JSTOR Digital Stewardship Services: Content accessibility recommendations

The goal of making your distinctive and archival collections accessible is to enable researchers, regardless of disability, to research, learn from, and teach with your collections. 

Accessibility is contextual. Different people have different needs, depending not only on their disabilities, but their level of expertise, where they're encountering your content, and how the content is being used. We encourage you to consider the different context(s) relevant to your institution and what accessibility measures will be needed to support those contexts. 

The recommendations in this article are drawn from the Web Content Accessibility Guidelines (WCAG) as well as other established accessibility best practices. When relevant, the specific WCAG Success Criteria(s) will be noted in each section. These recommendations are written with the JSTOR platform in mind, but are transferrable to other display and access platforms you may use as well. 

Images

To make an image accessible, you must provide text alternative(s) that serves the same purpose, and conveys the same information, as the image.

This means someone who cannot perceive the image should understand what the image is and why they are encountering it. In the context of archives and special collections, that could mean:

• Why is this image part of this particular collection?
• Why is this image relevant to their research or class?
• What is this image about? 

The most common text alternative for images is alt text. All images you publish to JSTOR or display elsewhere should have alt text. Alt text is typically 1-2 sentences and captures the most important information in the image.

Complex images, such as maps, diagrams, charts, infographics, and in some cases, artistic images like paintings, have meaning that is difficult or even impossible to convey in one or two sentences. In addition to the alt text, these images require a long description, which can be presented in many formats.

Alt text

Generally, alt text should:

• Be associated with the image programmatically, rather than presented separately on the page.
• Be 1-2 sentences long, maximum.
• Be consistent with your other descriptive item metadata.
• Include only the information needed to understand the image.
• Not include speculation or interpretation about the image.

If your display platform doesn't allow you to associate alt text with the image directly, you can provide it via a clearly-labeled field such as "Alt text" or "Accessible image description".

Long description

Images like maps, posters, charts, and art images that are too complex to adequately describe in 1-2 sentences need additional description to be made accessible. This is typically called a long description.

There is no formal definition or format for a long description. Techniques to provide a long description are listed below. These are non-exhaustive, and may be used in combination or individually, as appropriate. 

Long description metadata field

There are no requirements for the length and content of a long description as long as it conveys the same information as the image. For some images, such as charts, it may be appropriate to include a dedicated, full-length description of the image content.

You can add a custom long description field in Stewardship. Alternatively, your existing Description field can be used for long description, for example, including transcription of handwritten elements or more detailed explanation of the image's materiality. 

Item-level descriptive metadata

In many cases, your item-level metadata will be sufficient as long description when used in combination with alt text. This means in addition to making your content more discoverable, rich item-level metadata can greatly improve your images' accessibility. 

For a list of all default fields when publishing to JSTOR, including field definitions, see JSTOR fields and attributes. We also offer mapping recommendations when publishing to JSTOR for the following project templates:

• JSTOR template [Tiers 2-3]
• Visual Resources template [Tiers 2-3]
• AI-Compatible Dublin Core template [Tier 3]

Text transcripts

Any text in an image must be provided in an accessible format alongside the image. This can be a custom metadata field that includes the text or a link to the text in an accessible format, such as a .txt or .doc file, or a dedicated web page.

If you are using a text transcript as your long description and not including any other long description materials, the text transcript should include any visual details that are necessary to the meaning of the image.

For additional information on displaying text transcripts, see Images of text: Text transcripts.

Images of text

Unlike complex images, images of text do not contain meaningful visual (non-textual) information. They are text that has been rendered in an image format, such as a scan of a handwritten letter or printed document. 

All text in the image should be presented in an accessible alternative format, such as a text transcript. The transcript can be shared a custom metadata field that includes the text, or a link to the text transcript in an accessible format, such as a .txt or .doc file, or a dedicated web page. 

We recommend additionally including details about the placement and fidelity of text, as the document's materiality is often meaningful in the archival and distinctive collections context. For example, if part of a handwritten letter has been damaged, or if there is marginalia, noting these details will enhance the accessibility of your content.

Text transcripts

Text transcripts must be shared in an accessible file format. In general, we recommend displaying text transcripts in a format that allows users to control how the text is displayed, such as part of the metadata you publish to JSTOR.

If an image of text is being displayed online as an image file (e.g. a JPG, PNG, TIF) it must have alt text that links to or indicates where the accessible text transcript can be located. For example: "A handwritten letter by Mary Shelley. Full text transcript in the Description field."

For additional examples of how to display text transcripts on the JSTOR platform, see Transcripts and accessibility. While these examples use JSTOR Seeklight-generated transcripts, the techniques are also applicable for manually-created transcripts and can be used by both Tier 2 and Tier 3 participants.

PDF documents

In order to be considered accessible, PDF documents must be machine readable. This means a screen reader or text-to-speech tool can read the text and understand the structure within the document.

The process for making text machine readable is called OCR (Optical Character Recognition.) All textual content types published to JSTOR have OCR automatically applied to them if there is no existing OCR in the document. Since this is an automated process, we recommend double-checking the downloaded PDF to ensure the OCR meets your accessibility standards.

In addition to OCR, a PDF needs:

• Proper tags, which are structural metadata that define content within the document.
• A logical reading order, so assistive technology understands the order in which the content on the page should be read.
• Alt text for any meaningful images within the document. 

PDFs and text transcripts

An accessible PDF doesn't need a text transcript. However, we recommend providing a text transcript even for accessible PDFs as text provides more robust accessibility options than PDFs.

A/V content

Like images, audio and video content requires a text alternative that presents equivalent information for the content. What text alternative(s) are needed depend on the type of audio or video you are working with.

Captions

Used for: Videos with audio, including spoken word and non-spoken audio content (music, sound effects, etc.). 

Captions can be open, meaning always visible, or closed, meaning a user must choose to enable them.

Captions should:

• Include all meaningful information, including sounds and dialogue.
• Be presented in the same language as the video.

When publishing to JSTOR, your video's transcript file is used as closed captions within the JSTOR video player, in addition to displaying on the page as an accessible text transcript.

Descriptive transcript

Used for: Audio with no video component (audio-only); Video with no audio component (video-only); Video with audio.

Audio with no video component should always be accompanied by a descriptive text transcript. A descriptive transcript can also be used for videos with or without audio.

The descriptive transcript should:

• Convey the same information as the video content, including any text or graphics, settings, the actions and expressions of people, animals, etc.
• Be presented in the same language as the video or page.

You can use the same techniques used for text transcripts to display your descriptive transcripts when publishing to JSTOR.

Audio description track

Used for: Video with no audio component (video-only); Video with audio.

In addition to, or instead of a descriptive transcript, you can provide an audio description track for videos with or without spoken word content. 

The audio track should:

• Convey the same information as the video content, including verbally describing any text or graphics, settings, the actions and expressions of people, animals, etc.
• Be presented in the same language as the video or page.

An audio track, while satisfying WCAG standards, won't be accessible to users who are both unable to see or hear your content. Therefore, we recommend always supplying a text transcript whether you additionally provide an audio track or not.

Descriptive metadata

This section includes accessibility considerations for your item-level metadata, regardless of content type.

Provide contextual links

Stewardship supports the use of hyperlinks in specific fields and in your collection and institutional descriptions. The purpose of the link should be clearly communicated to the user from the link text alone.

The link text should not:

• Include the word "link"
• Use the phrase "click here" or the word "here".
• Present the "naked" URL, for example: "https://www.example.org"

If the link will open in a new page, include that information in the link text. For example:

• A link to an accessible text transcript: "Accessible text transcript (this will open in a new page)."
• A link to a related item: "Other examples of gothic novels include The Phantom of the Opera (1909-1910)."

Define important vocabulary terms

Many terms used to describe your content may be unfamiliar to novice researchers, students, or others. If your collection is open to all levels of researcher, consider including a short definition when using uncommon or field-specific terminology, or provide a resource that introduces these terms, especially if they're critical to the collection.

Use consistent terminology

Using consistent phrasing and structure in item descriptions across a collection will help users understand how your items relate to each other and to the collection as a whole. This is especially useful on JSTOR, as users navigate JSTOR collections at the item level.

This includes using predictable naming conventions for titles and components.

Don't rely on visual or spatial cues only in description

Spatial or visual information can be useful when describing an item, but you shouldn't rely solely on these details as not all users can perceive them. When possible, provide an alternative explanation. For example:

• "Signature in bottom left, underneath the image in the margins of the page."
• "A figure stands in the middle, with an arm stretched above them toward the right side of the frame." 

Include information about color only when it's relevant and meaningful to understanding the item.