took longer to complete the study than the older remote group, in-laboratory group: M = 42.00 min, SD = 7.06 min; remote group: M = 34.50 min, SD = 12.40 min; F (1, 40) = 3.16, p = .08. While these observations are consistent with the literature regarding difficulty with technology among the aging population in general (e.g., Nahm et al., 2011), this characteristic of our older in-laboratory group contrasts with that of the remote group. Remote participants recruited from Prolific.co are typically experienced remote research participants and are skillful at navigating remote testing platforms and using computer technology. Furthermore, in contrast to the in-laboratory group's struggle with technology (and unfamiliar equipment), the remote older participants should find the test platform easy to navigate, particularly with a computer and accessories they are familiar with. These factors likely offset the potential remote versus in-laboratory performance gap. As a result, the performance of our older remote group was on par with that of the in-laboratory group in easy to moderately adverse noise conditions, and slightly better in the most adverse condition (with a nonsignificant difference in accuracy, 37.2% vs. 29.1%). In practice, this finding demonstrates the importance of using a test protocol that matches participants' level of technology use to minimize their effort in completing the task. Future research should also consider including a computer proficiency test or questionnaire to measure and control for participants' proficiency with computer technology.

Due to restrictions posed by the COVID-19 pandemic, the sample size of the in-laboratory groups was small in this study. Therefore, the current findings should be considered preliminary. Nevertheless, this study contributes to a fast-growing literature comparing remote and in-laboratory performances on auditory tasks by shedding light on factors that influence performance on self-paced remote speech perception tasks, particularly with older adults. Specifically, these data suggest that across participant groups, performance gaps between remote and in-laboratory groups are due to variability beyond basic demographic variables (e.g., age, gender, education). In older groups, these results indicate that proficiency with computer technology is likely to be a primary factor that affects performance on remote tasks. Older participants recruited through web-based services typically have higher proficiency with technology than those recruited locally. With younger adults, who are likely to have comparable technology proficiency across remote and in-laboratory groups, factors such as headphone quality and attention to the task should be tested as next steps. These preliminary but novel findings should be examined in a future study with a larger and more balanced sample size.

Acknowledgments

The authors thank Reshma Reji and Caitlyn Dececco for their assistance with the data collection and processing; Caitlin Sharkey, Alyssa Liusie, and Lauren Calandruccio for their help with the stimuli preparation; and Sam Yelman for the assistance with the setting up the computer-based testing platform. This work was supported by National Institutes of Health Grant R21DC017560.

Funding Statement

The authors thank Reshma Reji and Caitlyn Dececco for their assistance with the data collection and processing; Caitlin Sharkey, Alyssa Liusie, and Lauren Calandruccio for their help with the stimuli preparation; and Sam Yelman for the assistance with the setting up the computer-based testing platform. This work was supported by National Institutes of Health Grant R21DC017560.

References

  1. Abrams, H. B. , Bock, K. , & Irey, R. L. (2015). Can a remotely delivered auditory training program improve speech-in-noise understanding? American Journal of Audiology, 24(3), 333–337. https://doi.org/10.1044/2015_AJA-15-0002 [DOI] [PubMed] [Google Scholar]
  2. Acoustical Society of America, Psychological and Physiological Acoustics Task Force. (2020). Remote testing Wiki. https://www.spatialhearing.org/remotetesting/
  3. Agrawal, Y. , Platz, E. A. , & Niparko, J. K. (2008, 2008). Prevalence of hearing loss and differences by demographic characteristics among US adults: Data from the National Health and Nutrition Examination Survey, 1999-2004. Archives of Internal Medicine, 168(14), 1522–1530. https://doi.org/10.1001/archinte.168.14.1522 [DOI] [PubMed] [Google Scholar]
  4. Anwyl-Irvine, A. L. , Massonnié, J. , Flitton, A. , Kirkham, N. , & Evershed, J. K. (2020). Gorilla in our midst: An online behavioral experiment builder. Behavior Research Methods, 52(1), 388–407. https://doi.org/10.3758/s13428-019-01237-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Archibald, M. M. , Ambagtsheer, R. C. , Casey, M. G. , & Lawless, M. (2019). Using zoom videoconferencing for qualitative data collection: Perceptions and experiences of researchers and participants. International Journal of Qualitative Methods, 18. https://doi.org/10.1177/1609406919874596 [Google Scholar]
  6. Binng, D. , Splonskowski, M. , & Jacova, C. (2020). Distance assessment for detecting cognitive impairment in older adults: A systematic review of psychometric evidence. Dementia and Geriatric Cognitive Disorders, 49(5), 1–15. https://doi.org/10.1159/000511945 [DOI] [PubMed] [Google Scholar]
  7. Brown, V. A. , Van Engen, K. , & Peelle, J. E. (2021). Face mask type affects audiovisual speech intelligibility and subjective listening effort in young and older adults. Cognitive Research: Principles and Implications, 6(1), 49. https://doi.org/10.31234/osf.io/7waj3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cooke, M. , Barker, J. , Garcia Lecumberri, M. L. , & Wasilewski, K. (2011). Crowdsourcing for word recognition in noise. In Trancoso I. (Ed.), Proceedings of the 12th Annual Conference of the International Speech Communication Association (pp. 3049–3052). International Speech Communication Association. [Google Scholar]
  9. Cooke, M. , & García Lecumberri, M. L. (2021). How reliable are online speech intelligibility studies with known listener cohorts? The Journal of the Acoustical Society of America, 150(2), 1390–1401. https://doi.org/10.1121/10.0005880 [DOI] [PubMed] [Google Scholar]
  10. Cruickshanks, K. J. , Wiley, T. L. , Tweed, T. S. , Klein, B. E. , Klein, R. , Mares-Perlman, J. A. , & Nondahl, D. M. (1998). Prevalence of hearing loss in older adults in Beaver Dam, Wisconsin: The epidemiology of hearing loss study. American Journal of Epidemiology, 148(9), 879–886. https://doi.org/10.1093/oxfordjournals.aje.a009713 [DOI] [PubMed] [Google Scholar]
  11. Diggle, P. , Diggle, P. J. , Heagerty, P. , Liang, K. Y. , & Zeger, S. (2002). Analysis of longitudinal data. Oxford University Press. [Google Scholar]
  12. Hartshorne, J. K. , Tenenbaum, J. B. , & Pinker, S. (2018). A critical period for second language acquisition: Evidence from 2/3 million English speakers. Cognition, 177, 263–277. https://doi.org/10.1016/j.cognition.2018.04.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Hazan, V. , & Baker, R. (2011). Acoustic-phonetic characteristics of speech produced with communicative intent to counter adverse listening conditions. The Journal of the Acoustical Society of America, 130(4), 2139–2152. https://doi.org/10.1121/1.3623753 [DOI] [PubMed] [Google Scholar]
  14. Keidser, G. , Naylor, G. , Brungart, D. S. , Caduff, A. , Campos, J. , Carlile, S. , Carpenter, M. G. , Grimm, G. , Hohmann, V. , Holube, I. , Launer, S. , Lunner, T. , Mehra, R. , Rapport, F. , Slaney, M. , & Smeds, K. (2020). The quest for ecological validity in hearing science: What it is, why it matters, and how to advance it. Ear and Hearing, 41(Supplement 1), 5S–19S. https://doi.org/10.1097/AUD.0000000000000944 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Knight, S. , & Heinrich, A. (2019). Visual inhibition measures predict speech-in-noise perception only in people with low levels of education. Frontiers in Psychology, 9, 2779. https://doi.org/10.3389/fpsyg.2018.02779 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Luthra, S. , Peraza-Santiago, G. , Saltzman, D. , Crinnion, A. M. , & Magnuson, J. (2021). Lexically-mediated compensation for coarticulation in older adults. Proceedings of the Annual Meeting of the Cognitive Science Society (Vol. 43, No. 43). [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Mayo, C. , Aubanel, V. , & Cooke, M. (2012). Effect of prosodic changes on speech intelligibility (1706–1709). 13th Annual Conference of the International Speech Communication Association. Interspeech 2012, 9–13 September 2012, Portland, Oregon. [Google Scholar]
  18. Mayo, L. H. , Florentine, M. , & Buus, S. (1997). Age of second-language acquisition and perception of speech in noise. Journal of Speech, Language, and Hearing Research, 40(3), 686–693. https://doi.org/10.1044/jslhr.4003.686 [DOI] [PubMed] [Google Scholar]
  19. Milne, A. E. , Bianco, R. , Poole, K. C. , Zhao, S. , Oxenham, A. J. , Billig, A. J. , & Chait, M. (2021). An online headphone screening test based on dichotic pitch. Behavior Research Methods, 53(4), 1551–1562. https://doi.org/10.3758/s13428-020-01514-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Nahm, E. S. , Bausell, B. , Resnick, B. , Covington, B. , Brennan, P. F. , Mathews, R. , & Park, J. H. (2011). Online research in older adults: lessons learned from conducting an online randomized controlled trial. Applied Nursing Research, 24(4), 269–275. https://doi.org/10.1016/j.apnr.2009.09.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Polsinelli, A. J. , Kaszniak, A. W. , Glisky, E. L. , & Ashish, D. (2020). Effects of a brief, online, focused attention mindfulness training on cognition in older adults: A randomized controlled trial. Mindfulness, 11(5), 1182–1193. https://doi.org/10.1007/s12671-020-01329-2 [Google Scholar]
  22. Potgieter, J. M. , Swanepoel, W. , Myburgh, H. C. , & Smits, C. (2018). The South African English smartphone digits-in-noise hearing test: Effect of age, hearing loss, and speaking competence. Ear and Hearing, 39(4), 656–663. https://doi.org/10.1097/AUD.0000000000000522 [DOI] [PubMed] [Google Scholar]
  23. Remillard, M. L. , Mazor, K. M. , Cutrona, S. L. , Gurwitz, J. H. , & Tjia, J. (2014). Systematic review of the use of online questionnaires of older adults. Journal of the American Geriatrics Society, 62(4), 696–705. https://doi.org/10.1111/jgs.12747 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Sassenberg, K. , & Ditrich, L. (2019). Research in social psychology changed between 2011 and 2016: Larger sample sizes, more self-report measures, and more online studies. Advances in Methods and Practices in Psychological Science, 2(2), 107–114. https://doi.org/10.1177/2515245919838781 [Google Scholar]
  25. Shafiro, V. , Hebb, M. , Walker, C. , Oh, J. , Hsiao, Y. , Brown, K. , Sheft, S. , Li, Y. , Vasil, K. , & Moberly, A. C. (2020). Development of the basic auditory skills evaluation battery for online testing of cochlear implant listeners. American Journal of Audiology, 29(3S), 577–590. https://doi.org/10.1044/2020_AJA-19-00083 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Shen, J. , Calandruccio, L. , Liusie, A. , DeCecco, C. (2022). The influence of linguistic properties on recognition of naturalistic speech in noise [Poster presentation] . Speech In Noise Workshop, Virtual Meeting. https://2022.speech-in-noise.eu/ [Google Scholar]
  27. Slote, J. , & Strand, J. F. (2016). Conducting spoken word recognition research online: Validation and a new timing method. Behavior Research Methods, 48(2), 553–566. https://doi.org/10.3758/s13428-015-0599-7 [DOI] [PubMed] [Google Scholar]
  28. Smits, C. , Merkus, P. , & Houtgast, T. (2006). How we do it: The Dutch functional hearing-screening tests by telephone and Internet. Clinical Otolaryngology, 31(5), 436–440. https://doi.org/10.1111/j.1749-4486.2006.01195.x [DOI] [PubMed] [Google Scholar]
  29. Thorén, E. S. , Andersson, G. , & Lunner, T. (2012). The use of research questionnaires with hearing impaired adults: online vs. paper-and-pencil administration. BMC Ear, Nose and Throat Disorders, 12(1), 1–6. https://doi.org/10.1186/1472-6815-12-12 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Van Engen, K. J. , Baese-Berk, M. , Baker, R. E. , Choi, A. , Kim, M. , & Bradlow, A. R. (2010). The Wildcat Corpus of native-and foreign-accented English: Communicative efficiency across conversational dyads with varying language alignment profiles. Language and Speech, 53(4), 510–540. https://doi.org/10.1177/0023830910372495 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Wolters, M. , Isaac, K. , & Renals, S. (2010). Evaluating speech synthesis intelligibility using Amazon's Mechanical Turk. Proc. 7th Speech Synthesis Workshop (SSW7). [Google Scholar]
  32. Woods, K. J. , Siegel, M. H. , Traer, J. , & McDermott, J. H. (2017). Headphone screening to facilitate web-based auditory experiments. Attention, Perception, & Psychophysics, 79(7), 2064–2072. https://doi.org/10.3758/s13414-017-1361-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Zeger, S. L. , & Liang, K. Y. (1986). Longitudinal data analysis for discrete and continuous outcomes. Biometrics, 42(1), 121–130. https://doi.org/10.2307/2531248 [PubMed] [Google Scholar]

Articles from Journal of Speech, Language, and Hearing Research : JSLHR are provided here courtesy of American Speech-Language-Hearing Association

RESOURCES