Programmed instruction

1. Introduction

Version followed: Skinner's linear programmed instruction as set out in "Teaching Machines" (Skinner, 1958) and used in Holland & Skinner's The Analysis of Behavior (1961); Crowder's branching ("intrinsic") variant (1958–1962) as the main alternative; and the Personalized System of Instruction (PSI, Keller Plan) as described for the Arizona State PY 112 course in Keller's "Good-bye, teacher…" (1968). Pressey's 1926 self-scoring machine is noted as a forerunner only.

What it is, plainly: this is an instructional technology, not a psychotherapy. It teaches knowledge and verbal or skill repertoires to learners; it has no treatment protocol, no diagnosis-specific procedures, and no clinical trials as a therapy. Its authors wrote for schools, universities, the military and industry [source: Wikipedia, Programmed learning; Molenda 2018]. A psychotherapist uses it to build teaching materials and courses: psychoeducation modules, between-session skills workbooks, parent-training or carer courses, group-programme handouts, and training for staff or trainees. Every clinical application below is marked [synthesis].

Core idea of change: learning is behaviour change produced by reinforcement. If material is broken into steps small enough that the learner almost always answers correctly, each step requires an active response, and each response is confirmed immediately, the learner is reinforced continuously and the repertoire is shaped toward the target behaviour [source: Skinner 1958]. Teaching is the arrangement of these contingencies, and the programme itself, not the learner, is held responsible when errors occur: high error rates mean the programme must be revised [source: Skinner 1958; Molenda 2018]. PSI applies the same logic at the scale of a course: small units, mastery before advancing, self-pacing, and immediate feedback from a human proctor [source: Keller 1968].

2. Framework

There are two related structures: building and running a programme (Skinner/Crowder), and running a PSI course (Keller). Neither has "sessions" in the therapy sense.

A. Programmed instruction: development and use

Phase	Goal	Typical format and size	Criterion to move on
1. Specify terminal behaviour	State in observable terms what the learner will be able to say or do at the end; write a pre-test and an equivalent post-test	Objectives list; pre- and post-tests of equivalent difficulty	Every objective is testable [source: Wikipedia, Programmed learning; Molenda 2018]
2. Analyse and sequence	Break the repertoire into every step needed to reach it; order simple to complex	Outline of steps; decide linear or branching (Skinner's rule: linear if possible, branching if necessary)	No step requires something not yet taught [source: Skinner 1958; octel 2013]
3. Write frames	Write each step as a frame that evokes a correct constructed response, using prompts, then fade them	Frames grouped into sets; answers shown only after responding	Draft complete with review sets [source: Skinner 1958]
4. Developmental testing	Run with a small number of target learners; record every response	A few dozen learners' responses were enough for Skinner to make large improvements	Frames with frequent errors rewritten; a stated standard reached, e.g. the US Air Force "90/90" (90% of learners reach 90% of objectives) [source: Skinner 1958; Molenda 2018]
5. Self-paced use	Learner works through sets alone	Holland & Skinner: 53 sets in 14 parts; estimated under 15 hours, realistically about 20 hours over several weeks	Learner error rate stays low (Holland & Skinner users told to seek help if missing roughly 3 or more in 10) [source: Ledoux 2017]
6. Post-test	Compare with pre-test	Equivalent form	Objectives met; if not, revise programme, not just learner [source: Wikipedia, Programmed learning]

B. PSI (Keller Plan) course

Phase	Goal	Format and numbers (Keller's PY 112)	Criterion to move on
1. Course design	Choose materials, divide into units, write study questions and several equivalent readiness tests per unit	30 units in fixed order, roughly one homework assignment each; about 30 study questions per unit	Tests and guides ready for all units [source: Keller 1968]
2. Staffing	Recruit and brief proctors and assistants	About 1 proctor per 10 students; proctors meet weekly with staff	Proctors know the material and grading rules [source: Keller 1968]
3. Unit cycle (repeats)	Study unit → take readiness test → proctor scores at once and discusses → pass or review and retest	Tests: about 10 fill-in items plus one short essay; offered in class and at extra times	Unit perfection (mastery) before next unit; failure carries no penalty [source: Keller 1968]
4. Lectures/demos	Motivate students who have earned access	Seven in one term; optional; not examined	Open only to students who have reached a given unit [source: Keller 1968]
5. Final exam and grade	Certify the whole course	Units counted 75%, final 25% (later sections: 50% reading units, 20% labs, 30% final); final drawn largely from readiness-test items	Course units completed [source: Keller 1968]

Pace: students set their own pace; Keller noted that an A grade required completing about two units a week. Students who had not passed a readiness test within two weeks were advised to withdraw [source: Keller 1968]. Exact pass thresholds for readiness tests vary by source: Keller required "unit perfection" and reported that more than four errors meant the student was told simply "not ready"; later PSI descriptions use 90% or similar [source: Keller 1968; Wikipedia, Keller Plan; knowledgejump PSI] [uncertain: exact original threshold, as the primary text was read only via a summarising host].

Clinical scale [synthesis]: for a psychoeducation module, think 5–15 frames per concept and one PSI-style "unit" per session or per week, with a brief mastery check at the start of the next session.

3. Guiding principles

Small steps – each step must be one the learner can almost certainly take; if they can't, the step is too big, not the learner too slow [source: Skinner 1958].
Active, constructed response – the learner writes or says an answer at every step; recall, not recognition; thinking an answer does not count [source: Skinner 1958; Ledoux 2017].
Immediate confirmation (knowledge of results) – show the correct answer right after the response, every time [source: Skinner 1958; Wikipedia, Programmed learning].
Low error rate – design for near-errorless learning; motivation should come from success, not from inducing mistakes [source: Skinner 1958].
Self-pacing – the learner, not the class, sets the speed: "different pace, same learning" [source: Keller 1968; Wikipedia, Keller Plan].
Mastery before advancing – no moving on until the current unit is mastered; failure is information, not penalty [source: Keller 1968].
Fixed, logical sequence – each step depends on earlier ones; no skipping or peeking ahead in linear programmes [source: Skinner 1958; Ledoux 2017].
Measurable objectives – specify the terminal behaviour before writing anything [source: Molenda 2018; Wikipedia, Programmed learning].
The programme is accountable – error records are data for revision; test, revise, re-test [source: Skinner 1958; Molenda 2018].
Written word as main channel – the content lives in texts and study guides so learners can re-read at will [source: Keller 1968].
Lectures motivate, not inform – live presentations are rewards and inspiration, not the source of examinable content [source: Keller 1968].
Personal-social reinforcement – a human proctor scores, discusses and praises immediately [source: Keller 1968].
Branching for diagnosis (Crowder variant) – errors are used to route the learner to remedial explanation, and learners who know material can skip it [source: knowledgejump Branching; Molenda 2018].
Clinical stance when borrowing these tools [synthesis] – you are a teacher here, not a tester: keep the learning task separate from the therapeutic relationship, and treat any client error as a cue to re-step your material.

4. Intervention inventory

Techniques A1–A17 are for building and using programmed materials (Skinner, Holland & Skinner, Crowder); B1–B8 are for running a PSI course (Keller). Example language is illustrative and adapted to a clinical teaching context [synthesis] unless quoted.

A1. Specify the terminal behaviour and write pre/post tests

Purpose / when: first step of any programme; defines what "learned" means and gives you a yardstick [source: Wikipedia, Programmed learning; Molenda 2018].
Steps: (1) List what the learner should be able to say or do at the end, each as an observable action (define, identify, give an example, carry out). (2) For each objective write one or more test items. (3) Split items into two equivalent forms: pre-test and post-test. (4) Give the pre-test to locate the starting point ("initial behaviour").
Example (illustrative): objective for an anxiety psychoeducation module: "Given a short description of a situation, the client names the trigger, the body sensation and the avoidance behaviour."
Watch for: vague verbs ("understand", "appreciate"); if you cannot write a test item, the objective is not yet specified.

A2. Analyse and sequence into small steps

Purpose / when: after A1; produces the order of frames [source: Skinner 1958; octel 2013].
Steps: (1) Work backwards from each objective, listing every sub-skill or term the learner must already have. (2) Order from simple to complex so each frame depends only on earlier frames. (3) Group frames into sets (Holland & Skinner use numbered sets within parts, with review sets at the end of some parts) [source: Ledoux 2017]. (4) Decide linear or branching: Skinner's rule, linear if possible, branching if necessary [source: Skinner 1958].
Watch for: steps that pass the developer but fail learners (detected only by A15); jargon introduced before it is taught.

A3. Write a linear frame with a constructed response

Purpose / when: the basic unit of a Skinner-style programme; every step of content [source: Skinner 1958].
Steps: (1) Write one or a few sentences carrying one new element. (2) Leave a blank the learner must fill by writing a word or phrase (not choosing from options), placed so the whole frame must be read to fill it. (3) Put the correct answer where it is hidden until the learner has responded (next page, under a mask, or revealed on screen). (4) Include prompts (A5–A8) so the response is almost certain to be correct.
Example (illustrative): "When we stay away from something that frightens us, the fear drops for a moment. Staying away is called ____." (answer: avoidance).
Watch for: Skinner preferred recall over recognition and warned that multiple-choice wrong options can strengthen wrong forms [source: Skinner 1958]. A blank that can be filled without reading the frame teaches nothing.

A4. Immediate confirmation and repeat-until-correct

Purpose / when: after every response [source: Skinner 1958].
Steps: (1) Learner writes response first. (2) Then uncovers the answer. (3) Correct: move on. (4) Incorrect: Holland & Skinner users cross out, write the correct answer while re-reading the frame, and work out what produced the error [source: Ledoux 2017]. (5) In Skinner's machines, correctly answered frames dropped out and missed frames came round again until answered correctly [source: Skinner 1958]. On paper or screen, collect missed frames and re-present them at the end of the set.
Example (illustrative): "Write your answer before you look. If it's different, cross it out, write the right one, and read the frame again."
Watch for: peeking before responding defeats the method [source: Ledoux 2017].

A5. Formal prompts

Purpose / when: early frames for a new term; the frame supplies part of the response's form [source: Skinner 1958].
Steps: give letters, a rhyme, or the word itself to copy; e.g. ask the learner to copy the word in the first frame, then supply some letters in later frames. Skinner's spelling example opens by asking the pupil to copy the target word after a definition [source: Skinner 1958].
Example (illustrative): "Breathing slowly from the belly is called d_ _ phragmatic breathing."
Watch for: if the prompt alone produces the answer, the learner is copying; it must be faded (A9).

A6. Thematic prompts

Purpose / when: to cue a response by meaning rather than form; bridges from copying to genuine recall [source: Skinner 1958].
Steps: use a recurring descriptive phrase or context that is associated with the term (Skinner's example uses a stock description of a light source to cue the term).
Example (illustrative): "The automatic, quick thought that pops up before the feeling is the ____ thought." (automatic).
Watch for: thematic cues that point to more than one correct word; add context until only one fits.

A7. Synonym prompts (familiar term before technical term)

Purpose / when: introducing technical vocabulary [source: Skinner 1958].
Steps: (1) Use the everyday word the learner already has. (2) Pair it with the technical word. (3) Over later frames ask for the technical word alone.
Example (illustrative): "Doing more of the things you used to enjoy, often called 'getting active', is known in therapy as behavioural ____." (activation).
Watch for: learner continuing to give the familiar word; add another paired frame.

A8. Priming, blocking and conspicuous errors

Purpose / when: fine control within a frame [source: Skinner 1958].
Steps: Priming: use the target word, or a phrase that leads into it, earlier in the same frame so it is easy to evoke. Blocking: add words that rule out a plausible wrong fill-in (Skinner adds a word that makes a synonym redundant). Conspicuous errors: structure the blank so that a wrong answer would show up twice (e.g. the same missing letter in two places).
Watch for: over-prompted frames that produce "correct" answers with no learning; check by asking for the item later unprompted.

A9. Vanishing (fading) of prompts

Purpose / when: to transfer control from the prompt to the learner's own recall; core of every sequence [source: Skinner 1958; octel 2013].
Steps: (1) Start with heavily prompted frames. (2) In each successive frame remove more of the prompt (letters, words, cues). (3) End with a frame demanding the full response with no prompt. Skinner's verse example: show a line with unimportant letters missing; the learner supplies them; repeat for each line; bring back earlier lines with more letters missing; after about 20–24 frames the learner recites all four lines unaided [source: Skinner 1958].
Example (illustrative, coping-statement card): frame 1 shows the whole statement with a few letters missing; frame 4 shows only first letters; frame 6 shows only the first word.
Watch for: a jump in errors when a prompt is removed means the fade step was too large; insert an intermediate frame.

A10. Spacing, reseeding and review sets

Purpose / when: to keep earlier material active and build durable recall [source: Skinner 1958; Ledoux 2017].
Steps: (1) After teaching a term, bring it back several frames later with other items in between. (2) Reuse terms from earlier sets in later sets (Skinner calls it reseeding). (3) Close a part with a review set that calls on everything taught so far.
Watch for: terms that never recur will decay; track which items have appeared how often.

A11. Rotating prompting techniques

Purpose / when: throughout a long programme [source: Skinner 1958].
Steps: vary which kind of prompt (formal, thematic, synonym, priming) controls the response across frames so learners don't learn the programme's habits instead of the content.
Watch for: learners getting answers right by pattern (e.g. the blank is always the last word); change blank position and prompt type.

A12. Worked template: teaching a difficult word or term (Skinner's spelling sequence)

Purpose / when: teaching a single hard term; illustrates A5, A8 and A9 together [source: Skinner 1958].
Steps (paraphrasing the structure of Skinner's six frames for one word): (1) definition plus copy the word; (2)–(3) copy related root words that show its parts; (4)–(5) fill in letters, using repeated letters so errors are conspicuous; (6) spell the full word unprompted in a sentence.
Watch for: any frame in the six with errors; revise that frame.

A13. Learner-use procedure for a programmed text (Holland & Skinner)

Purpose / when: give these instructions to anyone using a printed programme or workbook [source: Ledoux 2017].
Steps: (1) Read the whole frame, not just around the blank. (2) Write the answer on a separate sheet, one per line, headed with the set number. (3) Then check against the answer box. (4) If wrong: cross out, write the correct answer while re-reading; if right by guess, look back at earlier frames to see why. (5) Do not look ahead. (6) Keep answer sheets. (7) Take short breaks each hour. (8) If missing about 3 or more in 10, check that procedures are being followed and get help.
Example (illustrative): "Write it down before you check. The sheet is for you, and for me to see where the workbook needs fixing."
Watch for: answer sheets that are blank or perfect without corrections (may mean peeking).

A14. Response records for the programmer

Purpose / when: during all use; the data source for revision [source: Skinner 1958].
Steps: collect every response (Skinner's machines punched holes or kept written paper strips). Tabulate errors per frame across learners.
Watch for: frames where many learners err in the same way (sequence fault) versus scattered errors (individual difficulty).

A15. Developmental testing and revision

Purpose / when: before releasing any programme and after each revision [source: Skinner 1958; Molenda 2018; Wikipedia, Programmed learning].
Steps: (1) Run the draft with a small group from the target population. (2) Record all responses (A14) and pre/post scores. (3) Rewrite frames with frequent errors; lengthen sequences where errors cluster; split steps; add or slow fading. (4) Re-test. Skinner reports that one or two revisions based on a few dozen learners' responses produced large improvement. (5) Optionally set a release standard such as 90/90.
Example (illustrative), to pilot clients or colleagues: "You're testing the workbook, not yourself. Every mistake shows me a page I need to rewrite."
Watch for: revising only from your own reading; the method's point is empirical revision from learner data.

A16. Branching ("intrinsic") frame, Crowder variant

Purpose / when: when learners differ widely in prior knowledge or when common misconceptions are predictable; allows skipping and remediation [source: knowledgejump Branching; Molenda 2018; Wikipedia, Programmed learning].
Steps: (1) Present a page or paragraph of content. (2) Ask one multiple-choice question whose wrong options represent likely misunderstandings. (3) Each option sends the learner to a different page or screen. (4) Correct option: confirm and continue. (5) Wrong option: a page that explains the specific error and gives remedial teaching, then returns the learner to the original question to choose again. Crowder published this in TutorText books and used it on his AutoTutor machine.
Example (illustrative): content on anxiety and avoidance, then: "Avoiding the shop makes the fear go down in the long term: true / false." Choosing "true" leads to a page explaining short-term relief versus long-term maintenance, then back to the question.
Watch for: Skinner's objection that wrong options can strengthen wrong responses [source: Skinner 1958]; keep distractors as real misconceptions, and make remedial pages specific to each one. Scrambled page numbering in TutorText books is commonly described but is not established [uncertain].

A17. Self-scoring test-and-teach drill (Pressey forerunner)

Purpose / when: drill and self-testing of already-taught material [source: Wikipedia, Programmed learning]; Pressey's 1926 papers are not drawn on [uncertain].
Steps (as described secondarily): learner answers multiple-choice items; the device scores immediately and the learner cannot advance until the right answer is chosen; items answered correctly drop out. Modern equivalent: flashcard or quiz software with immediate feedback [synthesis].
Watch for: recognition-only learning; pair with constructed-response frames.

B1. PSI course design: units and study guides

Purpose / when: setting up any self-paced mastery course or group programme [source: Keller 1968; knowledgejump PSI].
Steps: (1) Choose the content (Keller: texts and readings; later PSI adopters used films, audio, programmed texts, experiments, interviews). (2) Divide into self-contained units of a size comparable to one homework assignment (Keller: 30 units, fixed order). (3) For each unit write a study guide: brief introduction, the reading, and a list of study questions (about 30 per unit in Keller's course) whose answers are in the reading. (4) Hand out the guide for each unit only when the previous one is passed.
Example (illustrative), group-programme leader: "Each week's booklet has a short reading and questions at the back. When you can answer them, you're ready for the check-in quiz."
Watch for: units too large to master in a sitting; study questions not answerable from the reading.

B2. Readiness tests with equivalent forms

Purpose / when: the mastery check at the end of every unit [source: Keller 1968; Wikipedia, Keller Plan].
Steps: (1) Write several equivalent forms per unit (Wikipedia gives three as typical). (2) Keller's format: about 10 fill-in items plus one short essay question, answered in a booklet. (3) Draw items from the study questions. (4) Offer testing at several times in the week, not only in class.
Watch for: forms of unequal difficulty; students memorising one form (rotate forms).

B3. Immediate proctor scoring and discussion

Purpose / when: every time a test is handed in [source: Keller 1968].
Steps: (1) Proctor scores at once, with the student present. (2) Marks each answer and may accept a student's reasoned defence of an answer. (3) Discusses wrong or weak answers; a student with no errors may still be asked to explain some answers. (4) Records pass/fail. (5) Staff later re-check the booklets. (6) Disputes go to an assistant or the instructor.
Example (illustrative): "This one's right. Tell me in your own words why the fear goes down if you stay in the situation."
Watch for: proctors marking leniently; staff spot-checks catch drift.

B4. Not-ready outcome: review and retest

Purpose / when: whenever the mastery criterion is not met [source: Keller 1968; Wikipedia, Keller Plan].
Steps: (1) Tell the student they are not yet ready; no penalty is recorded. (2) Point to weak areas (Keller reports that with many errors no item-by-item review was given). (3) Send the student to review for at least 30 minutes. (4) Retest on an equivalent form; no limit on attempts.
Example (illustrative): "Not ready yet, and it doesn't count against you. Spend half an hour on sections 2 and 3, then take a different version."
Watch for: repeated failure on the same unit (see section 5).

B5. Proctor selection, duties and meetings

Purpose / when: staffing a PSI course or a peer-led programme [source: Keller 1968; Wikipedia, Keller Plan].
Steps: (1) Select proctors for mastery of the content, maturity, understanding of beginners' difficulties, and willingness to help (Keller used undergraduates who had done the course; later PSI distinguishes internal proctors, advanced classmates, from external ones). (2) Assign about 10 students each. (3) Duties: hand out study materials, score readiness tests, tutor, discuss, record progress, provide social reinforcement. (4) Weekly staff meeting (Keller: about 4 hours proctoring and 2 hours of meetings per week, for course credit).
Watch for: proctors who lecture rather than check; proctors not briefed on scoring rules.

B6. Lectures and demonstrations as motivators

Purpose / when: occasional events that reward progress [source: Keller 1968].
Steps: (1) Schedule a few (Keller: seven a term). (2) Open each only to students who have passed a stated unit. (3) Make attendance optional. (4) Never examine their content.
Example (illustrative), carer course: "Everyone who's finished unit 4 is invited to the Q&A with a former participant."
Watch for: turning them into required content delivery, which breaks the self-pacing.

B7. Grading, final exam and course rules

Purpose / when: whole-course assessment [source: Keller 1968].
Steps: (1) Weight completed units heavily (Keller's described version: units 75%, final 25%). (2) Build the final largely from readiness-test items. (3) Restrict incomplete grades to special cases (Keller reported this regularised study habits). (4) Advise withdrawal for students who pass no unit within two weeks.
Watch for: in clinical use grading usually does not apply; use mastery records only [synthesis].

B8. Study hall, records and staff roles

Purpose / when: running the course day to day [source: Keller 1968].
Steps: (1) Offer a staffed study room (recommended, not required). (2) Classroom assistant keeps progress records for every student and confers daily with the instructor. (3) Instructor selects and organises material, writes tests and the final, gives lectures, settles disputes.
Watch for: progress charts that show stalled students early; act on them (section 5).

Clinical uses of A and B [synthesis]

None of the following is described in the primary sources; they are derived applications: programmed psychoeducation workbooks (A1–A15) for anxiety, depression or medication understanding; fading sequences (A9) for memorising coping statements or safety-plan steps; branching modules (A16) for correcting common myths; PSI-structured skills groups or parent courses with weekly mastery checks (B1–B4) and peer helpers (B5); training trainees in a therapy's concepts via a programmed text before supervised practice.

5. Handling problems

Procrastination and dropout in self-paced courses. The main documented weakness of PSI [source: Wikipedia, Keller Plan; knowledgejump PSI].

Keller's own controls: an explicit pace-to-grade link (about two units a week for an A), advising withdrawal if no unit is passed in two weeks, and restricting incomplete grades [source: Keller 1968].
Deadline contingencies from later research: target dates per unit with bonus points for early completion and penalties for lateness ("conditional" deadlines) reduced procrastination compared with advisory-only deadlines in one web-based course, without lowering satisfaction [source: Majchrzak 2001]; other reports simply award bonus points for timely tests [source: knowledgejump PSI]. Treat these as later additions, not Keller's original design.
Clinical version [synthesis]: agree a target unit per week, review the progress record at the start of each session, and praise completion specifically; if two target dates pass, discuss obstacles rather than adding pressure.

Learner repeatedly fails a readiness test or makes many errors in a programme.

Programme principle: high error rates are a programme fault; split the step, add prompts, slow the fading, add review frames (A2, A5–A9, A15) [source: Skinner 1958].
Learner-level: check the learner is following the procedure (writing before checking, not skipping) and get help at around 3 or more errors in 10 [source: Ledoux 2017]; proctor tutoring on weak areas, mandatory review time, retest on a new form [source: Keller 1968].
Clinical version [synthesis]: check reading level, concentration, and current distress before assuming a content problem; offer the material aloud or in shorter sittings.

Boredom or feeling patronised by small steps. Not addressed directly in primary sources [uncertain]. Branching lets learners who already know the material skip it [source: knowledgejump Branching; Molenda 2018]; a pre-test (A1) can place learners further along [source: Wikipedia, Programmed learning]. [synthesis]: offer a placement quiz and let competent clients jump ahead; name the format's purpose up front.

Peeking and copying answers. Defeats the method [source: Ledoux 2017]. Skinner's machines physically prevented it [source: Skinner 1958]. [synthesis]: on paper, put answers on a separate page or the back; on screen, reveal only after typing; explain why writing first matters rather than policing it.

Proctor problems (lenient scoring, poor tutoring, too few proctors). Keller used weekly proctor meetings, staff re-checking of test booklets, and a route of appeal to assistants or the instructor [source: Keller 1968]. Workload is a known reason PSI declined [source: Wikipedia, Keller Plan]. [synthesis]: in clinic groups, use co-facilitators or graduates of the group as proctors with a one-page scoring key and a short debrief after each session; computerised quizzes can take over routine scoring (Wikipedia notes technology may ease testing and administration).

Over-reliance on text (literacy, language, visual or cognitive impairment). Keller's model leans on the written word [source: Keller 1968]. [synthesis]: substitute audio, pictures, or spoken frames; keep the same structure of small step, response, confirmation.

Anxiety about being tested; overwhelm. Not addressed in the sources [uncertain]. Keller's no-penalty retest rule is the built-in protection [source: Keller 1968]. [synthesis]: call checks "practice" or "check-ins"; keep units short; stop when distress rises and return to therapy work; never make access to therapy contingent on passing a module.

Non-adherence to between-session workbooks. [synthesis]: reduce unit size; agree when and where the client will work (a "study hall" equivalent such as 10 minutes in the waiting room, after Keller's study-hall idea, B8); begin each session with the brief mastery check so completion has an immediate consequence (proctor-style discussion and praise).

Ruptures (client feels like a pupil, power imbalance, "homework shame"). Not in sources [synthesis]: acknowledge the feeling, explain that errors indicate where your material needs work (the programme is accountable, A15), invite the client to edit or co-write frames, and keep the teaching component clearly separate from relational therapy work.

Distressing content in psychoeducation (trauma, suicide, self-harm). Not in sources [synthesis]: keep such content in session rather than self-paced; programmed formats can teach the steps of an agreed safety plan (e.g. by fading, A9) but do not replace a clinician's current, standard risk assessment and collaborative safety planning, including lethal-means counselling about safe storage and removal.

Modality-specific limits. Later research found that small steps, strict linear order and knowledge of results were not essential for many tasks and that confirmation is not always reinforcing [source: Molenda 2018]; therefore [synthesis] treat Skinner's rules as defaults to test with your own learners, not fixed requirements.

Sources

Sources:

Books and papers cited:

Skinner, B. F. (1958). Teaching machines. Science, 128, 969–977.
Skinner, B. F. (1954). The science of learning and the art of teaching. Harvard Educational Review, 24, 86–97 [uncertain].
Skinner, B. F. (1968). The Technology of Teaching. Appleton-Century-Crofts [uncertain].
Holland, J. G., & Skinner, B. F. (1961). The Analysis of Behavior: A Program for Self-Instruction. McGraw-Hill [used via Ledoux 2017].
Keller, F. S. (1968). "Good-bye, teacher…". Journal of Applied Behavior Analysis, 1, 79–89.
Crowder, N. A. (1960). Automatic tutoring by intrinsic programming. In Lumsdaine & Glaser (Eds.), Teaching Machines and Programmed Learning [uncertain].
Pressey, S. L. (1926). A simple apparatus which gives tests and scores – and teaches. School and Society, 23, 373–376 [uncertain].
Markle, S. M. (1969). Good Frames and Bad: A Grammar of Frame Writing (2nd ed.). Wiley [uncertain].
Molenda, M. (2008). The programmed instruction era. TechTrends, 52(2), 52–58.

Further reading (not drawn on) [uncertain]: Fuller & Winch (2005) PSI review; PubMed Central copy of Keller 1968; Europe PMC.

← Previous · II-026Precision teaching Next · II-028 →Clinical behaviour analysis