The DOI matches its Crossref record, with a small difference in title or year; 31 of the 31 references it cites were found in a registry or an index; 5 of the 26 sources its claims cite were checked for support.
Typing Performance of Blind Users
Open accessAbstract
Non-visual text-entry for people with visual impairments has focused mostly on the comparison of input techniques reporting on performance measures, such as accuracy and speed. While researchers have been able to establish that non-visual input is slow and error prone, there is little understanding on how to improve it. To develop a richer characterization of typing performance, we conducted a longitudinal study with five novice blind users. For eight weeks, we collected in-situ usage data and conducted weekly laboratory assessment sessions. This paper presents a thorough analysis of typing performance that goes beyond traditional aggregated measures of text-entry and reports on character-level errors and touch measures. Our findings show that users improve over time, even though it is at a slow rate (0.3 WPM per week). Substitutions are the most common type of error and have a significant impact on entry rates. In addition to text input data, we analyzed touch behaviors, looking at touch contact points, exploration movements, and lift positions. We provide insights on why and how performance improvements and errors occur. Finally, we derive some implications that should inform the design of future virtual keyboards for non-visual input.
Overview
CoChat Summary
## TL;DR
This paper reports a two-month longitudinal study of five novice blind smartphone users, combining in-situ logging with weekly lab text-entry trials on a touchscreen QWERTY keyboard. Users improved over time, but slowly: average speed rose from 1.6 WPM to 4 WPM, and the paper argues that gains came from better spatial models of the keyboard and more efficient touch exploration.
## Key findings
- Participants improved on average by 2.4 WPM over eight weeks, from 1.6 WPM in week 1 to 4 WPM in week 8; the effect of Week on WPM was significant (F1,7=12.329, p<.001), with an average learning rate of 0.3 WPM per week.
- Total error rates fell from 26% (SD=11.7%) to 7.4% (SD=1.7%) over eight weeks (F1,7=4.176, p<.001); uncorrected error rates averaged 1.6% by week 8.
- Substitutions were the most common error type: they decreased from 24% (SD=12%) to 6% (SD=1%) but remained higher than insertions and omissions (F2,8=125.321, p<.001).
- Touch behavior also improved: by week 8, users landed within the intended key boundaries 48% of the time and within the intended or adjacent key 91% of the time; visited keys per keystroke dropped from 4.9 to 2, and target re-entries from 6.6 to 0.8.
## Methods
- Longitudinal study with 5 novice blind users (4 male, 1 female), recruited from a local training institution; all were legally blind, experienced desktop screen reader users, and had no prior smartphone or touchscreen screen reader experience.
- Participants used a Samsung S3 Mini with Android 4.1, TalkBack enabled, and a background logger (TinyBlackBox) for two months; researchers collected in-situ usage plus eight weekly lab sessions of 20 minutes each using unconstrained sentence transcription in Portuguese.
- Analysis used input-stream measures (WPM, total/uncorrected/corrected error rates, character-level error types) and touch measures (landing positions, movement distance/time, PL/TA ratio, pauses, visited keys); statistical testing used Shapiro-Wilk, then mixed-effects models with Week as a fixed effect and Participant/In-Situ Usage Time as random effects.
## Limitations
- The paper states the sample was only five novice blind users and says results should be replicated with more experienced blind typists.
- The authors note that allowing participants to use the phone “in-the-wild” and running weekly lab assessments may have influenced learning, so the reported weekly performance may not represent a truly natural learning experience.
- Evident from the text: the study is limited to a single device/keyboard setup (Samsung S3 Mini, default Samsung QWERTY, TalkBack) and to Portuguese, so generalization to other devices, keyboards, languages, or screen readers is not established.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
3 Valid · 6 Not valid · 11 Not checked · 133 Own results and interpretations
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“However, in doing so these works neglect to justify how and why differences exist between interfaces; for instance, in character-level errors [13, 27].”
What the sources say
- Not checked — no full text of the sourcecites [13] MacKenzie, I.S. and Soukoreff, R.W. 2002. A character-level error analysis technique for evaluating text entry methods. Proceedings of the second Nordic conference on Human-computer interaction (2002), 243--246. source
- Disputedcites [27] Wobbrock, J.O. and Myers, B.A. 2006. Analyzing the input stream for character- level errors in unconstrained text entry evaluations. ACM TOCHI, 13, 4 (2006), 458--489. sourceExperimental — accuracy not yet measured
Section 1.1 (pp. 4–5) says aggregate error rates are too coarse to identify problematic characters and presents character-level analysis as the paper’s remedy. The Conclusion (p. 30) says its methods provide designers with more character-level error data. Although the paper does not itself explain a specific difference between two interfaces, it does not neglect the cited sentence’s example of character-level errors. (LLM, conf=0.82)
Checked by a reasoning model — may be wrong.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Total error rates tend to 7.4%.”
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Moreover, Figure 3 shows that error rates are starting to stabilize around that value.”
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“To fix the error, several characters, including correct characters, are usually deleted.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“A detailed inspection of logs files show that editing operations, such as cursor movement, were never used.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
82 more claims
Create a free account to see all 153 claims and to verify the 1 reference still open.
- Probably not a substantial claim. 28
- It may state two separate claims as one. 45
- The quote only partly states this claim. 1
Citations
Cites
Linear Mixed Models for Longitudinal Data
Verbeke, G. (1997) · Lecture Notes in Statistics · doi:10.1007/978-1-4612-2294-1_3 · Cited by 3,506
Generalized, Linear, and Mixed Models
McCulloch, C. E., & Searle, S. R. (2000) · Wiley Series in Probability and Statistics · doi:10.1002/0471722073 · Cited by 2,994
Text Entry for Mobile Computing: Models and Methods, Theory and Practice
MacKenzie, I. S., & Soukoreff, R. W. (2002) · Human-Computer Interaction · doi:10.1207/s15327051hci172&3_2 · Cited by 571
Text Entry for Mobile Computing: Models and Methods,Theory and Practice
MacKenzie, I., & Soukoreff, R. W. (2002) · Hum. Comput. Interact. · doi:10.1080/07370024.2002.9667313 · Cited by 483
Accuracy measures for evaluating computer pointing devices
MacKenzie, I. S., Kauppinen, T., & Silfverberg, M. (2001) · Proceedings of the SIGCHI Conference on Human Factors in Computing Systems · doi:10.1145/365024.365028 · Cited by 448
Metrics for text entry research: an evaluation of MSD and KSPC, and a new unified error metric
Soukoreff, R. W., & MacKenzie, I. S. (2003) · Proceedings of the SIGCHI Conference on Human Factors in Computing Systems · doi:10.1145/642611.642632 · Cited by 398
Generalized Linear Mixed Models
McCulloch, C. E. (2001) · Encyclopedia of Environmetrics · doi:10.1002/9780470057339.vag009 · Cited by 295
Analyzing the input stream for character- level errors in unconstrained text entry evaluations
Wobbrock, J. O., & Myers, B. A. (2006) · ACM Transactions on Computer-Human Interaction · doi:10.1145/1188816.1188819 · Cited by 200
Logarithmic Transformations in ANOVA
Berry, D. A. (1987) · Biometrics · doi:10.2307/2531826 · Cited by 195
(2007) · Cited by 192
Cited by
Accessible Touchscreen Technology for People with Visual Impairments: A Survey
Grussenmeyer, W., & Folmer, E. (2017) · ACM Transactions on Accessible Computing · doi:10.1145/3022701 · Cited by 86
“Nicolau et al. [2015] recorded the learning of typing on a mobile touchscreen by novice users.”
BackgroundMobile device accessibility for the visually impaired: problems mapping and recommendations
Damaceno, R. J. P., Braga, J. C., & Mena-Chalco, J. P. (2018) · Universal Access in the Information Society · doi:10.1007/s10209-017-0540-1 · Cited by 73
“Furthermore, there are omission errors in which one character is not entered because of forgetfulness or lack of the user’s awareness [39].”
BackgroundOpen Challenges of Blind People Using Smartphones
Rodrigues, A., Nicolau, H., Montague, K., Guerreiro, J., & Guerreiro, T. (2020) · International Journal of Human–Computer Interaction · doi:10.1080/10447318.2020.1768672 · Cited by 53
“Several studies focused in understanding text-entry needs on touchscreens (H. Nicolau et al., 2015b; A. Rodrigues et al., 2016) while others looked for alternative interaction methods, many of which based on Braille (Azenkot, Wobbrock et al., 2012; H. Nicolau, Montague et al., 2014; H. Nicolau et…”
MethodSelf-citationSmartphone Usage by Expert Blind Users
Jain, M., Diwakar, N., & Swaminathan, M. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445074 · Cited by 41
Accessible Gesture Typing for Non-Visual Text Entry on Smartphones
Billah, S. M., Ko, Y.-J., Ashok, V., Bi, X., & Ramakrishnan, IV (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300606 · Cited by 21
“A longitudinal study of the typing performance of blind users is reported in Nicolau et al. [16].”
Background
Metadata
- Authors
- Nicolau, H., Montague, K., Guerreiro, T., Rodrigues, A., & Hanson, V. L.
- Affiliations
- Rochester Institute of Technology (US)
- University of Dundee (GB)
- University of Lisbon (PT)
- Venue
- Proceedings of the 17th International ACM SIGACCESS Conference on Computers & Accessibility - ASSETS '15
- Publisher
- ACM Press
- Published
- 2015
- Pages
- 273-280
- OpenAlex
- W2001516180
- Cited by
- 37
- References
- 31
- Citations by year
- 2026: 2 · 2025: 3 · 2024: 1 · 2023: 2 · 2022: 3 · 2021: 4 · 2020: 2 · 2019: 4 · 2018: 2 · 2017: 8 · 2016: 4 · 2015: 2
- FWCI
- 8.90
- Type
- conference paper
- Language
- en
- Open access
- Locations
- International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) (publishedVersion)
- Discovery Research Portal (University of Dundee) (submittedVersion, other-oa)
- Funders
- Fundacao para a Ciencia e a Tecnologia (SFRH/BD/103935/2014, UID/CEC/00408/2013)
- Research Councils UK (EP/G066019/1)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Tactile and Sensory Interactions
- Social Sciences › Human Factors and Ergonomics › Digital Accessibility for Disabilities
- Health Sciences › Health Professions › Occupational Therapy › Assistive Technology in Communication and Mobility
- Keywords
- blind users, typing performance, substitution errors, touch behavior, virtual keyboard, visual impairment, novice users
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-09-30, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-09-30 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Typing Performance of Blind Users. https://app.cochat.ai/papers/10.1145/2700648.2809861
- The paper
- Nicolau, H., Montague, K., Guerreiro, T., Rodrigues, A., & Hanson, V. L. (2015). Typing Performance of Blind Users. Proceedings of the 17th International ACM SIGACCESS Conference on Computers & Accessibility - ASSETS '15. https://doi.org/10.1145/2700648.2809861
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.