Speech-language pathologists routinely measure behavior before treatment so that changes that occur during treatment can be compared to starting points.  Administering probes prior to the start of treatment is part of establishing baseline.  By establishing a firm baseline and then re-measuring performance over the course of treatment, we get a picture of a client’s performance.  The methodology we use to do this—and the subsequent interpretations of the data for confirming treatment efficacy—comes from single-subject design, also called small-N (small number of subjects) research.  

 

Single-subject design examines an individual’s response over time. Unlike large group research, in which there is a control group whose performance can be compared to that of the experimental group, small-N research does not have a control group per se.  Instead, the individual functions as his/her own control.  This means that the individual’s responses are examined at different points in time, so that the individual is being compared to himself rather than to control subjects.  By measuring a person’s behavior systematically and repeatedly, trends in performance can be identified and interpreted.

 

 

 

Prior to treatment, when probes have not been administered previously, the clinician should administer a probe three times to observe the trend.  Later, during the course of treatment, the clinician needs to administer the probe at intervals (but not three times in a row each time), and the results of each probe can be compared to previous results to get the overall trend.  At the end of treatment, the clinician may wish to administer probes several times in a row to make sure results are stable.

 

Clinicians may find it troublesome to administer three pretreatment probes because such measurement delays the onset of treatment.  This can be costly in both time and money. There are a few options when faced with this problem.  One possibility is for the clinician to take into account all known information to confirm the baseline, even with fewer probes.  Thus, for example, a clinician who found 0% correct production of /s/ in a first word-level probe may examine the spontaneous speech (also showing 0%) and productions on standardized testing (also at 0%) to confirm that the target of /s/ is absent.  This is not as thorough as repeating the word-level probe three times but may be an acceptable compromise if the error type appears quite consistent.  This is an example of a choice you might make clinically that would be different from that made in a research setting.

 

Another possibility is for the clinician to continue the probe of some targets (those whose baselines are less clear) while beginning  treatment on the targets that are more obviously stable.  Finally, it is easy to administer probes in one session and readminister them at the beginning of the following session, thus collecting two data points without a significant delay in treatment.  The key is to remember that the baseline measures should establish the pretreatment level with confidence so that subsequent changes may be interpreted.   Although three baseline probes may seem excessive in the case of an obviously absent target behavior, they will provide valuable information for targets that are produced some of the time.

 

Once treatment has begun, we measure repeatedly over a stretch of time and observe the changes.   Even though changes may not be observed in our early measures, we keep measuring so that we can see when changes do begin and how rapidly they accrue.  We do not want to wait a long time and hope to see a large change just towards the end of treatment; rather, we want to see when changes occur (their timing) and how quickly or slowly they occur (rate of change).  We observe these changes by implementing our probes at regular intervals.  

 

 

This leads to the obvious question, “How often do we probe during the treatment process?”  The answer to this depends on the client, the nature of the disorder, and the clinician’s expectations.  How well does the client handle probes?  How many behaviors do you hope to probe each time?  What is known about the nature of change with this disorder and the treatment method used?   How is this child responding in treatment sessions, and what is likely to be revealed through probes?Remember that probes are designed to give us information that will help us make good decisions about the course of treatment and allow us to document the treatment’s efficacy.  We are not administering probes just for the sake of saying we’ve done it; we want the information so we can adjust and trust our treatment.

 

I find I get good information by implementing basic probes once per month (e.g., the target in unpracticed responses at the same level being practiced in sessions, plus one untreated behavior), with more extended probing every two months (e.g., more linguistic levels, untreated targets, and control behaviors).  For a client who is progressing unusually rapidly in sessions, I may probe more often, since generalization patterns and changes in control behaviors may allow me to shorten the duration of the treatment.  If I do not see progress in session-to-session treatment data (what is being practiced in the session), I may decide to forego probes and, instead, modify the treatment for better results.  Note that even when I am not administering structured probes, I am listening to spontaneous productions to note any changes (even if spontaneous speech is a level more advanced than that being practiced).

 

To summarize:  we establish baseline at the start of treatment by measuring treated and untreated behaviors, and we repeat the measurements multiple times to establish a trend.  Then during the course of treatment we repeat probes at regular intervals.  Behaviors should be stable (not changing systematically) prior to treatment; once treatment is implemented, we hope to see positive changes in relation to the baseline.  Probes will reveal the amount of change (how much changed there is compared to what was seen at the last data point) and the rate of change over time (how long it takes to start seeing change, and how much change occurs at each measurement).  

 

I will conclude this entry by commenting briefly on my experience using probes.  At times in my work I have cut corners (hasn’t everyone?) and not been willing to take the time to establish a clear baseline.  Sometimes my resistance has been because the target seems so obvious, and in other cases the target was so complex it seemed to be too much work to figure out a system for clear measurement!  I can say with authority that I’ve never really gained time by cutting corners on measuring baseline (which also involves setting up probes I will use later on).  I have, however, wished that I had more clear baseline data or that I had taken time to figure out my measures more carefully.  Sometimes what I thought was an obvious baseline turned out to be more complex (for example, with effects of phonetic or pragmatic context); sometimes the response to treatment was confusing and I wished I had more clear evidence of the pre-treatment status.  

 

I used to worry about taking several sessions (after diagnostic evaluation) to document a clear baseline or taking time later on for probes.  I worried that families might object or would think they were not getting their money’s worth.  I no longer have this worry.  I find that obtaining clear data to support my treatment decisions increases my confidence that I am providing good service and improves communication with families because I can show them good data.  

 

Happily, just as our clients benefit from practice, so do we.  By administering probes regularly, I’ve become more skilled and efficient, as have the many students I’ve supervised over the years. The better we get, the less time the probes take. In addition, I believe that taking data helps train our eyes and ears by forcing us to make judgments about responses.  Although I don’t have data to prove it (!), I like to believe that disciplining ourselves to take good data not only serves the client’s best interests but also helps keep us “tuned up” as speech-language pathologists, alert to details that improve our understanding of our clients.