Second Cycle

Second user-centered-design cycle: Research and improvement (partially complete)

The key objective of the second cycle in the development of AVoS is to improve the protoype and allow for more sophisticated user interactions. Again, before doing any implementation work, it is important to re-assess the needs of the target group. From the evaluation in cycle 1, there are indications as to which improvements of AVoS are desirable. However, the evaluation has been conducted among only 15 school children. This explains the need to evaluate a more global demand of an auditory vocabulary and spelling trainer. To this end, an online survey has been conducted, using the barrier-free tool The survey has been completed by 88 visually impaired adults out from which 77 sets of data could be evaluated. The youngest test participant was 10 years old, the oldest was 74 years old and the mean age was 38.4 (SD=13.794). The majority of test subjects had achieved a university degree (29.5 %) or completed high school (16.7 %). Also, most participants specified their braille literacy as “very good” (67.9 %). Test subjects were asked to answer a total of 18 questions concerning their vocabulary learning and computer usage habits, and their usage of vocabulary learning software. Further, in an open-ended question, the subjects could indicate which features were important to them in a CAVL application. 

(1) What are typical vocabulary study and computer usage habits among visually impaired language learners?

The results from the online survey show that language learners with visual impairments struggle significantly more with acquiring spelling of an L2 vocabulary item than with acquiring its meaning. Consultation of the chi-square distribution χ2 (df=1)=12.737 shows that the probability of observing this difference is significant on a level of p<0.01. In terms of finding spelling mistakes in a text on the computer screen, more test participants indicated that they find the mistake from the mispronounciation of the text-to-speech synthesis rather than finding it with the braille display. On a scale from 1 (never) to 7 (always) of how often they apply either method, the mean for the screen reader was 4.59 (SD=1.814) and for the braille display it was 4.19 (SD=2.134). Tested against the middle value 4 (sometimes), the deviation was significant in the case of the screen reader (t(67)=2.674, p<0.01 (two-sided)). Further, there is a statistically significant correlation between how well visually impaired people know the braille script and how much they like to engage in foreign language study F(1;72) = 4.052, p<0.05 (two-sided).

(2) Which features are desirable in an auditory vocabulary and spelling trainer among adult learners?

68 test subjects answered this question and mentioned the following features as making a CAVL application usable for them: pronounciation help 48.5%, orthographic feedback 30.9%, screen-reader friendly 25%, context of the vocabulary item 20.6%, flash card method 23.5%, simplicity of the program19.1%, braille line access 17.6%. Further, some test participants noted grammar help, display of progress, online dictionary, free of charge, and rich contrast for the display. 

(3) How can the findings from the user survey be implemented into the AVoS prototype?

Based on the results from cycle 1 and the results from the online survey, the following additional features should be included in the next prototype of AVoS: Pronounciation help. Provide formant text-to-speech synthesis in the according language to allow a more natural pronounciation. Intermediate pronounciation. Include a function to listen to the word ‘so far’. Visually impaired people can easily pick out spelling errors from the misprounounciation of their screen reader. This feature could help them to verify their input easily before submitting their answer. Braille line access. Usability for screen reader and braille display. While the screen reader is often prefered, good accessibility via braille line is also needed. Context. Present vocabulary in the context of a sample sentence. This could be achieved via external links to online dictionaries.