If I'm being honest, data collection is something I genuinely struggle with. Not because I don't believe in it. When I actually do it consistently, I feel it. There is nothing like having real numbers and updated probes to show that what you're doing is working. It feels like the work finally has evidence behind it.
But here's the part nobody talks about enough: I struggled because I didn't always know how to score what I was hearing.
A child working on /ʃ/ produces something between /s/ and /ʃ/. I can't give them a plus. But a minus feels like I'm erasing something that actually happened. Something that matters.
That in-between production? It does matter. A lot.
The Problem with Plus and Minus
The +/- system is simple. That's the appeal. Correct or incorrect. Right or wrong.
But speech sound development isn't binary, and a system that treats it like it is will mislead you.
Research from Munson, Schellinger, and Carlson (2012) puts words to something a lot of us have felt intuitively. Traditional phonetic transcription, including the correct/incorrect system most of us were trained to use, often fails to capture the gradual changes that happen during sound acquisition.
When we rely on transcription alone, a child's progress can look like this: wrong, wrong, wrong, right. But that's not what actually happened. What actually happened was: wrong, slightly less wrong, closer, even closer, right. And we missed all of it.
Phonetic transcription can suggest sudden acquisition. Kids don't suddenly acquire sounds. They approach them.
What's Actually Happening Between Wrong and Right
Sound acquisition is gradual. Along the way, children produce what researchers call intermediate forms.
A child working on /ʃ/ might produce [sju]. That is not an /s/. It's also not a /ʃ/. But it is closer to the target than where they started. That production is evidence of learning.
A correct/incorrect system marks it wrong. And if we're only looking at the score column, we miss the shift entirely.
Munson and colleagues found that a visual analog scale gives clinicians far more useful information than a simple plus or minus. Rather than forcing a binary judgment, a visual analog scale lets you rate where a production falls on a continuum. How much does this sound like the target? Rate it.
That rating, averaged across a session, becomes a data point. String those data points together over time and you get a progress curve that actually reflects what's happening in your student's speech.
What Visual Analog Scaling Looks Like in Practice
I want to be clear: this is not a complicated system.
Munson and colleagues describe it as straightforward to implement, with high intra-rater reliability and strong correlation with acoustic measures. You don't need software. You don't need to print a new form every session.
What you need is a scale, descriptions of what each point means, and a consistent way to record the average.
Here's an example using singletons:
- 1 = true phonemic substitution (e.g., [t] for /s/)
- 2 = not quite a true substitution
- 3 = between the target and another sound
- 4 = close to the target
- 5 = acoustically correct production
You rate productions across the session, record the average, and that average becomes your data point. The child who scores a 2.8 one week and a 3.6 the next? That is progress. It doesn't show up in a +/- column. But it's real, and now you can see it.
The Scales That Made All The Difference
I adapted these forms with permission from the work of Jen Richard, MA CCC-SLP at SLPath.com to work with the Lingua Tokens line.
For singletons (like /s/ and /l/), the scale runs from 1 to 5.
For 2-element clusters (like /fl/ and /fr/), the scale runs from 1 to 7, allowing for more nuance in capturing where a production falls along the cluster acquisition path.
For 3-element clusters (like /str/ and /spr/), the scale also runs from 1 to 7.
Each form includes descriptions for every point on the scale so you're not guessing. You listen, you rate, you record the average. That average is your data point.
Why Your +/- Data Might Be Selling Your Students Short
It works for individual sessions. It also works for groups. If you have four students and two sessions a week, you can rotate through them. Every student gets solid data every two weeks without anyone losing their mind.
A Note on the Research
Munson and colleagues are clear that visual analog scaling still needs broader empirical validation in clinical settings. They are not presenting it as a finished, fully-proven system. They are advocating for clinicians to try it, document their experiences, and contribute to the growing evidence base.
I share that framing here. This is a tool I believe in because it fits what I know about how speech develops and what I've seen in my own sessions. I'd encourage you to try it and decide for yourself.
Download the Free Scale
The Data Collection Analog Scale is a free download at https://lingua-tokens.com/collections/freebies. No strings. Just a resource I wish I'd had earlier.
If you try it, I'd love to hear how it goes. Find me on Instagram at @linguatokens
Reference: Munson, B., Schellinger, S. K., & Carlson, K. (2012). Measuring speech-sound learning using visual analog scaling. Perspectives on Language Learning and Education, 19(1), 19–30.
0 comments