Spring is a time of renewal and, to me, it really feels that way this year. I hope it feels that way to you, too.
As the Northern Hemisphere is renewing and reinventing itself, as speech-language pathologists, it’s a good time to re-evaluate and reconsider how we are serving our clients.
Here’s a question for you: What is a Type I or Type II error in a clinical setting?
Type I or Type II errors? Do those terms ring a faint bell from your undergrad or grad school days? Maybe you’re remembering it has something to do with research (but that can’t possibly have much to do with your clinical practice, right)?
First, welcome to the world of partial semantic (vocabulary) knowledge. You know you’ve heard those terms before, but maybe you don’t remember much more about them. This is a VERY common phenomenon in vocabulary acquisition and one we’ll come back to in a later post. Partial semantic knowledge affects your performance on standardized vocabulary tests.
Which leads me back to Type I and Type II errors in a clinical setting. They are ‘research’ terms, but they are very much a part of the evaluation process of clients. As a researcher, I have both statistical and experimental design tools I can use to control the likelihood of committing a Type I or Type II error. As an SLP, I have Evidence-Based Practice (the family’s perspective, my clinical expertise, and the external scientific evidence) and the ASHA Code of Ethics.
A Type I error is a ‘false positive.’ A Type I error would be committed when the SLP diagnosis a client with a disorder they don’t have. There is a high risk of committing a Type I error clinically when there is a mismatch between the client’s dialect and the SLPs. Maybe you’ve seen this before. I certainly have. I have seen typically developing children who didn’t speak Mainstream American English placed in speech-language services in school settings when they simply demonstrated a language difference but were incorrectly diagnosed with a disorder – in other words, when the SLP committed a Type I error. Put yourself in the family’s shoes for a moment. Imagine a professional telling you your child had a language impairment. But, it’s not true. The SLP is the one who made the mistake by committing a Type I error.
One could argue that early intervention (EI) is based entirely on a high likelihood of Type I errors. We know from the research that 50% of ‘late talkers’ will ‘catch up’ by school entry. The problem is we’re not very good at figuring out who will and who won’t catch up. So, we commit Type I errors by enrolling very young children in EI ‘to be on the safe side.’
A Type II error is a mistake in the other direction. A Type II error is a ‘false negative.’ A Type II error is committed when the SLP decides a client doesn’t have a communication disorder when, in fact, they do. A Type II error is more of a possibility with very young children or with individuals with milder disorders. A Type II error also could be simply an artifact of a client’s performance on the day of the evaluation. Maybe they had a really good day and we weren’t able to detect the presence of milder impairments.
Type II errors are exactly the types of errors EI is trying NOT to make. If there is any question about whether or not a young child is demonstrating language delays, the child is enrolled in EI. These programs are going out of their way to make sure they are not committing Type II errors by enrolling a percentage of children who likely do not really need the services. Once again, we aren’t really good at predicting which children will or won’t catch up. From a social policy perspective, it is better to commit Type I errors with very young children than to commit Type II errors.
Type I and Type II errors are very real threats which need to be minimized in every SLP’s clinical practice. Here are some steps you can take to inoculate yourself from committing Type I and Type II errors in your evaluation of clients:
To decrease the likelihood of committing a Type I error:
- Ask questions about the family’s cultural and linguistic background. What ‘type’ of English do they think they speak?
- Know your standardized tests. Read the examiner’s manuals. Many popular standardized language tests have scoring conventions for African American English and/or Spanish-influenced English. But, you must read the examiner’s manual to find them!
- Know your standardized tests. Some tests should not be used with people who identify as using non-mainstream dialects of English. Read your examiner’s manual!
- Be prepared to use measures like a language sample or classroom observation to document functional language skills
- The family is the ‘expert’ on their loved on; get their perspective on whether or not they see a problem with the person’s communication skills
To decrease the likelihood of committing a Type II error:
- Talk with the family about their concerns about their loved one. Ask follow up questions. Ask the family to bring in video clips of their concerns.
- Don’t rely solely on the numbers from standardized tests. They don’t tell the whole story. Be prepared to evaluate the person across multiple tasks/contexts
- Know the limits of your standardized tests. Some tests simply don’t go into enough depth to pick up more subtle concerns
- Offer to re-evaluate in 3-6 months if the family still has concerns