Our two workstreams separate our reports and URL analysis from the image and video assessments. Here we outline the methodology and datasets used in each of these processes.
This year we have been able to provide more detailed analysis focusing on different areas of our work. Image, video and Multichild analysis are a new focus for 2024. We have created six dataset tags to clearly identify what data and what part of our workstream is being referred to throughout all sections of our report.
We assess child sexual abuse material according to the levels detailed in the Sentencing Council's Sexual Offences Definitive Guideline. The Indecent Photographs of Children section (Page 34) outlines the different categories of child sexual abuse material.
When our analysts assess a report or URL, the age classification is based on the youngest child visible in the imagery; for example, a URL containing criminal imagery of a 2-year-old, a 7-year-old and a 13-year-old would be assessed as ‘0-2’ to reflect the age of the youngest child.
The same approach is applied to severity, with the most severe category of abuse visible in the imagery on the URL being recorded. In a composite image or video showing every category of abuse (from A to C), the analyst would log an assessment of Category A.
Where some reports include multiple images or videos displayed on a single URL, the same rule of ‘youngest visible child’ and ‘most severe visible category’ is applied; however, the two may not be relevant to the same image.
Sex can be recorded as boys, girls, both or – in rare cases - unidentified. Where boys are recorded this indicates ‘only boys’ have been seen, this is the same with girls, however if the report is recorded as ‘both’ it may be that both ‘boys only’ and ‘girls only’ imagery was present or it could be that boy, and girl, victims were present together in the same image or video.
Below are the definitions of the dataset tags you will see throughout this report.
Reports analysis
Reports analysis
Most reports contain a URL (website address), however not all reports do.
Newsgroups ‘predate’ the world wide web and are not URLs, however the IWF is able to receive reports of newsgroup content. Any child sexual abuse material found within a newsgroup does not contribute to the URL analysis data.
Children using child reporting services, such as Report Remove and Meri Trustline, can report non-URL content, such as images or videos that children can submit directly to the IWF that may not knowingly be hosted online.
In both cases criminal imagery is still downloaded and assessed by the IWF Taskforce and therefore a report is created and assessed using the methodology above.
URL analysis
URL analysis
One URL (a website address) can represent a webpage, or an online download that contains either 1 or hundreds of child sexual abuse images and videos. Within the imagery pertaining to that URL, the IWF analyst will record the youngest age they view, and the most severe category depicted using the methodology above.
Some URLs actioned by the IWF are recorded as ‘gateways’ to child sexual abuse material, which means they don't directly show this imagery, but provide a direct link to it. Hosting data for the gateway and the URL showing the criminal material is recorded.
If an IWF analyst sees an instance of ‘self-generated’ content within the imagery pertaining to the URL in question, they will record the URL as containing self-generated material. The actual amount of self-generated content pertaining to one URL may be one instance, or multiple instances. In some webpages or downloadable files, it can be common to find a combination of 'self-generated' content and not 'self-generated' criminal content.
URL analysis tag is used in this report when we are only referring to URL hosting data. When there is no URL, we will use the ‘report analysis’ tag as this demonstrates we are including data on the additional reporting methods, such as child reporting service and or newsgroups.
When our image assessors review criminal imagery they always record the most severe category of sexual abuse visible. For individual images they also record the age and sex of each child in the image. For videos, only a severity assessment is given.
Below are the definitions of the dataset tags you will see throughout this report.
Image analysis
Image analysis
Imagery that is downloaded by our analysts is assessed by our image assessors.
Image analysis is distinct from URL analysis. Our image assessors are able to record each child present in an individual child sexual abuse image. This is different from the record made by an IWF analyst assessing a URL, where they record only the youngest age and most severe category seen within the entire collection of imagery associated with that URL.
For every image they view, the age and sex of each child is recorded, this is also referred to as Multichild classification. One image can feature one child, or many children and all these details are recorded. If one image features more than one abuse scenario, image assessors will record the most severe category seen in the image only.
Collage images are also seen. These comprise of multiple different images joined together as one image. Each image within the collage can depict a distinct child involved in a distinct sexual abuse scenario. To support image assessors’ welfare and to ensure efficiency, the IWF gives images of this kind the most severe category assessment only, but no individual child data is recorded.
Sexual activity metadata is also recorded for images. This provides a more detailed tag which helps to identify the sexual activity seen within the category recorded. For example, for Category A we might record a tag of bestiality, penetration or sadism or degradation. This helps to align our assessments with different international legal requirements across the world.
Video analysis
Video analysis
Videos depicting child sexual abuse can vary in length, from seconds long to over an hour of abuse material. They can feature just one victim, or a compilation of many child victims from different abuse scenarios. To support efficiency, and to protect the wellbeing of the Image Classification Assessors who view the material, we attribute only one assessment to these criminal videos: the IWF gives images of this kind the most severe category assessment only, but no individual child data is recorded.
Images and video analysis
Image and video analysis
Where we can, we show image and video data together for severity assessment to provide a complete picture across all media formats. This enable us to show all category data together, this is the only assessment field that is recorded in the same way across all image formats.
When IWF analysts review URLs that show child sexual abuse they can be images or videos all of which will be downloaded and assessed by our image assessors.
Multichild analysis
Multichild analysis
Many images of child sexual abuse depict multiple children, and image assessors record the age and sex for every child that is visible in an image. By recording child data for all child present in each unique image, we are able to more accurately represent the number of children seen in this criminal imagery.
For efficiency and welfare, we do not record ages of children present in video content, or multiple-image collages. For that reason, children present in videos and multiple-image collages are not represented in Multichild analysis data.