How I used Data to Study for the Exam – Building Fluency
Updated: Dec 16, 2022
In my previous blog post “How I used Data to Study for the Exam – What to Study?”, I talked about how I used data to determine what to focus my studies on. In this blog post, I’ll talk about another way I used data when studying, specifically using the say-all-fast-minute-each-day-shuffle technique (SAFMEDS) to help build and fluency. I chose to use SAFMEDS because I had a personal goal of being able to quickly recall the critical features associated with a term.
The term SAFMEDS was coined by Ogden Lindsley in the late 1970s (Calkin, 2003). SAFMEDS were developed as an improvement to flashcards (Quigley, Peterson, Frieder, & Peck, 2018). The process of using SAFMEDS includes a measure of behavioural fluency which is a combination of accuracy and speed of responding (Binder, 1996). More recently behavioural fluency has been described as effortless, accurate, and almost automatic responding (Beverley, Hughes, & Hastings, 2009, p. 236). A full discussion on the evidence supporting the use of SAFMEDS and fluency building is beyond the scope of this blog. In a detailed review of the evidence for procedures and outcomes of SAFMEDS the authors found that the procedure used for SAFMEDS varied across implementations (Quigley et al., 2018). For clarity, I will be using the term SAFMEDS to refer to the daily practice of fluency building and monitoring progress; I did not follow the procedures exactly as described in the literature.
Before I go on, I should mention that building fluency was the last step in my studying chain. Prior to adding a term into my SAFMEDS deck, I had done all of the work needed to “learn” the concept. I considered a concept “learned” when I could: tact the critical features of the concept, list examples, and non-examples, and apply the concept to novel examples. I plan to address my approach to “learning” the concepts in a future blog; for now, I’m highlighting the fact that there’s a big gap between the “What to Study” step and the “Building Fluency” step.
The SAFMEDS Cards
My approach varied over the years. I had originally started using Quizlet (www.quizlet.com) to make my cards. I would print two-sided cards on paper, cut them out and use those during the timings.
Some advantages of Quizlet:
· I could transfer terms to and from an Excel spreadsheet
· I could listen to Quizlet play the cards (e.g., say term then say definition) while I was driving
· The text was always neatly written.
Some disadvantages with using Quizlet:
· It was time-consuming to print and cut out the cards
· I could see pretty easily through the paper so my responding was under faulty stimulus control at times
· Quizlet was limited in what I could put on the cards (e.g., really hard to put pictures of the schedule of reinforcement responding)
· I was not getting any additional learning opportunities (e.g., I was not writing them out myself)
In the end, I chose to use 3x5” cue cards. I wrote the “term” on the front blank side of the card and the “definition” on the back lined side of the cards.
Tool for Organization
The main tool I used to keep track of my SAFMEDS was a spreadsheet. The main column headings of my spreadsheet were:
· Term: The term, concept, category header, or AKA to be studied
o Examples: “conceptually systematic”, “levels of understanding”, “Mentalism AKAs (6)”
· Cooper or Other's Definition: The exact definition from the textbooks
o I wanted to be able to refer to the textbook definition to ensure I was not oversimplifying the definition (e.g., omitting some critical features)
· Cindy's SAFMEDS Definition: The ‘definition’ I created for the term
o Most often, the critical features of the concepts served as my ‘definition’ side of the SAFMEDS card
o I rarely used sentences and kept the definition as brief as possible
o The definition I used sometimes evolved over time based on a new understanding of the concept. When this happened, I created a new card and threw out the old one
In the spreadsheet, I had an additional column titled “Card Made (y/n)” to indicate if I had already made my SAFMEDS card and added it to the card bin.
Although I didn’t start out this way, I ended up having two card bins: one for acquisition and one for maintenance.
In the acquisition bin I had the following categories labeled on the card bin dividers:
· Future: The SAFMEDS cards prepared ahead of time I had not used yet
· Practice: The SAFMEDS cards I was practicing (or intending to) each day
Maximum of 10-12 cards in this section
·For each card, I would practice reading the term, reading the definition then saying the definition out loud until I thought I “knew” it
·At the end of the practice session, I would test myself. I would read the term and say the definition out loud before checking the definition on the back of the card
· If I got the definition right, I moved the card into Fluency
· If I got the definition wrong, I kept the card in Practice
·I added a new card to Practice for each one I moved to Fluency
Fluency: The SAFMEDS cards that had passed the Practice test.
·Add the new cards from the practice session completed moments before
·I would test myself the same way I had in the Practice session, except I timed myself. I usually used a two-minute timing. I would read the term and say the definition before checking the definition on the back of the card
·I placed each card in either a correct or incorrect pile.
·At the end of the timing, I counted the corrects and incorrects and recorded the data on the standard celeration chart (SCC)(stay tuned for my upcoming blog on how to use the SCC!)
· If I got the definition right, I kept the card in Fluency
· If I got the definition wrong, I moved the card back to Practice.
· About once per week I moved the cards I was fluent at (speed plus accuracy) into the maintenance bin
With the maintenance cards, I ran a maintenance check about once per week. The maintenance check was very similar to the Fluency testing (e.g., see term, say definition) except I tested each card in the deck - this resulted in much longer timings (one timing was almost 10 minutes). The reason I did this was to make sure I reviewed each term at least once per week. I do not consider these timings true SAFMEDS because of their lengths; nevertheless, I still collected data and graphed the results on the SCC.
· If I got the definition right, I kept the card in Maintenance.
· If I got the definition wrong, I moved the card back to Practice.
· Much of the literature I’ve read says to start with the whole deck of terms. This was not possible because 1. I didn’t have all the terms when I started, and 2. There are hundreds of terms and there’s no way I would learn them if I tried to study them all at once.
· I added the Maintenance section. I’ve not seen this in the literature, usually, all cards are kept in the full deck. I added the maintenance section because:
1. I disliked reviewing things I already knew
2. I could get more practice with the terms I did not know well
3. It kept my acquisition sessions shorter and hence the response effort lower
4. I liked seeing the maintenance deck grow (positive reinforcement for the win)!
The graphed results of my daily fluency timings are shown in Figure 1. I ran one to three timing trials each session and charted my best time on the SCC. For my aim (not shown on the chart) I used the aim required during my master’s program of 16 per minute. Visual analysis shows that my rate of responding during the daily fluency timings remained at a level of 15-18 definitions per minute. This informed my studying and told me what I was doing resulted in my performance remaining stable and at the desired level – yeah!
Figure 1: Daily Frequency Timings
The graphed results of my weekly maintenance checks are shown in Figure 2. The number on the phase line label indicates the number of cards in the maintenance deck. The number of cards in the deck more than tripled in three weeks! Visual analysis reveals that my rate of responding for the first six sessions is between 15 to 18 responses per minute which is at the desired level. There were four weeks between the second last and the last session. My rate of responding for the last session did drop and the errors went up. This information suggests a lack of retention of information and had I continued, I would likely have either increased the number of practice sessions or limited the duration of my trials in order to increase my rate of responding.