30 March 2016

Kick It Up a Notch

Muriel Rukeyser says that The universe is made of stories, not atoms. As a scientist, I might disagree. But as an educator working with data and children, I am inclined to draw that same conclusion.

I tell stories with data all the time. It's my responsibility to look for connections and identify patterns. For the most part, however, it's focused on those tales that I am asked to tell. What are the characteristics of students who are never able to meet the standards? Does better attendance lead to better achievement outcomes? Which schools or programs best support closing the achievement gap?

We are data rich, and information poor, in my district. Perhaps you are in yours, too. So for next year, I am planning a new challenge: Ten data stories in ten months. But I don't want these to be our regular stories, such as how students performed on the spring assessments. Instead, I am looking to use some untapped sources of data.
  • For example, a month of use of the district board room. Are there connections between our stated goals and priorities and how we use this resource? How do we use time, which seems to be the most precious commodity? 
  • Or perhaps I could dive into the class of 2017 with the lens of those students who spent all 13 years in our district. 
  • What would a text analysis of a month of school newsletters reveal?

I need at least seven more ideas like these. I'd like a few "spares" because I don't know how many will be viable once I get into the actual project.

These stories will be told in two ways. First, I plan to use a bulletin board at the district office to display some offline data. I like the idea of data you can touch. Can I use different grits of sandpaper to indicate different levels? What about some 3-D paper techniques, like you'd see in a pop-up book, to illustrate change? There will be a companion web page for every data story---something that those with online access can explore to get details.

I hope to get a jump start on these over the summer. Ten is an ambitious goal...perhaps I might have to back off from that. I chose ten because we have school during ten months of the year...and we have ten schools. Perhaps December and June will be "short stories" given the number of actual school days in each.

I am excited to kick things up a notch. I have seen any number of schools with a "data wall," but none who have data stories. What stories would you want to see about your school or district? Where would you shed some light or reveal some data? What are the questions no one seems to ask about your classroom or students?

26 March 2016

Increasing Understanding with Data Displays

Commenters on the last post are looking for suggestions of better strategies for highlighting student data. As students, we often learn rules for communicating with text. We know to put a capital letter at the beginning of a sentence, and end a sentence with a punctuation mark. We learn about grouping sentences of a similar topic together---and even signalling those groups/paragraphs with indentation. We have an entire grammar system when it comes to text. It's all built on enhancing the dialogue between a writer and a reader.

Visual communication is not so different. There are some basic rules, although we don't seem to teach them in the same way as we do for writing. The purpose behind these principles is the same as for text: We are trying to enhance understanding between the author and the audience.

If you were given a page of text and a highlighter, and then tasked with identifying the most important ideas, would you highlight every word on the page? Probably not. Why? Because when everything is highlighted, it is as if nothing is highlighted. Your brain cannot identify what is most worthy of attention. So why do so many educators insist on doing things like this with their data?

From strictly a visual processing standpoint, this approach has a variety of problems. The most obvious is the non-stop highlighting. If one of our goals is to identify students in need of support or identify patterns, the approach at the left is not helpful. We should be as strategic with our highlighter here as we are with text. Only those students in need of greatest support should catch our eye. This is not to say that heat maps can't be useful as data visualizations, but most education related data doesn't connect to their purpose. We need to match the right visual with the right goal, just as we might match our writing style (informative, persuasive...) to the outcome we wish to achieve.

The second problem with this example is that it has both numbers and colors. Working memory can hold about 7 items. There are 38 numbers in the Fall and Winter columns---far beyond what we could remember, let along compare in our heads. It's great to use highlighting to reduce that cognitive load, but it also means the numbers need to be hidden so we can look for patterns. When we leave the numbers there, we start devoting mental processing to things like figuring out cut scores or how far away a particular student was from the next achievement level. We're distracted from the more important conclusions about student performance.

I could put on my ranty-pants about the color choices here, but if you're interested on color perception and how it relates to your design choices, you can visit my post on my other blog (which is devoted to data viz for the classroom). Head on over there anytime for all sorts of ideas to transform your data.

Another issue with this example is that every cell has lines printed around it. This is called enclosure, and like signals such as line length, color, or position, your brain "sees" it as a way to pay attention to what's important. (To learn more, hit teh googles for pre-attentive attributes.) Enclosing everything is as confusing as highlighting everything. Let's look at enclosure another way.

Here's a typical data table. There's some students and then some item analysis for each student. How long does it take to recognize the patterns here? What if we take off some of the lines and, instead of highlighting, simply grey out the zeros?

One might argue that this type of representation isn't as sexy as the circus-like highlighting in our original example, but it sure makes things a lot easier to understand.

And that's really the bottom line with data visualization. It is intended to be a bridge between the raw data and meaning that we elicit. If the visualization gets in the way of that, then we are at risk of making the wrong conclusions or even taking the wrong actions based on those data. We can apply some principles to our numbers, much as we do with our words, to help our audience---even if it just a party of one.

13 March 2016

The Right Kind of Sticky

At one point in my post-secondary education, I took a class on working with gifted students. As college students, we took turns modeling different types of differentiated lessons with our peers. I really don't remember much about the details, except for this one example where the "teacher" asked us to think of words that "described how the soldiers at Gettysburg felt." And the first word that popped into my head was sticky. Apparently, this was not the sort of feeling that the facilitator had in mind.

I was thinking about this story earlier this week because I have been pondering sticky ideas---and, in particular, what we do with the ones that have become old and gummy, but are still hanging on long past the time we have moved to something better.

I had the privilege of attending the Tapestry Conference this week. As far as I could tell, I was the only K - 12 educator there. That's not surprising, given that it was not a conference for educators. It's goal was to bring together people who use data to tell stories. The most common question I was asked was "Are there others like you---in other districts?" This was a hard question to answer. Yes, there are people who work with assessments in every district (no matter how small), but the data part only comes into play once districts reach a certain size...and even then, I haven't run across very many who tell data stories.

Instead, I see lots of spreadsheets that are coded in shades of red, yellow, and green. This is a sticky idea---and one which might have been the best option 8 - 10 years ago, but it is certainly not considered best practice (let alone effective) now. And yet, it's so ubiquitous that I don't know how we will ever manage to shift away from it. Are we really so frightened of change that we would rather to hang onto the only thing we know than make sure we have the best option available? What does it say about us when we work in a profession devoted to learning and yet we are unwilling to learn and adapt?

I understand how comfortable it is to be in a box of your own design. Whatever passes for normal in ones' world is what gets maintained. I am certainly guilty of choosing safe over new. I'm trying to stretch more this year. Tapestry was one way to do that, and Eyeo will be next. There are all sorts of education-specific conferences I love to attend. I learn a lot of things that support my day-to-day work. But I can't pretend that there isn't more out there to explore or learn. I can go to a data conference and find things to apply to my work. I can learn to unstick myself from routines and ideas, at least for a little while. How do I help others do the same?