• If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • You already know Dokkio is an AI-powered assistant to organize & manage your digital files & messages. Very soon, Dokkio will support Outlook as well as One Drive. Check it out today!


SAFMEDS on the Web

Page history last edited by Regina Claypool-Frey 15 years, 5 months ago


Return to Home Page Special Topics Menu SAFMEDS Issues


SAFMEDS on the Web

Guidelines and Considerations for SAFMEDS

by John W. Eshleman, Ed.D.




Dr. Ogden R. Lindsley coined the acronym SAFMEDS, and developed the technology in the 1970's and 1980's. SAFMEDS have been a staple of Precision Teaching ever since. In recent years, Dr. Stephen A. Graf has continued to promote SAFMEDS in a series of annual workshops held at the conventions of the Association for Behavior Analysis.

The acronym SAFMEDS refers to a specific learning procedure, developed by Lindsley, that one does with cards that resemble flashcards. Lindsley's acronym stands for:



Say All Fast a Minute Every Day Shuffled


In some respects SAFMEDS share outward similarities to flashcards. One typically makes SAFMEDS out of paper cards. The cards are put together into a deck, like a deck of playing cards:




One writes text on each side, although the front side can also contain drawings or pictures.

Behaviorally-speaking, one uses SAFMEDS mainly to establish a sequelic "intraverbal" verbal behavior repertoire. This follows B.F. Skinner's terminology in Verbal Behavior (1957). In plain English "intraverbal" refers to a see/say repertoire. One might also describe SAFMEDS as "paired-associate" learning. One "pairs" the answer on the back of a card with the statement or question on the front. All of these ways of referring to SAFMEDS are appropriate.

The basic procedure for doing SAFMEDS sounds simple. With SAFMEDS one sees what is on the front of a card, says what's on the back, then quickly moves on to the next card. You do SAFMEDS brief, timed sessions, often no more than 30 seconds or 1 minute. Sometimes people conduct "sprints" in shorter 15-second timings. Regardless of the timing length, the main objective centers around building both one's speed and accuracy. The overall goal to seek: stable, fluent performance. That is, we want "solid" learning that persists. We want learning that lasts, not learning that evaporates. SAFMEDS can deliver on that goal of long-term retention!

Lindsley realized, when teaching his college classes, that the traditional way of using flashcards just does not work very well -- or not as well as it could. In coining SAFMEDS, Lindsley embedded the suggested procedure within the very acronym. Originally pronounced like "safe-meds," the pronunciation changed to "saff-meds." Each letter in the acronym states an important procedural point:

  • S -- say -- the learner should say the responses out loud. Not "think" the responses silently.
  • A -- all -- one works with the entire deck of cards as a unit. Versus peeling off the first 10 cards only.
  • F -- fast -- in the timed sessions, work through the cards as quickly as possible. Not slowly like flashcards.
  • M -- minute -- the brief, timed sessions are just that, brief! Either a minute, a half-minute, quarter minute, etc.
  • E -- every --
  • D -- day -- do the brief, timed sessions every day, at least once a day.
  • S -- shuffle the cards before doing a timing. Just like you shuffle a deck of Bicycle playing cards.



Creating SAFMEDS


Structural Definition: SAFMEDS consist of a deck of cards. Each card has text printed on both sides. One side is considered the front, the other the back. (If a "tact" relation between a graphic, or photo, of some object and its name, or other property, is to be learned, then the front would consist of a picture and the back the text.) Usually, however, one puts text on both sides.

The front may consist of a short statement with a word, phrase, or other segment of text blanked out. When you blank out a word, you put the missing word on the back of the card (Graf, 2000).

Many people construct SAFMEDS using 3" by 5" index cards. Graf (2000), however, suggests a smaller size more approximate to business cards in size. For some reason I haven't yet determined, I prefer the 3" by 5" cards. Graf (personal communication March 30, 2000) notes that 4 or 5 test subjects were found to work through the smaller, business-card size cards more quickly than with the 3" by 5" cards.

Some Construction Criteria and Considerations. According to Lindsley, Graf - -(2000)- -, and my own experience:

  • keep the text on the back side be relatively brief (e.g., one word, or even an abbreviation), not long.
  • make any blanks consistently the same length (e.g., 5 underscore characters): _____.
  • put blanks toward the end of the text on the front side.
  • highlight salient text on cards that are otherwise too similar in wording.
  • definitional cards that have only one card per definition (e.g., only one blank) probably need as many blanks (and thus as many cards) as critical attributes of the definition, or key words of the definition.
  • avoid extensive wording on both fronts and backs.



Examples and Nonexamples of SAFMEDS


Let's start with a nonexample of a SAFMEDS card. The nonexample contains problems one should avoid when designing SAFMEDS:

SAFMEDS Nonexample:
Front of card: Back of card:

Criticisms: The text on the front side runs too long. The designer should put the blank toward the end of the statement. The front side contains irrelevant text (the name of Fuller's book). The designer wrote the statement on the front in the passive voice. The answer on the back side runs too long. The back side contains two answers. The semicolon on the back side that separates the two answers may prove confusing, as might the ampersand. The second answer contains a redundancy (the back side repeats the word 'frequencies').

Front of Card: Back of Card:

Comments: Note that in the revised cards, we have cut down considerably on the wording. A good rule to follow: The fewer the words the better the SAFMEDS card becomes. Note, also, that the answer on the back is written larger and in a darker color. This makes the answer easier to see. If the other information is important, such as the title of Fuller's book in this example, then you could put that information on a separate card.



Stimulus Control Considerations:


  • Follow some basic Direct Instruction considerations. If examples of a concept are to be used to teach the concept, then nonexamples of the concept must also be used. The examples should include "far out" examples containing the different varying attributes of the concept. The nonexamples should be "close in" nonexamples, each nonexample differing in, or lacking, one and only one of the critical attributes of the concept.
  • Keep the general, overall length of text on the fronts about the same for all cards. That is, do not have a deck with one or two unusually-lengthed wordings. These cards will quickly stand out. Moreover, the learner will quickly respond to the irrelevant stimulus dimension of text length: "That's the card with lots of words...."
  • Avoid exotic, unusual words on the front where possible. If not possible, then reuse the same exotic word again in several cards. This will help to dilute its uniqueness. Re-using an unusual word will thereby avoid a stimulus control problem of "see an unusual word/say response that should go with the whole front." The exception, of course, is where the unusual word is the subject of the definition.
  • Apply the "black-out" ratio to card design. Irrelevant stimuli on the front of the card should be "blacked out" (a method originally developed by Dr. Jim Holland and associates). (See examples above.) Avoid unusual punctuation, marks, placement of text, and so on. If a card gets smudged or dog-eared, replace it with a new card.

Click here for illustrations of some stimulus control problems: SAFMEDS Issues



SAFMEDS Deck Considerations.


A deck of SAFMEDS may be any number of cards. Often, decks have between 50 to 100 cards. Some people make monster decks having 200 or more cards. I've made decks as small as 20 cards. The actual number of cards in a deck depends on the objectives one has written beforehand. It also depends on the set of facts one has. A set of alphabet SAFMEDS would have 26 cards. The number of cards also depends on the set of related objectives that may comprise a "module." Decks that are small (e.g., 20 cards or fewer) may be combined with other small decks if appropriate. Decks that are too large (over 200 cards) probably ought to be split into more manageable sizes.





The procedure for working with SAFMEDS appears straightforward:

(1) Before each session working with a deck of cards, shuffle the cards. Shuffling prevents the "serial learning" effect. "Serial learning" results in irrelevant intraverbal control by one's own previous spoken response. Eventually, with "serial learning" one could dispense with the cards altogether. That would be great if that is one's purpose. Saying the alphabet, in sequence, represents an example of "serial learning." If that is your goal, great. We want people to know certain things in order. On the other hand, simply because one can say something in an ordered series does not mean that the person can respond to stimuli other than one's previous spoken response. In short, just because you can say the alphabet does not imply you can read the letters! Shuffling the cards "randomizes" their sequence.

(2) You may study the cards before a timed session. Please note that if you do, then this study forms part of the actual learning. It could be where the bulk of the learning actually takes place. Dr. Guy Bruce has suggested that if we want to know more about learning, then we should capture and record this study time. Lindsley himself has suggested that the study behavior can be put on the same chart where one puts the SAFMEDS frequencies.

(3) Set timing sessions from 15 second sprints to 30 second sessions or up to a 1 minute maximum. I have done 2-minute sessions, and two minutes in the world of SAFMEDS is a long period of time. It's kind of like the endless "2 minute drills" in NFL games! As an alternative, one could work through all of the cards in a deck and time how long that takes. In this latter case, the Record Floor for the behavior on a Standard Celeration Chart will differ for each timing.

(4) The learner should be the person holding and flipping the cards. This forms a key difference from flashcards. Often, with flashcards you find one person holding and flipping the cards, and the second person -- the learner -- seeing the card fronts and saying answers. SAFMEDS differ from that. With SAFMEDS the learner holds the cards him or herself.

(5) Set a timer and begin. You can use a stopwatch for the timer. At least use a watch with a second-hand. A digital watch should have the seconds displayed.

(6) For each card, repeat the following: See what is on the front of the card; then say what is on the back before turning the card over. Some people advocate not turning the cards over during a timing. They suggest that you do not flip the cards over, but simply slide the card off the top of the deck as you respond to it.

(7) Turn the card over after saying the back. You may check your own progress, or a partner can determine whether you were correct or not.

(8) Correctly answered cards should be put into one pile, errors into another. Sometimes "skip" responses are put into a separate pile from errors, because a "skip" (e.g., saying "don't know") may not be the same as an error.

Corrects and Errors piles.
Corrects: Errors:

(9) Two methods of timing:

(a) Work through all of the cards in the deck, and when a response has been given to the last card ,turn off the timer and note the elapsed time. Pro: You work with each card. Con: Times will be of differing and nonstandard lengths, making charting more difficult.

(b) Work through as many cards as you can in 30 seconds, one minute, or however long the timing session interval has been set. Pro: A standard elapsed time makes charting results easy. Con: You may not get to all of the cards in the deck.

(10) At the end of the timing, count up the numbers of corrects and errors.

(11) Tally the numbers of corrects and errors.

(12) Chart the numbers of corrects and errors.

(13) You may do one timed session per day, or you may wait a few moments and do another timed session right away. Shuffle the cards before the next timed session!

(14) When working with the cards, make sure that you SAY the response out loud. This works better than silently thinking the response.

(15) Work through a deck as fast as you can; do not dwell on any given card. If you don't know it, say "skip," or "go," or "don't know" and move on to the next card. Lindsley has suggested saying "GO!" when you don't know. Remember, skips may be counted separately from errors if you wish. "If you don't know, say 'go'!"



(16) After several timed sessions you may notice certain "problem cards" that seem difficult to learn. You have several options at this point. (a) You may try to study these cards both after and before a timed session. (b) You may reexamine these cards to see if the wording on them could be improved, or whether they have or lack some other stimulus property making them difficult. (c) You may refer back to the objective related to this card and rework the objective, and from the reworked objective make some new cards.

(17) After you have attained 100% mastery and get every card correct, continue working on them to build up your speed. Ideally, there should be a speed goal set.

(18) If you begin a timing and for some reason get all flustered after a few cards, and your "time is gone," you may stop, reshuffle the cards, take a breath, and begin again. There's nothing wrong with starting over! Try not to do this too often, though.



SAFMEDS Performance Criteria


Graf(2000) suggests that with well-constructed SAFMEDS learners ought to be able to reach and exceed 50 per minute correct. The 40-49 per minute range would be considered good "B" level performance. Again, this depends on the characteristics of the cards. Generally speaking, the fewer the words on the card, the more cards that can be done in a unit of time. So, keep that in mind when setting performance aims. Also, for the aims, have an expert work through the cards, and try working through them yourself.



Mini Review of Graf's (2000) Text


Steve Graf's (2000) book and computer template goes into far more detail than conveyed on this page. The text covers

  • the historical precedents of SAFMEDS
  • a comparison between SAFMEDS and flashcards
  • a list of contact people who have experience working with SAFMEDS
  • the construction criteria and characteristics for making SAFMEDS
  • a classification system for SAFMEDS
  • a set of example SAFMEDS for each classification
  • criteria for nonexamples of SAFMEDS, and examples of same
  • step-by-step guidelines for making good SAFMEDS
  • using a computer Word file template for making SAFMEDS
  • a set of general rules and tips for using SAFMEDS
  • how to run a daily classroom SAFMEDS exercise
  • a list of performance standards, and
  • the future for SAFMEDS.

The text has pages set aside for exercises, questions, and comments. While the text can be self-explanatory, Graf created it to accompany a workshop he conducts (see below). Frankly, this leads me to an observation about teaching SAFMEDS: A learner needs a live teacher to model SAFMEDS, to show what a completed deck looks like, to help guide learners in making SAFMEDS, giving them correction and feedback. Graf once had -- and may still have -- a one-minute VHS video of one of his students from the mid-1980's, Ken Etzold, who set a world-class performance in 1984: 97 correct, 0 errors in 57 seconds. The video shows what SAFMEDS are all about. One can see what fast means. Likewise, a live teacher can do the same. I found such demonstration important when I have taught people how to work with SAFMEDS. People otherwise do not understand what "fast" means.

Even when you show and tell people what "fast" means, they still may not understand it. Or they may think, "I can't ever go that fast." They may think you have some unusual ability. No. They have the same ability, but may not know it. I have found that showing the Etzold video was not sufficient. Modeling the behavior was not sufficient. Necessary, yes. But not sufficient. A live teacher can (a) continue to model the behavior multiple times, (b) encourage students to go faster, and (c) give them immediate feedback about their speed. Note that the word 'encourage' contains the word 'courage.' Learners may need some of what Lindsley calls "curricular courage" to accept going faster, and to realize what it really means.

Likewise, a live teacher can show what it means to say responses out loud. Given the background most people probably have with ordinary flashcards, they are prone to "think" answers silently, or to whisper them. No, say the cards out loud! A live teacher can both help demonstrate this and give people feedback about their loudness. If necessary, a teacher can help learners work on their pronunciations of words, too. Steve Graf has found this to be an important, and often-overlooked objective in college teaching -- especially in this day and age of the unreliable and often degraded verbal repertoires students are apt to have.

A live teacher can also inspect the work of the students in the class or workshop. The instructor can give feedback about instructional design flaws and considerations. And point out and reward the right things the student does, too.

Therefore, practice with instruction, with feedback, and with correction fills a major gap that neither texts nor web pages can fill. A live teacher can also help students unlearn the way they learnt to do flashcards. However, as a reference source, and a source that can help as a reminder once one has learned how to develop, produce, and use SAFMEDS, Graf's text comes highly recommended.





Graf, S.A. (1994). How to Develop, Produce and Use SAFMEDS in Education and Training. Poland, Ohio: Stephen A. Graf, Ph.D. publisher. (Available for $25 from Zero Brothers Software, 7779 Lee Run Road Poland, OH 44514; Email: Zerobros@aol.com. Package includes Word for Windows, or Macintosh, SAFMEDS template. Please specify which system.)


Skinner, B.F. (1957). Verbal Behavior. New York: Appleton-Century-Crofts.




Note: These guidelines for designing, developing, and doing SAFMEDS are adapted, in part, from Steve Graf's (2000) textbook How to Develop, Produce and Use SAFMEDS in Education and Training, but are also based on my own 21+ years of experience making, using, teaching about, and working with SAFMEDS.


Historical:Steve Graf's ABA SAFMEDS Workshop (2000)


Presented at the Association for Behavior Analysis Convention. Marriott Wardman Park Hotel, Washington, D.C. Workshop # 31. 5/27/2000 8:00 AM - 11:00 AM. Delaware A. (Note: You may be able to obtain a copy of Graf's workshop materials from him directly. Contact him at Zerobros@aol.com.

How to Develop, Produce and Use SAFMEDS

STEPHEN GRAF, Ph.D. -- (Youngstown State University) --

Description - SAFMEDS, an acronym for "Say All Fast Minute Every Day Shuffled," represents a procedure to promote fluency in verbal behavior developed by Ogden Lindsley 20 years ago. Learners see information on the front of a card and attempt to say a brief response that can be found on the back. Use of the technique provides an easy way to monitor the learning of large numbers of facts and concepts. Use provides enhanced retention of material with better application to new material, as well as focused performance under pressure.

Objectives - Participants should be able to: develop sets of SAFMEDS that conform to the rules provided for content areas in which the participants teach or train; produce SAFMEDS masters from the MSWord templates provided them as part of their materials; reproduce sets of SAFMEDS for students or trainees; and implement use of SAFMEDS in various settings, structuring practice timings to interface smoothly within classroom or training sessions.

Activities - Participants will practice with conversion of different types of information to SAFMEDS; discuss using the templates provided to produce SAFMEDS masters to be printed as decks on card stock; discuss alternatives form making SAFMEDS available to students, clients or trainees; demonstrate how to run timings with small or large groups; and discuss how to incorporate SAFMEDS into larger patterns of learning.

Audience - Participants should either have competent subject matter repertoires and be interested in promoting fluent verbal behavior in specialty areas for students, clients or employees they teach or train; or want to improve their own repertoires in a content area.

Members - $70 Non-Members - $85

Please refer to the Association for Behavior Analysis website for further information regarding how to sign up and participate. A complete year 2000 ABA convention program can be found there as well. Dr. Graf has conducted this same workshop several years running, and may do so again next year. Please check theABA website for further details.


"As I get around to it...." I will add more SAFMEDS references to this page as time goes along. There are several Lindsley references that I need to cite. I don't have my PT References list handy, but will dig it up. These are references to published and presented works. I will be glad to add links to sites that have software that generates SAFMEDS. Send me the link information if you have it! Meanwhile, have you done your SAFMEDS today?


The information on this page is for educational reference and illustrative purposes only.

This page does not attempt to teach how to make or use SAFMEDS.




Return to Home Page Special Topics Menu SAFMEDS Issues


URL: http://standardcelerationcharttopics.pbwiki.com/SAFMEDS+on+the+Web

E-mail: standardcharter@aol.com

Webmaster: John W. Eshleman, Ed.D.

Copyright 2000 by John W. Eshleman, Ed.D. (Just the form the information is presented in on this page)

Revised -- April 5, 2000.

Site mirrored and updated by Regina Claypool-Frey -- October, 2008


Comments (0)

You don't have permission to comment on this page.