Accessibility Audit 2021

Introduction

This audit was completed in July 2021.

People may not have a choice when using a public sector website or app, so it’s important they work for everyone. The people who need them the most are often the people who find them hardest to use.

Regulations came into force for public sector bodies on 23 September 2018. They state that you must make your website or mobile app more accessible by making it ‘perceivable, operable, understandable and robust’. The full name of the regulations is the Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018.

Overview

The Calderdale Next Chapter website has 133 active pages based on page views taken from website analytics between 1st January 2021 and 5th July 2021.

Website analytics provide a wider insight into how our page numbers evolve since launch. They may include pages that are no longer published and miss pages that have received no recent traffic.

The Calderdale Next Chapter website is made up of 3 core page types(published amount in brackets):

Articles (61)
Projects (19)
Unique or basic page (7)

Unique/basic pages do not follow a pre-defined layout like Articles or Projects.

3 different component types:

Header
Footer
Project carousel

Components are small pieces of self-contained design and functionality. They are managed in one place but used across multiple pages. This is beneficial as we can test smaller samples and apply one fix which updates all instances.

There are 161 published PDFs, an increase of 117 since the last assessment.

Sample

Criteria

We use a representative sample of pages and features to test how accessible our websites are. Our sample takes into consideration:

Pages that share the same source.
Reusable components that share the same source.
Unique features that use bespoke functionality to complete tasks.
Popular user journeys based on website analytics.

We believe this is the most effective approach to test our system as it allows us to test the widest range of layouts and functionality without having to run extensive tests on multiple comparable pages.

Inventory

Test process

All pages listed in the inventory are tested using the same process. They are first scanned using tools that can quickly detect common issues. We use a combination of:

WAVE
Axe dev tools
Google Lighthouse

An additional manual check is carried out aided by the NVDA screen reader application to address concerns that automated tools are unable to process.

We use Adobe Acrobat DC to test PDF documents. We first use the built-in accessibility checking function to quickly detect common issues. We then do a manual check of the document which includes checking the document tags, reading order, colour contrast, grammar and more.

Results

Our audit has identified areas where there is room for improvement and barriers that need overcoming. The results from this audit will be shared with the system and content owners so that remedial action can be taken.

Some errors may affect more than one page where they share the same source code such as in global elements like that header and footer. Where the same error is detected, it will only be recorded once against the page where it is first encountered.

Homepage

The homepage is an example of a unique page. No critical errors were detected in the automated check of the page. Google Lighthouse scored the page 100% for accessibility.

Automated checks highlighted:

2 ‘Read more’ links in the news section of the page as being both redundant and suspicious.
Skip to main content link is marked as content sat outside of a landmark region.

A manual test of the page uncovered:

Multiple links to the news page with different descriptions.
Skip to main content link goes to a new page rather than the main region of the page.
Some images have alt text that is not suitable or necessary.
For the interactive map, NVDA unhelpfully narrates “graphic To get missing image descriptions, open the context menu.” Followed by “clickable” 16 times.

Projects

The projects page is an example of a unique page. No critical errors were detected in the automated check of the page. Google Lighthouse scored the page 100% for accessibility.

Automated checks of the page highlighted no areas of concern.

A manual test of the page uncovered:

Images with alternative text is are unnecessary and distracting in the project link tiles.
Incorrect semantics used for the tiles. It would benefit from being a list of links without and not headings.

News

The news page is an example of a unique page. No critical errors were detected in the automated check of the page. Google Lighthouse scored the page 99% for accessibility.

Automated checks highlighted:

Missing heading level 1.
Unnecessary alt text.
Skipped heading level in the pagination.
Redundant links (image/heading of each news item).

Manual checks didn’t find any additional areas of concern.

A629 Phase 1B Supporting Documents

This page is an example of a basic page. No critical errors were detected in the automated check of the page. Google Lighthouse scored the page 100% for accessibility.

Automated checks highlighted:

2 redundant links – adjacent links have the same destination but upon inspection these have different labels so they may have been misconfigured.

Manual checks uncovered:

Ambiguous link labels – multiple links to “report” and “minutes”.

A629 Orange Street to Ogden

This is an example of a project page which shares its design and functionality with all other project pages. No critical errors were detected in the automated check of the page. Google Lighthouse scored the page 97/100 for accessibility.

Automated checks highlighted:

List that doesn’t contain List Item elements.

Manual checks uncovered:

PDF document hyperlinks have file name descriptions which aren’t user friendly.
The map embed can receive focus – this could be a hindrance for screen readers.

Have your say on improvements around North Halifax

This is an example of a news article page which shares its design and functionality with all other news articles. One contrast error was detected in the automated check of the page. Google Lighthouse scored the page 95% for accessibility.

Automated checks highlighted:

Colour contrast is insufficient between link colour and grey background.
There is a skipped heading detection due to the title being h1 and footer links being h3.

Manual checks uncovered:

Unnecessary alt text on the thumbnail.
Full URL used as link description multiple times.

Public engagement analysis - A629 Calder and Hebble junction

This is an example of a PDF file linked to from a project page.

An automated check of the document highlighted:

20 instances of images with no alternative text or decorative marking.
16 images not visible.
Insufficient metadata – no title or subject and ambiguous author.

A manual check of the document uncovered:

The document has no tag structure.
There are no semantic headings or easy means of navigation due to absence of tags.
Headings appear to be plain text but they are embedded images.
10 incorrect use of tables.
NVDA starts by reading the duplicated header of each page repeatedly for every page before it starts reading the content of page 1.

West Yorkshire-plus Transport Fund projects: Engagement approach

This is an example of a PDF file linked to from a project page.

An automated check of the document highlighted:

Insufficient metadata – no title or subject and ambiguous author.

A manual check of the document uncovered:

The entire document is a table according to the tags, but NVDA is unable to find any tables on the page.
The table doesn’t have any semantic table headings (TH).
The table is missing many cells (white cells), this is causing the blue cells to shift to the right pairing the values with incorrect headings.
The 3 stages paragraph do not belong in the data table.

Bridge Scheme Order – Sealed

This item is a PDF file. An automated check of the document highlighted an excessive quantity of errors that could not be visibly attributed to page content; this is likely due to the nature of the document being a scanned image. One common error noted is the absence of document metadata such as title, subject and author.

A manual check of the document uncovered:

The document is upside down.
The content cannot be read by NVDA.
The document requires users to have knowledge of third-party software to adjust orientation and/or scan the text in the image. I could only manage to get text recognition working for the first paragraph. OCR (Optical Character Recognition) cannot guarantee accuracy.

2020 Supplemental CPO A629 1b Statement of Reasons

This item is a large PDF file made up of 31 pages. An automated check of the document highlighted:

Insufficient metadata – no title or subject and ambiguous author.
Bookmarks failed – document over 21 pages without bookmarks.
16 figure related alternative text errors.
5 table summary errors.

A manual check of the document uncovers:

Although there appears to be a visual menu – there are no links or bookmark functionality.
The numbers in the menu don’t correspond with page numbers which may cause confusion.
There are very few semantic headings. The headings that do exist are sporadic at tagged incorrectly.
The contents menu is 20 separate paragraphs instead of a semantic list or Table of Contents.
The reading order does not match the visual flow of the document.
Lots of empty content – for example, the document begins: “blank blank graphic light vertical blank blank borough council of Calderdale……… HIGHWAYS ACT 1980 blank blank 2020”.
Single lists are split into multiple separate lists when over multiple pages
There are several images of graphs – does each graph have an accessible means of conveying the same information?
There are image(s) of table(s) instead of semantic tables that can be parsed.
Improper use of tables in appendices.

Piece Hall Environs Supplementary planning document

This is a large PDF file made up of 53 pages. The checks on this document were cut short due to its size and the quantity and frequency of issues discovered. An automated check of the document highlighted:

Tagged PDF failed (no tag structure)
Primary language failed
Title failed (insufficient metadata)
Bookmarks failed
Tagged content failed (1112 instances)
Tab order failed
Figures alternative text failed
Nested alternative text failed
Hide annotations failed
Other elements alternative text failed (43 instances)
Tables – rows failed
Tables – TH and TD failed
Tables – Headers failed
Tables – Regularity failed
Tables – Summary failed
Lists – List items failed
Lists - Lbl and lbody failed
Headings appropriate nesting failed

A manual check of the document has uncovered:

Large blocks of uppercase text.
Spelling and grammar mistakes. For example “Halfiax” heading on page 9.
No semantic elements such as headings and lists.
Insufficient colour contrast of text.
Lots of double slashes that appear to be a style element but are announced as “slash slash” by NVDA.
The reading order is inaccurate. Text is narrated in a seemingly random order which is confusing and bordering on impossible to understand.
NVDA gets stuck in places.
The Piece hall in the bottom right of all pages with a styled P character is narrated as “Pee-Piece hall”.
Reliance on sensory characteristics to find information.
Inaccurate use of sensory characteristics. For example, page 9 tells readers to look to the right to see an illustration, where the illustration is on the left.
Incorrect use of tables.

Recommendations

Based on the assessment results, the following actions are recommended to improve the accessibility of the website:

System owner to update website code to address technical issues highlighted
System owner to explore pages outside of the test sample for any other areas that can be improved
Content owner to schedule remedial action for inaccessible PDF documents prioritising them based on publish date or popularity.
Content owner to publish all future updates in one of the following formats in order of most appropriate:
- Native web page built with semantic HTML in accordance with WCAG 2.1;
- Semantically tagged PDF files designed in accordance with WCAG 2.1;
- Partially compliant document or webpage that demonstrates an attempt has been made to meet basic accessibility requirements where possible. This must include a fully accessible counterpart uploaded alongside the original.