What's the evidence? What’s in a caption?

February 2018

What's the evidence?

What would be a good caption for this photo? How’s this: A water wheel is a traditional form of generating power. The falling water turns the wheel and rotates a shaft, generating energy for activities like milling, grinding, sawmilling, gold stamping, and pumping. /Photo by dnoob83.

Water gushing out of a spinning water wheel

Do you read captions? Did you read the one above? I never thought about whether I do, but David Ogilvy, advertising guru of Ogilvy and Mather, notes captions are read four times as much as body text (Ogilvy, 1983/2011, p. 77). Or is it twice as much (p. 107)? The point is that they’re read more, as shown in a Gallup survey (cited in Ogilvy, 1983/2011).

Film director Errol Morris describes their power: ‘If you want to trick someone with a photograph, there are lots of easy ways to do it. You don’t need Photoshop. You don’t need sophisticated digital photo-manipulation. You don’t need a computer. All you need to do is change the caption’ (cited in Kleon, 2008). Here are examples of photos with different captions.

What do captions do? That doesn’t seem to be spelt out, but from what I gather, they:

  • add information. ‘The caption should add context to the image’ (Krueger, 2017). ‘Each must quickly tell the reader what the picture itself cannot say’ (Melton, 2012).
  • guide understanding. Images can be interpreted differently or viewed in isolation. Apparently, ‘[in] proposal writing, more so than in other writing, captions serve as interpretive screens for the reader . . . . A caption should provide information; it should tell the reader what to look for and what to see in the visual’ (Woolever, 2008, p. 361).
  • intrigue. Every caption ‘should also intrigue in a way that makes [readers] look back at the picture because they just learned something they didn’t know before they read the caption’ (Melton, 2012).
  • may facilitate search engine optimization (Enzo Creative, 2013).

What evidence is there on captions? What advice on writing them? How are captions different from alternative (ALT) text? And what does it mean for your next caption?


We know that images are helpful, captions get read, and placement matters.

Images help

Houts, Doak, Doak, and Loscalzo (2006) reviewed health education research that compared text with text plus pictures. Users given text with pictures were more likely to pay attention to, understand, remember, and follow instructions. While this research did not suggest how to write captions, it demonstrated that they add information and guide understanding.

The authors also suggest that captions can help reach a wider audience than the body text. They describe an American Geriatrics Society’s booklet. While the body text was written at tenth grade reading level, the captions were written at second grade level. This meant ‘people with only minimal reading skills are able to understand this key part of the message without being able to understand the more complex explanatory text’ (Houts et al., 2006, p. 182). Interesting concept, but their article didn’t say if this strategy worked. Would less able readers have been put off by the booklet and not bothered to look at the pictures?

Captions get read

In an eye-tracking study, linguists Faria, Baptista, Luegi, and Taborda (2006) asked subjects to look at photos, which they later had to describe in as much detail as possible. Results suggest the subjects did indeed read the captions, since they used words in the captions when describing the photos. We know reading captions was not at the expense of looking at the photos, because the subjects’ descriptions included details that were only in the image.

Treat this evidence with caution, however. The authors did not say how many subjects they had, or who they were. There were few enough that the authors felt the experiment should be repeated on a larger group. Nor was there enough information on the experimental procedure for the research to be repeated by someone else, a key expectation of research reports.

Placement matters

Communications scholars O’Donnell and Willoughby (2017) looked at the use of captions to convey sexual health messages on Instagram. They placed information in several ways:

  1. In photo captions
  2. Split between the photo and the caption
  3. Embedded only in the photos
  4. Not at all – the comparison group only had background images.

You could argue that these weren’t strictly ‘captions’, since they didn’t elaborate on the photos. Rather, they were messages placed with photos. That probably doesn’t matter too much for our discussion, but feel free to look at the images and decide for yourself.

Participants deemed information most effective when it was embedded in the photos. Ooh, I can hear protests that this isn’t practical. The good news is, placing information in captions worked second best, and there was no statistical difference between the effectiveness of these two placements. (Statistical difference: whether it could be due to chance.)

Advice: Add value

There’s plenty of advice on captions, mainly for photos, but the one thing that came through time and again was to add value. The Guardian and Observer Style Guide (2016) puts it plainly: ‘Captions are an opportunity to give readers further information, rather than insulting their intelligence by stating the obvious.’

The BBC (n.d.) provides an example of adding value: ‘A caption should be more than a literal description of the picture; it should add value (eg: George Smith and Terry Jones: Long-time friends).’

What exactly needs to go into captions? There’s advice on the information needed, referring to parts of the image, placement, and punctuation.

Information in captions

Captions need to be self-sufficient, accurate, and concise.

  • Self-sufficient: Celia Elliott, technical communication lecturer, tells us scientists ‘read’ a journal article by reading the abstract, scanning the conclusions, and glancing at the figures and tables. ‘Thus, figures must “stand alone”; a reader should be able to understand the main message in a figure without having to read the text’ (Elliott, 2017; p. 3; Woolever, 2008).
    • Elliott (2017) recommends telling readers what the figures show before explaining what they mean: ‘Always give the “title” of the figure first. Tell the reader what she’s looking at before you start explaining the details’ (p. 11). For example, introduce a figure with ‘Graph of …’
    • Editor of the Northwest Scholastic Press, Rob Melton (2012), goes further, telling journalists to include ‘names, stories, dates, places, significance. The 5W’s and H.’
  • Accurate: ‘Fact-check all quantitative information such as spelling, names (of people, places, and things) and titles, and data, and double-check that you describe action or procedures accurately’, says Mark Nichol, contributor to Daily Writing Tips (n.d.).
  • Concise: The Oxford Style Manual (OSM) says to ‘delete the words The diagram shows … when the representation is obvious’ (‘Captions’, 2012, p. 309). And the BBC (n.d.) recommends ‘Picture captions in news stories should be no longer than two lines, or one line for large pictures’.

Don’t forget to acknowledge your source at the end of the caption (‘Captions’, 2012).

How do we put all the advice together? Melton considers National Geographic captions ‘excellent’ so I hopped onto their website and found a story of a ‘fairy-tale castle’ with this caption:

Neuschwanstein Castle is the most visited castle in Germany and one of the most popular tourist destinations in Europe. Photograph by Roberto Moiola, Robert Harding/National Geographic.

Referring to parts of an image

What if you need to refer to specific parts of an image? Contrary to O’Donnell and Willoughby’s (2017) Instagram research, experts don’t recommend embedding text in images (Swisher, 2012). ‘Remove or reduce labelling on the figure where possible by including explanations in the caption’ to allow copy editing (‘Captions’, 2012, p. 309; Elliott, 2017).

That’s the straightforward advice. How to do it is less straightforward. The OSM proposes using terms ‘such as above, below, top, bottom, left, right, and clockwise’ (‘Captions’, 2012, p. 310). But Elliott (2017) advocates labelling images with letters or numbers instead, and referring to the letters or numbers – if images get moved around, spatial references become meaningless.

Which advice to follow probably depends on your style guide and situation. If you only have one image or if you control what happens to the images then above, left, and clockwise may be fine. But if you’re working with several images or do not have full control of how they are arranged, you might prefer labels.


Captions go below or beside images, graphs, and charts, but above tables (‘Captions’, 2012; Woolever, 2008). Why treat tables differently? The Chicago Manual of Style (n.d.) explains:

[Often] tables have other matter that appears below (sources, notes). Often, too, qualifying information essential for interpreting the table (measures, dates, etc.) appears in the title, so it makes sense to put it at the top where the reader will see it first. A figure caption can go above, below, or on the side of the figure. Its placement is usually a design decision, not an editorial one.


Ending captions is less clear-cut than I thought, and no, it’s not a UK vs US English debate. According to the OSM, ‘Captions traditionally end with a full point, whether they are full sentences or not’ (‘Captions’, 2012, p. 309). On the other hand, the Telegraph (2008) says ‘Picture captions do not have full stops at the end’, which the BBC echoes (n.d.).

What does your style guide say?

Captions: Not alt text

Captions aren’t alt text. Alt text applies only in online environments, and in contrast to captions, it does describe the image, the way you might describe it to someone over the phone. Screen readers read alt text to those with visual impairments, sighted readers see it when images do not display, and search engines use it to index images (Moz, n.d.).

Your next caption

How will you write your next caption? Elliott tells us to try to make the figures and their captions firstly tell the story, and secondly entice the reader to read the whole article (2017, p. 1).

Will you do that? If you’d like to find out more, do read:


BBC (n.d.). BBC News style guide: C. Retrieved Dec 24, 2017, from http://www.bbc.co.uk/academy/journalism/news-style-guide/article/art20130702112133556

Captions. (2012). In New Oxford style manual. Oxford: Oxford University Press.

Chicago manual of style online (n.d.). Capitalization in titles of works. Retrieved from http://www.chicagomanualofstyle.org/qanda/data/faq/topics/CapitalizationTitles/faq0021.html

Elliott, C. M. (2017). Effective figure captions for technical documents . Retrieved from https://courses.physics.illinois.edu/Phys496/fa2017/Lectures/FigCaptions.pdf

Enzo Creative (2013). 3 lessons from David Ogilvy’s book “Ogilvy on Advertising” . Retrieved from http://www.enzocreative.com/blog/?p=2224

Faria, I. H., Baptista, A., Luegi, P., & Taborda, C. (2006). Interaction and competition between types of representation: An example from eye-tracking registers while processing written words and images . Retrieved from www.ucp.pt/site/resources/documents/ICS/GNC/ArtigosGNC/AnaMariaAbreu/E_FaBaLuTa06.pdf

Guardian (2016). Guardian and Observer style guide: C . Retrieved Dec 24, 2017, from https://www.theguardian.com/guardian-observer-style-guide-c

Houts, P. S., Doak, C. C., Doak, L. G., & Loscalzo, M. J. (2006). The role of pictures in improving health communication: A review of research on attention, comprehension, recall, and adherence. Patient Education and Counseling, 61(2), 173-190. Retrieved from https://pdfs.semanticscholar.org/51ae/e47655338b5b415b8cd28a4efd59752a3f8d.pdf

Kleon, A. (2008). The power of captions: Words added to pictures . Retrieved from https://austinkleon.com/2008/09/01/the-power-of-captions-words-added-to-pictures/

Krueger, V. (2017). 6 tips for writing photo captions . Retrieved from Poynter website https://www.poynter.org/news/6-tips-writing-photo-captions

Melton, R. (2012). Follow these simple techniques to write the perfect caption every time to intrigue, inform readers . Retrieved from Northwest Scholastic Press website https://nwscholasticpress.org/2012/09/30/follow-these-simple-techniques-to-write-the-perfect-caption-every-time-to-intrigue-inform-readers-2/

Moz (n.d.). Alt text. Retrieved from https://moz.com/learn/seo/alt-text

O’Donnell, N. H., & Willoughby, J. F. (2017). Photo-sharing social media for eHealth: Analysing perceived message effectiveness of sexual health information on Instagram. Journal of Visual Communication in Medicine , 40(4), 149-159. Retrieved from http://www.tandfonline.com/doi/abs/10.1080/17453054.2017.1384995

Ogilvy, D. (2011). Ogilvy on advertising. Retrieved from Kotui. (Original work published 1983).

Swisher, V. (2012). Writing for global readiness: What technical writers need to know . Retrieved from TechWhirl website https://techwhirl.com/writing-global-readiness-technical-writers-need-know/

Telegraph (2008). Telegraph style book: Cc. Retrieved Dec 24, 2017, from http://www.telegraph.co.uk/topics/about-us/style-book/1435309/Telegraph-style-book-Cc.html

Woolever, K. (2008). Writing for the technical professions (4 th ed.). New York, NY: Pearson Longman.