Rethinking Data-Collection in Speech Therapy - Part 1

Any SLP I know collects data. Whether we do it for ourselves, or we do it because we’re required to by the powers that be, we all do it. And so we should. However, MORE data is not necessarily BETTER. In fact, the opposite can be true. Here’s why.

Good data is…

VALID: it measures what it says it does

SENSITIVE: it will show change when change is being made

REPLICABLE: another person can look at it and know what it means (and then collect the same data)

RELIABLE: we can collect data over multiple sessions and know we’re looking at the same thing

In short, we can accomplish this by being REALLY clear about what we’re collecting data on in a given moment, and then sticking to it. If we say we are collecting data on the “production of R in the initial position of single words, given a model only,” then that’s what we need to do. No extra prompts, no feedback, no sentences, nothing that isn’t what we said we’d do.

Makes sense right? Right.

Sounds like great therapy right? Not so much.

Good therapy is…

DYNAMIC!

We know that as the fantastic SLPs we are, we are reassessing our targets, level of input of support, and the responses we require ALL. THE. TIME. Every second I’m doing therapy, I’m trying to find that sweet spot between not too easy and not too hard. It’s literally my JOB to do that.

On some productions, I give feedback, on others I don’t. Sometimes, I have them try the word in a sentence. Sometimes we break it down and really drill a specific tricky syllable.

But of course, while being dynamic during therapy makes me a great therapist, it really messes with my data.

So I’m stuck with a catch-22 where I have to collect data…but great data means crappy therapy, and great therapy means crappy data.

So what do I do?

I stop trying to do both at the same time.

Instead of trying to collect a million data points across the entirety of every speech session, I’ve changed the way I think about data and therapy.

At any given point in time, I am EITHER collecting data, OR giving therapy. Not both.

I have on either my data hat or my therapist hat. Not both.

I’m either taking measurements or making change. Not both.

This doesn’t mean I do less therapy (in fact, I probably do MORE).

This doesn’t mean I have worse data (in fact, it’s BETTER).

It does mean that I designate specific sessions, minutes, or stimuli where am measuring progress and I do everything I can to preserve the reliability of my data…by resisting the urge…just for a little bit…to get in there and elicit change. Yes, believe it or not, I actually shut the heck up. I let them show me what they can do, track what that is, and keep my mouth shut.

You data should measure the child's performance, not yours.

And then, I put down my clipboard and become the world’s most engaged SLP. Need a little extra practice on that sentence? Fine, let’s do it. Nailed that production? Awesome! Let’s see if you can do the next level. Struggling with this concept? No problem, I have another idea we can try. I can focus 100% of my attention on the student and their needs, without worrying about checks and minuses.

Remember, your data should measure the child’s performance, not yours.