Writing photo captions: the five Ws and the ethics

A photograph without a caption invites the viewer to guess. For organisations, a wrong guess can matter: a health worker mistaken for a patient, a training session read as an emergency, or a place name that puts someone at risk. Caption writing is the point where a picture becomes evidence.

Ethical communication guides treat captions as essential, not optional. The Dóchas Guide to Ethical Communications advises including captions with images as a rule, unless there is a specific and justifiable reason not to.¹ UNICEF's ethical reporting guidelines ask organisations to always provide an accurate context for a child's story or image, and say that complete captions should accompany all multimedia.²

The IPTC, which maintains the photo metadata standard used by cameras, news agencies and image software, describes the caption as covering the who, what (and possibly where and when) and why of what is happening in the photograph.³ It can include people's names, their role in the action and the location.³ In practice, each caption should answer five questions:

The level of detail depends on the image. The IPTC notes that it varies between documentary and conceptual photographs.³ A field photograph used as evidence of a programme needs all five Ws; an abstract texture used as a page background does not.

• Who: full names and titles of the main people, if they have agreed to be named.
• What: the action in the frame, described plainly.
• Where: the community, district and country, as precisely as is safe.
• When: the date the photograph was taken.
• Why: one line of context that explains why the moment matters.

Journalism training offers clear rules that work just as well for organisations. IJNet's guide advises complete sentences in the present tense, or at least a present-tense first sentence if the rest cannot stay that way.⁴ Identify people and places clearly, with professional titles and the formal name of the location; in group photographs, identifications typically go from left to right.⁴ Include the date the picture was taken, and mark older images as file photos.⁴

Keep it brief: most captions are one or two short declarative sentences, with a third only when complex context is needed.⁴ Avoid stock verbs such as 'looks on' or 'poses', do not guess what someone is thinking or feeling, and do not call a picture beautiful or dramatic; let the image do that.⁴ Above all, spell names correctly.⁴

Compare 'Happy mother at the clinic' with a caption built on the five Ws: 'Community health worker [name] weighs a newborn at a health post in Nyamagabe District, Rwanda, on [date].' Then add one sentence of context, supplied and checked by the programme team.

Captions get lost when images are copied, emailed and uploaded. The answer is to store them inside the file. IPTC photo metadata has a Description field for the caption, plus fields for the description writer, creator, credit line, copyright notice, date created and location.³ The IPTC recommends starting the location with the most specific level, the sublocation, and working up through city, province and country.³ It also separates Location Created, where the photograph was taken, from Location Shown, what the image depicts.³

A caption is not alt text. The IPTC explains that alt text is read aloud by assistive technologies, while the caption is usually shown below the image and provides the facts about it.³ It recommends keeping alt text short, a couple of sentences at most.³ On a web page, the W3C's alt decision tree advises an empty alt attribute when an image's content is already presented as real text nearby, so screen reader users do not hear the same information twice.⁵

Before you write a name, check that you may. Dóchas advises establishing and recording whether participants wish to be named, and always acting accordingly.¹ It recommends naming regions, localities or communities, and recording names, situations and locations as accurately as possible, while keeping anonymity where it protects participants.¹

Children need particular care. UNICEF says to always change the name and obscure the visual identity of any child identified as, among other categories, a victim of sexual abuse or exploitation, a perpetrator of abuse, or a child charged with or convicted of a crime.² It also says that showing a child's home, community or general whereabouts must not endanger them.² A caption naming a child's village can undo everything a photographer did to hide a face.

Captions shape how people are seen. UNICEF warns against categorisations or descriptions that expose a child to negative reprisals, and lists HIV status among the labels that require a child's identity to be obscured unless there is fully informed consent.² Describe people by what they do rather than by a condition or a need: a farmer, a student, a nurse.

Dóchas cautions against presenting a small slice of a situation that reinforces stereotypes, and against generalising diverse people as 'they' or 'these people'.¹ If a caption is significantly edited, it advises saying so, and editing should not change the original meaning.¹ The same honesty applies to the image itself: IJNet advises labelling a manipulated photograph as a 'photo illustration' in the caption or credit.⁴

Captions are easiest to write on the day. In our own practice, we note names, spellings, roles and consent decisions during the shoot, then add captions to the IPTC fields before images leave the edit. Ask someone who knows the programme to check the context line, and fill in the Description Writer field, which records who most recently wrote or corrected the caption, so questions can be traced later.³

• Who, what, where, when and why answered
• Present tense, one or two short sentences
• People identified left to right, names spelled correctly
• Date taken included; older images marked as file photos
• Consent to be named checked and recorded
• No detail that could locate or stigmatise a child
• Caption saved in the IPTC Description field
• Separate, shorter alt text written for the web