Introduction
At the beginning of Blindness Awareness Month, which took place during October, I was left feeling frustrated and disheartened by the high level of inaccessible blindness content I saw being posted online. Each year it seems I speak out about the problem of inaccessibility within blindness spaces and every time I’m told it can’t be that much of an issue. I felt, and still feel, that it is a problem, but I had no hard data to back up what I was saying.
As a result, I decided to track content posted over on Twitter for Blindness Awareness Month to see how bad the inaccessibility problem within blindness spaces actually is. Here are my findings.
An explanation of my process
Firstly, I would like to say that whilst I took care to follow a fairly strict process, there may be content that I have missed. I was tracking all of this data by hand, as it were, and although I made the greatest effort to record my findings accurately, I was limited in some ways by the technology I was using. I did my best to search for content using various different text strings to try and find as much of it as I could, however there is a chance that I have also missed some posts.
In order to find content, I monitored the following text strings:
- Blindness awareness month
- #BlindnessAwarenessMonth
- Braille week
- #BrailleWeek2020
- #WhiteCaneAwarenessDay
In order for a post to be included in the data it had to contain the following:
- A mention of Blindness Awareness Month, white cane awareness day, or Braille Week, which takes place during the month.
- A photo, GIF or video.
I did not include posts which were only text based and did not contain some kind of media.
In order for posts to be considered accessible they had to meet the following criteria:
- Photos and GIFs must contain an alt text description which accurately describes the image, either using Twitter’s alt text feature, or in the body of the post.
- Videos should contain a transcript.
A video was not considered accessible if it included captions but no transcript, as this is still not accessible to deafblind people.
A photo was not considered accessible if the alt text field was filled, but it was used to add hashtags, or other data that did not describe the image, instead being used for SEO or other marketing purposes.
Blindness Awareness Month results
I tracked a total number of 399 tweets posted for Blindness Awareness Month. This works out at an average of 12.87 tweets posted per day. Some days during the month were far more active than others, for example on the 15th of October, which was designated as white cane awareness day, 60 tweets were posted, whereas on the 23rd and 24th of October, only 2 tweets were posted on each day.
The overall number of tweets which met the criteria to be considered accessible was 131, leaving 268 tweets that were not accessible. This means that 32.8% of tweets were accessible, and 67.2% were not accessible.
I tracked the number of accessible and inaccessible tweets posted per day as I hoped that I might see some interesting trends throughout the month. I had supposed, for example, that we might see a greater daily percentage of accessible content as the month progressed and organisations became aware that the content they had initially posted was not fully accessible to the blind and deafblind community. Whilst my hypothesis was based on a reasonable assumption it ultimately proved not to be the case.
Blindness Awareness Month data
The following table shows the data I gathered throughout the month, including the percentage of accessible and inaccessible content posted per day.
| Day | Accessible content | Inaccessible Content | Total | Percentage accessible | Percentage inaccessible |
|---|---|---|---|---|---|
| 1 | 4 | 12 | 16 | 25 | 75 |
| 2 | 4 | 14 | 18 | 22.22 | 77.78 |
| 3 | 4 | 1 | 5 | 80 | 20 |
| 4 | 2 | 3 | 5 | 40 | 60 |
| 5 | 6 | 26 | 32 | 18.75 | 81.25 |
| 6 | 9 | 19 | 28 | 32.14 | 67.86 |
| 7 | 10 | 17 | 27 | 37.04 | 62.96 |
| 8 | 7 | 20 | 27 | 25.93 | 74.07 |
| 9 | 5 | 4 | 9 | 55.56 | 44.44 |
| 10 | 5 | 4 | 9 | 55.56 | 44.44 |
| 11 | 5 | 2 | 7 | 71.43 | 28.57 |
| 12 | 6 | 7 | 13 | 46.15 | 53.85 |
| 13 | 4 | 8 | 12 | 33.33 | 66.67 |
| 14 | 6 | 7 | 13 | 46.15 | 53.85 |
| 15 | 21 | 39 | 60 | 35 | 65 |
| 16 | 5 | 13 | 18 | 27.78 | 72.22 |
| 17 | 2 | 2 | 4 | 50 | 50 |
| 18 | 1 | 5 | 6 | 16.67 | 83.33 |
| 19 | 2 | 7 | 9 | 22.22 | 77.78 |
| 20 | 0 | 10 | 10 | 0 | 100 |
| 21 | 4 | 6 | 10 | 40 | 60 |
| 22 | 2 | 7 | 9 | 22.22 | 77.78 |
| 23 | 0 | 2 | 2 | 0 | 100 |
| 24 | 1 | 1 | 2 | 50 | 50 |
| 25 | 2 | 1 | 3 | 66.67 | 33.33 |
| 26 | 3 | 3 | 6 | 50 | 50 |
| 27 | 4 | 2 | 6 | 66.67 | 33.33 |
| 28 | 2 | 4 | 6 | 33.33 | 66.67 |
| 29 | 0 | 6 | 6 | 0 | 100 |
| 30 | 3 | 12 | 15 | 20 | 80 |
| 31 | 2 | 4 | 6 | 33.33 | 66.67 |
| Total | 131 | 268 | 399 | 32.83 | 67.17 |
World Sight Day
World Sight Day falls during Blindness Awareness Month, on the second Thursday in October. I also tracked posts for World Sight Day, using the criteria mentioned near the start of this article to determine whether content was accessible or not. Unlike Braille week and white cane day, I did not include these within the Blindness Awareness Month statistics as content is overwhelmingly posted by those involved in the medical field in some way, and I felt like it might unfairly skew the stats. However, I thought it was also important to gather data for this event.
Process
All of the statistics relating to World Sight day were gathered in the days leading up to WSD, and on the day itself. In order to gather data, I searched for the following text strings:
- World Sight Day
- #WorldSightDay
- #WSD2020
The event attracts a high level of traffic on Twitter and I did not analyse every single tweet that was posted, as at some points in the day thousands of tweets were posted per hour. Instead, I analysed a number of tweets per hour, so that I wasn’t gathering data at only one point during the event. This was the best I could do, given the limited resources I had access to.
World Sight Day Results
I tracked a total of 827 tweets posted for World Sight Day. 61 of these met the criteria to be considered accessible, however 766 did not. This results in 92.6% of tweets being inaccessible and only 7.4% were accessible. As you can see, a far greater percentage of tweets posted for World Sight Day were inaccessible, compared with tweets posted for Blindness Awareness Month. However both events ended up producing a majority of content which was inaccessible.
World Sight Day data
The following table shows the data I gathered for World Sight Day.
| Accessible tweets | Inaccessible tweets | Total | Percent accessible | Percent inaccessible |
|---|---|---|---|---|
| 61 | 766 | 827 | 7.4 | 92.6 |
What the Results Don’t tell us
Whilst the results clearly show that the majority of content posted for both Blindness Awareness Month and World Sight Day is not accessible, they cannot indicate intent, or attitudes towards blindness, areas which I believe deserve further examination.
I would have loved to analyse the language used in the tweets to see whether there was any relation between accessible content and positive attitudes towards blindness, however as this was a personal project and I am also busy with other work this wasn’t possible.
It also cannot tell us why people do not make use of Twitter’s accessibility features. We do not know if they are intentionally excluding blind people or if they are unaware of alt text.
What we do know is that there is an accessibility problem within the blindness field and this should be taken more seriously in future than it has been until now.
Conclusion
As already mentioned, this study was limited in scope and only forms the beginning of the process we must follow if we are truly to understand the reasons for such high levels of inaccessibility within digital blindness spaces. Overall, content posted which claims to raise awareness of blindness is not accessible, which should be cause for concern.
Overall levels of accessible tweets are even lower. There is limited data that can tell us about how people outside of blindness spaces make use of Twitter’s accessibility features, however a study by Dr Cole Gleason at Carnegie Mellon University discovered that only 0.1% of tweets containing images also contained an alt text description. In comparison, the blindness field is doing quite well, however this isn’t much of an achievement when the bar is so low. I recommend reading the study examining how user provided descriptions have failed to make Twitter accessible, as it provides important data which we can use to make a case for improved accessibility features.
Ultimately, those working within the blindness field should be setting an example. It’s not enough to talk about how blind people should have rights, jobs and equal opportunities if you then deny us access to your advocacy. It means very little to raise awareness of blindness if you can’t make the most minimal effort to include us. After all, the problem is not my blindness but the many ways in which I am excluded by a world that is not accessible.
Discover more from Catch These Words
Subscribe to get the latest posts sent to your email.