Specific Learning Disorders in Reading (commonly known as Dyslexia) affect approximately 10 – 15% of the population (the official estimates are between 5 -12% but have been reported as high as 20%). The International Dyslexia Association gives the following definition:
“Dyslexia is a specific learning disability that is neurobiological in origin. It is characterized by difficulties with accurate and/or fluent word recognition and by poor spelling and decoding abilities. These difficulties typically result from a deficit in the phonological component of language that is often unexpected in relation to other cognitive abilities and the provision of effective classroom instruction. Secondary consequences may include problems in reading comprehension and reduced reading experience that can impede growth of vocabulary and background knowledge.
https://dyslexiaida.org/definition-of-dyslexia/
Although the term dyslexia was first used in 1887, this official definition was not established until the 1994 Definition Consensus Project.
As mandatory free public school education was established in the US over the 19th and 20th centuries by each state (Massachussets was the first in 1851, Mississippi the last in 1917) learning to read became the single most important skill any citizen could acquire in school. Learning to read is a skill made up of many complex neurological brain processes. This visualization, created in 2001 by Hollis Scarborough, breaks down the many elements a person needs to master in order to read proficiently:
by Hollis Scarborough, 2001
While for some the process of learning is relatively easy, for many people one or several of these elements is not taught to them effectively, which affects their entire reading proficiency.
The myriad of complex reasons that reading is often taught in ineffective ways for large percentages of can be better understood by listening to this podcast series called “Hard Words” by Emily Hanford from APM Reports. The accompanying article includes the graphic below from The Nation’s Report Card of students proficient in reading in the US, which puts reading proficiency below 40%. Also note that in several recent years proficiency actually went DOWN between 4th and 8th grades.
There is currently a movement in education to approach teaching literacy to all kids using the science of reading, which argues that teachers must understand the different brain processes involved in learning to read and be trained in techniques scientifically proven to give kids the skills needed from the beginning to be strong readers.
I have been in contact in recent months with two dyslexia advocates in regards to my eight year old son’s dyslexia. The training and experience that advocates have can range but their primary role is usually to support parents in getting educational or other support services for their kids from their school system. Often this involves trying to distill a lot of technical and medical knowledge into concepts that lay people can understand. At some point both advocates gave me infographics- one on structured literacy and another on cognitive profiles – that I found confusing, misleading, overly busy, and ultimately counterproductive to conveying the information they wished to.
I decided to redesign both of these graphics using their existing datasets but adhering to the guidelines of what makes good data visualization. I wanted to see how transforming the design would convey the data differently and, hopefully more clearly. However, before I start criticizing their designs I want to point out how underfunded disability education research is, and that advocates and teachers are saddled with wearing every hat imaginable in their efforts to support students. So I would like to clarify how important the contributions of the people who designed these are and how privileged I feel to be able to work with their data.
Seeing, Decoding, Understanding: Reading and Vis
There are some interesting parallels between the science of reading and data visualization. This chart describes the cognitive processes associated with reading:
https://www.edcircuit.com/cognitive-skills-reading/
In “Readings in Information Visualization, using vision to think” (Kard, S. T., Mackinlay, J. D., & Scheiderman, B., 1999), the authors conclude with the following elemental benefits of using information visualization to convey complex and vast amounts of data.
This medium has promise for five reasons:
It brings increased resources to the human in the form of perceptual processing and expanded working memory.
It can reduce the search for information.
It can enhance the recognition of patterns.
It enables the use of perceptual inference and perceptual monitoring.
The medium itself is manipulable and interactive.
Kard, M. (1999). Readings in information visualization: using vision to think. Morgan Kaufmann.
Roughly put, the goals of both are to make complex and multitudinous processing demands on the brain automatic enough that you can get to the meaning of the content.
Visualizations
To begin with I will describe what I think are problems that both visualizations have in common, and then delve into each separately. I will show the redesigns I created and describe the reasoning behind my design choices and the impact on the charts of user feedback that I received.
My first thoughts on seeing both visualizations were of a lecture in my info vis class on “data- ink”, and that any ink on an info vis page should be in service to the “sense and substance of the data, not to something else” (Tufte, E. R. (1989) The Visual Display of Quantitative Information. Taxon,38(3), 451.) There was a lot of extra and unnecessary ink happening in both of these, albeit in very different styles. I found it hard to understand the salient points of either graphic quickly because of this.
The Ladder of Reading – Before
Figures 4 & 5- The Ladder of Reading by Nancy Young
I find these two pages incongruous – while I understand that the Ladder of Reading is about kids it’s clearly not meant to be for kids, given the dense definitions on the second page. The sweet illustration is undermining the legitimacy of the data and the scientific basis of structured literacy. I also found a “ladder” to be a confusing metaphor. Science tells us that the cognitive processes that affect reading are neurobiological. You can’t climb your way from being dyslexic to being an effortless reader. The goal of structured literacy is not to turn kids from one to the other, but to best support each child based on their existing needs so that everyone can become a strong reader.
The chart itself is boxes within a box, each of which is a gradient. I was told by one advocate that this was meant to convey a spectrum from one profile to the next, but the gradients don’t transition well from one to the next since they jump from light to dark. Also the rounded corners are unnecessary . The percentages don’t really add up clearly and the main points – about structured literacy – are tangential. The entire second page is a dense mass of text.
The Ladder of Reading – After, The Circle of Reading
I began by transcribing the data from the original chart into a very simple .csv document. The only change I made was to choose single percentage numbers for all of them that would total 100% since the ranges would not work in Tableau Public, the visualization platform I was using to create the new chart. After importing the csv into Tableau I chose to create a pie chart because I thought a circle would be a more appropriate visual representation than a ladder. In any school a class of kids will be loosely comprised of this makeup and it’s crucial that all students are seen as a part of the whole, that different modes of learning are not viewed on a hierarchical scale based on how easy they are to master.
I next integrated the most relevant points into the pie slices; whether Structured Literacy was essential or beneficial, and then the list of instruction types they would either benefit from or require. The colors are based on a general coloring system I was given by one of the dyslexia advocates: blue for “above expectations” green for “meeting expectations” yellow for “below expectations” and red for “far below expectations. You can also see that the warm tones in the chart equal 50% and the cool tones are also 50%, which also aligns with “benefits” and “ requires”.
The goal of this chart is to persuade parents, administrators, teachers, and school boards that structured literacy, if used as part of the general education reading curriculum will benefit every student, not just kids with dyslexia. However without structured literacy, up to 50% of any given classroom will struggle to learn to read proficiently. Pushing kids out into specialized learning support (such as individualized instruction with a dyslexia specialist) is, overall, far more expensive and complicated than bringing Structured Literacy into the classroom, where it will benefit and support all students.
I added shortened, simplified versions of the Structured Literacy definitions as a single sidebar. This makes it easier to refer back and forth between the chart and the definitions. The simplicity of the chart breakdown suggests that it is intended for a fairly broad audience, not for dyslexia experts, and I thought the text should match the level of detail in the chart. As the chart is most often used as a printed handout or part of a slideshow presentation I designed it to be downloaded as a letter size PDF. Based on user feedback, in light of the data-ink rule I removed any borders, although I left a gray box on the text to to give it a shape separate from the chart.
The dyslexia expert I showed this to expressed that she would like to see a version of the chart using a color spectrum rather than hard percentages, since this would better reflect better the range of kids abilities, which I would like to try to tackle in the future.
Six Distinct Cognitive Profiles of Early Reading- Before
Jamie received this from a PATTAN Literacy Symposium in 2018
I received this from a dyslexia advocate who said this was one of her favorite charts, but when I spoke to her further on how she used it I found that she spent a lot of time explaining to people what the chart was saying. Ideally an info vis should be able to speak for itself, to support the advocates message effortlessly.
The most confusing aspect of this chart is that it appears to be longitudinal. The horizontal axis appears to be tracking data over a progression such as time. In fact these are cognitive test names that could have been presented in any order. The lines between datapoints could represent any relationship depending on the order of the tests. When I asked the advocate why she liked the lines she said it was because they showed how certain profiles have extreme deficits in specific areas of testing. I see what she meant, but I think this was expressed incorrectly as a line graph.
Additionally I found the addition of the specific Z-score numbers for each datapoint to be redundant, especially since there is no explanation of what a z-score is on the chart. A z-score is a standard deviation from the median with the median being 0, based on a specific bell curve. The advocate was correct that the z-scores are most informative not as raw numbers but in relationship to their scores in tests across a profile. The added blue highlighter and annotations of profiles at the end of the chart further showed me that the chart itself wasn’t doing the heavy lifting in terms of conveying the message regarding the relationship between tests, profiles and z-scores.
Again I took the existing datapoint from the chart and created a very simple csv, with a column each for Cognitive profile, Z-score and Test name. I imported the data into Tableau and began experimenting with various bar charts, which I thought would be the best way to represent the data.
I created 4 initial bar charts that were each organized differently by profile or test, horizontally or vertically, with or without legends. I sent all 4 versions to 6 different potential users; the two advocates and 4 parents.
Option 1 
Option 2 
Option 3 
Option 4
Figures 8, 9, 10, and 11 – 4 table options made in Tableau
4 of the six chose Option 1, which displays the bars as horizontal. I was told by the advocate that they would never focus on the data according to the test name rather than the reading profile as in Option 2, and that the test names should be a part of the chart, not as abbreviations or as part of a color coded legend in options 2 and 3. Most found that the 4th one was difficult to read with so much vertical text. So the first option was clearly the best for a multitude of reasons. The colors help to emphasize the overall shape of the graph and separate each profile, while the bars allow you to see specific differences in tests between profiles. As with my first chart I adhered to the colors that are generally used by dyslexia advocates, but I used different shades of each color than on my first chart since they are measuring different things.
Six Distinct Cognitive Profiles of Early Reading- After
As with my first chart I wanted to add context to the data. This chart I felt warranted more detailed definitions and explanations for the datasets. Again, my final version is designed to be printed as a single 8.5×11 page, although if kept in dynamic form of a Tableau chart you can hover over the bars to reveal specific Z-score numbers.
The overall shape of the bar charts communicates that a double deficit profile scores more than twice as badly overall in nearly every test as a high performer, that a phonological deficit effects more areas of the profile than a rapid automatic naming deficit, in which the profile is the same or even better than an average performer except in rapid automatic naming, where it is far below the median. Orthographic and phonological deficits show very different areas of weakness. While Phonological deficits overall have more negative Z-scores than orthographic deficits, the phonological profile scores above the median in Letter Sound Knowledge, it’s best score, which is the lowest score in Orthographic deficit.
The chart demonstrates how the 4 profiles with deficits are all different, supporting the belief that their deficits require targeted and specific support. Moreover it helps to disavow the vague and misguided notions that many people have about dyslexia; that it’s simply people reversing letters, or more insidiously, that it’s because students are lazy or haven’t been raised “properly” to value reading.
If we are going to address the abysmal literacy rates in this country it is essential that we can universally assess learning profiles as early as possible (there is currently no universal screening system for learning disabilities in the US), we have to integrate scientifically proven methods of teaching literacy in mainstream curricula that support as many learners as possible. We have to address learning disabilities with targeted, standardized, and measurable methods that we know will support specific deficits. Clear simple visualizations of the data on learning disabilities can help to further communicate these priorities.