And, yes, technically, they’re right.

However, that misses a big issue. There are 26 letters in the alphabet. There are 460ish GPCs (grapheme to phoneme correspondences) in English. A grapheme, remember, is a single letter or combination of letters (like “a” or “ough”). A phoneme is a single sound (there are 43 or 44 depending on your region). A GPC is the relationship between the two. So the relationship that “a” makes the sound “aaaa” and “ough” sounds like the long O when in a word like “though.”

The issue here, is that the same grapheme can map to multiple phonemes. The easiest (and most obvious) is that the vowels can be both long and short. But what about the words hear and bear? In these, the “ea” maps to a different sound in each (long e in hear and short e in bear).

Some diehard phonics people will likely argue that you can teach all sorts of rules around word origin, or how the other letters in the word affect sounds (because hear and heart also use different sounds). But, let’s be honest, you’re never going to be able to read anything if you before you can learn any new word you have to first know everything about it’s origin.

Side note: One of my favourite examples of phonics and how crazy it all is are the words colonel and kernel. They are pronounced the same! Even though the first half of each word is so incredibly different.

While phonics are incredibly helpful and needed (I am not arguing against phonics, I’m just arguing against phonics being the be all/end all), being able to read well means that you have been creating the mappings in your brain between words and their sounds/meanings. You are not decoding each word you come across one by one. You are only decoding words the first few times you come across them, until you’ve encountered them enough to have that mapping created.

Now, that could sound like support for whole word teaching (it’s not). What it is, is acknowledgement that teaching phonics will only get you so far. Even research has recognized that you don’t need to learn every single GPC. Instead, it has shown that you should aim to know the 64 most common (at minimum)1. That once you have those, the rest you can/will learn as you encounter words that use them. But that because so many graphemes can map to multiple phonemes, the best thing to do is to focus on the most common or frequent form and teach those first.

I feel like this is something most of us intuitively know. I mean, there’s the long running meme/joke about realizing your pronunciation of a word is way off because you learned it from reading. That’s a clear indication that you used the phonics knowledge you had and came up with a reasonable “guess” to the word. And you know what it? It doesn’t matter if you didn’t pronounce it properly the first time (or even million times) in your head. And when you do find out the correct pronunciation, you can correct yourself and move on.

I’ve been thinking about all of this a lot as I’ve been trying to figure out what order of GPCs I want to use in my little program. I had heard many swear by the UFLI curriculum, but when I looked into it, it didn’t really match what I expected. Many of the high frequent GPCs I expected come quite late, and some that I know have low frequency came much earlier. I looked at some other versions as well, and nothing really “clicked.” I think I’m going to go back to these 64 and start from there. And the 100 high frequency words.

For work recently, I’ve been doing a lot of digging into edtech apps and also standardized tests. One of the big things that I keep encountering, is that they all want to talk about how they’ve normalized their data based on all their players. So you can see how well your students are doing versus everyone else from the same grade (and often even the same time of year).

The thing is, we’ve been lowering the standards for success for years (way before COVID) and therefore, the average 3rd grader today is achieving a lot less than the average 3rd grader 10 years ago. And so adjusting the norms to todays students, just make students look like they’re doing okay, when in fact they’re achieving less, and doing worse.

I think norms are generally useless. Mainly because I think the benefit they bring is only about being able to brag about percentiles, and have no connection to actually adjusting learning outcomes. If you care about outcomes, you’d stop caring about what rank a student is, and instead just focus on what they know.

I would like us to be aiming for every kid to be able to master all the standards for grade 3, not just 50%. Now, we can get into a discussion about all “what about” questions. And sure, okay, maybe 100% is unrealistic to ever fully achieve, but that doesn’t mean we shouldn’t set the bar high.

The the more we continue to be okay because our students are doing “better” than others, and not actually evaluating what they know, all we’re going to do, is 10 years from now, have students that are doing worse, but getting the same percentile ranking.

It’s well past time we start expecting more.

Do you remember mad minutes? I have pretty clear memories of doing them for multiplication. I remember loving them, because I loved math and liked the challenge. I particularly liked it when my teacher would instead do an oral version where she’d read out the questions. Those I always got 100% on, because if I didn’t know it immediately, I’d write down the question and come back on an easier one. Yeah, I was that kid.

Lately, as I’ve finally got back to playing around with my little pet (edtech) project, I’ve been thinking about how important it was to do these timed tests. I originally was thinking about it from the learning to read space (which is where my head normally is), specifically about fluency. Because one part of the definition of reading fluency is reading speed.

(I also had some recent conversations I’d had with one sister and a nephew about the multiplication times tables and mad minutes.)

I’m a firm believer that elementary school needs to be focused on mastering skills. And I really do mean mastering. You are building your foundation for all future learning. We often talk about the shift from learning to read to reading to learn, which emphasizes the importance of mastering learning to read before this shift. We don’t talk as often about the importance of early math skills, and how much each new grade builds upon the previous. And how if you don’t just know that 3x5=15, but have to pause and think about it each time, how much harder it’ll be to keep up each year.

Having knowledge that you just know – knowledge that doesn’t take you time to recalculate or to slowly decode, but knowledge that is immediately available at your fingertips – allows your mind to think about and solve bigger problems.

Just because you can look things up or use a calculator, doesn’t mean you shouldn’t be working to have base knowledge down pat.

Quick Math and Reading Skills

All this to say, when I finally got back to my pet project, I’ve been working on my own version of “mad minutes” although I’ve stretched the idea. I started from the premise of making one for multiplication, and I realized that to cover the times table from 1x1 to 12x12 there are 78 distinct questions when you ignore ordering (1x5 == 5x1). So, I thought, why not just ask all 78 questions at once and just randomly flip the order of each factor?

From there, I ended up with 4 math versions:

1. Single Digit Addition (1+1 to 9+9)

2. Single Digit Subtraction (1-1 to 9-9)

3. Multiplication (1x1 to 1x12)

4. Division (1/1 to 144/12)

The first two have 45 questions and the last two 78. For each of the four, I made 2 versions — one where you see all the questions at once (the image at the top) and another where it shows them one at a time. For each version, it corrects as you go. For the one at a time, it requires you to get the correct answer before moving on.

You can play around with them here (yes I don’t have it at a nice url yet).

I then circled back to reading, and have been slowly working on that side. So far, there’s only one version available to, which cover the top 100 high frequency words (split into two sets of 50). Right now it’s set to 5 minutes, which feels too long for 50.

For the reading version, you listen to the word, and then have to pick the correct written word. It’s also available at the same site. (I’ve gone back and forth on whether this should be you see the word and then have to listen to 4 different options to pick the right one. And that version may still come about.)

I want to do some more that cover things like CVC or CVCe words. Or even just the base GPCs of a → ‘ah’.

A Starting Point

None of these are a full solution. They don’t teach. They are just a way for someone to practice a skill until they can do it quickly.

But, they’re a starting point for my project. When I think about early reading skills (or foundational math skills), I think a lot about how practice volume matters. That it’s important to stop looking for quick fixes and instead focus on real solutions. And real solutions when it comes to learning, mean spending time on intentional practice. And the more you practice, the better you get.

None of my little skill “tests” above, take more than 5 minutes to do. Imagine how much better you would get if you did one once a day for a month. Or how much better your students or kids would get.

1 month of the multiplication tables would take only 90 minutes (3 mins time 30 days). But at the end of it, you would answer upwards of 2,340 multiplication questions! Small investment, but big results.

I had done a lot of research into this area back when I made an app for pre-k to grade 2 students focused around the same topic. And I went back to a lot of that research again.

I had already heard about sight words from when my oldest niece and nephew had gone through Kindergarten. And I knew about the Fry word lists from a previous app I had worked on. But the more I dug into it, and the more research I read, I realized how much “we” had been swinging from phonics first to whole words as a reading strategies. And that if we really wanted to focus on getting people reading fast, we needed to take a multi-prong approach.

And, in the years since, I believe this even more than I did then (if possible).

So, as I started on this project, I knew I was going to tackle high frequency words as well as phonics and decodable words.

I went back and found the paper by Jonathan Solity and Janet Vousden. I love this paper because they give you two great starting points. First, because they include a list of the top 64 most frequent GPCs that are most crucial to learn to start decoding quickly. And second, because it also includes a list of the top 100 high frequency words. These two pieces form the foundation of my thinking.

GPCs

GPCs, or grapheme–phoneme correspondences, is the connection between a letter (or grouping of letters) to a single phoneme sound. The word “rug” is made up of three, as each letter in it is tied to a different phoneme sound. The word shrug is made up of 4, as “sh” is tied to a single sound.

What makes learning GPCs hard, is that the same letter(s) can map to more than one phoneme. It’s why we the word “read” can be read two different ways. The “ea” can be mapped to the “eh” sound or the “ee” sound.

In fact:

Gontijo, Gontijo, and Shillcock (2003) analysed 160,595 different word types (the number of unique words, i.e. the number of different words) and found that they can be represented by just 195 graphemes and 461 grapheme–phoneme associations.

Yikes! That’s a lot of possible combinations. However, as I mentioned above, the paper highlights the 64 GPCs that are the most frequent. If you know these ones, you can start to approach and decode a lot of words. Most words, really. Because the thing is, not all GPCs are created equal. Some are found very infrequently in words (like “dge” in fridge or judge). For these ones, it makes less sense to spend time teaching these up front, as compared to the basic sounds behind most constants and the long and short versions of the vowels.

High Frequency Words

High Frequency Words (HFWs) and Sight Words often get confused and tangled, but they do actually mean two different things. Sight words, are words that use tricky/rare GPC mappings, and so aren’t easy to sound out early on (think words like “was” or “the”). High Frequency words, on the other hand, get their name specifically because of how frequently they arise in text. The top 100 I mentioned above, account for about 50% of the words in most texts. 🤯

While it’s important that people learn how to sound out words and decode them (and need phonics to do so), mastering these top 100 words so that you just know them without thinking, is a crucial step to being able to read fluently and quickly. If you can already read half the text, you’ve unlocked a huge portion. And those top 100 give a lot of help to the context of what you’re reading.

My Approach

I knew from the start that I wanted to tackle this project by starting with a focus on the basic GPCs and HFWs. But, I also wanted to take it farther. Knowing that b makes the sound of “buh” is good, but you need to also put it together to start decoding words like bug or bat. So, that meant added in a third category for me, decodable words. And then, I decided that I’d actually like to take it even one step farther, and bring in sentences. I didn’t want to go past sentences because a) I already do reading comprehension at that level at my job (no compete) and b) it’s a much much bigger than the other four groups I mentioned combined and well, this is being done just by me as a hobby. So I wanted to try and keep it more manageable.

My first plan looks something like the chart below. Where players would start with both basic GPCs and HFWs available to work on. And then as they learn enough GPCs, it would unlock beginner decodable words. And when they have enough of those and enough HFWs, it would unlock sentences.

I’m not 100% sold on the plan, mainly because I think I’m going to divide it into more granular levels. There’s a sense of accomplishment to getting things done, and if the task is too big, it can feel too overwhelming and demotivating. So I don’t want someone to go in and feel they need to learn all 100 HFWs before they’ll be able to unlock sentences, and so on. So this will be modified, but gives some basic idea of the structure — “learn enough of the parts needed to master the next part, and it unlocks.”

Anyway, will have to see how it goes. For now, I’m just getting the basic app setup. I’ve got it to the point that it has the base GPCs and HFWs and you can do some “listen and read” and “read and listen” matching. But, more on that in the next post.

I’ve decided, as a way to try to hold myself accountable for working on it (and to remember why I’ve done what I’ve done), I’m going to use this to keep track of my progress. What I’m working on, why I’ve made decisions I’ve done, and the research I’ve done along the way.

I’m very much focused on the “make it work” and trying hard not to get caught up in the “make it pretty” (a very bad rabbit hole for someone not artistic).

For this app, I’ve decided to stick with the topic I’ve been more closely working around, which is reading. But I’m taking a bit of a different tack than I have in the past. Instead of focusing on kids (k12 is my favourite area), I’m going to focus on adults. And I’m going to focus on the foundational skills around decoding. That means helping people to understand the link between letters (graphemes) and sounds (phonemes) such that they have the skills needed to approach a new word and figure out what word it is (decoding).

I did work in this area almost 10 years ago (wow time flies), and at that time, I was focusing on young kids. There were certain things that needed to be taken into account with that audience that I aren’t true with adults (and vice versa).

For example, with the app for kids, there was an assumption/expectation that the kids were in school and thus supported in their use of the app with in-class learning. With adults, I’m going to assume that they would be choosing to use this on their own and have no support outside of the app.

With kids, there were issues around fine motor skills (would they be able to successful drag and drop with a mouse?) that should be less of an issue with adults.

And with kids, everything needed to be very fun and colourful. Now, I’m not saying adults want or appreciate being bored, but I do think the expectation level on fun (and what counts as fun) is different.

I’m assuming (and hoping) that adults who would be interested in what I’m creating, will be more self-motivated. And so I can structure the app and learning process with different goals than I would for kids. But… this is something I will need to explore more as I work on it.

Next post, will be about the basics I’m starting from.