Capstone: Educational Psychology in the Classroom

This is Unit Thirteen, part two, the capstone's last two arcs: testing, and the behavioral perspective, each ending in one classroom. Part one closed on four developmental theories used together in one lesson. Part two gathers the course's arcs on testing and on behavior, again restating entries taught earlier. Carry one idea through this part: a number or a consequence means what it does, not what it seems to mean, so measure before judging. At the end of each arc, one classroom shows how. By the end you will be able to describe a class's scores by center and spread. To choose the right reading of a score report. To answer the four questions every test score raises. To name the operant quadrant for a classroom consequence. And to build a behavior plan that measures, reads the function and fades.

Three entries on describing and relating scores, restating units two and twelve: distributions, correlation, and center and spread.

Distributions and averages.

EDP 1221 Distributions and Averages

Statistics begins with distributional thinking, since the whole distribution often reveals more than its center; a frequency distribution organizes scores, and the mean, median and mode summarize their center. Two classes can share a median of seventy, one bunched and one spread wide, and the mean is pulled toward outliers while the median is not.

Correlation, and its limits.

EDP 1222 Correlation and Its Limits

A correlation coefficient runs from minus one to plus one, its size giving strength and its sign direction; correlation allows prediction, as admissions offices use test scores to predict grades, but it does not show causation, because a confounding variable may drive both. Only an experiment with random assignment can establish cause.

Describing a class's scores.

EDP 1223 Describing a Class's Scores

Describing a class's scores takes a center and a spread. The mean adds the scores and divides by their number; the median is the middle score; the mode is the most frequent; the range is the highest minus the lowest; and the standard deviation gives the typical distance from the mean. Five scores, seventy, seventy-five, eighty, eighty and ninety-five, add to four hundred, so the mean, median and mode are all eighty and the range is twenty-five. Replace the seventy with a twenty, and the mean falls to seventy while the median stays eighty and the range grows to seventy-five. With an outlier, the median describes the typical student better.

Seven entries on reliability, validity and score reports, ending with the first payoff: reading a score report.

Consistency, measured several ways.

EDP 1224 Consistency Measured Several Ways

Reliability is consistency, estimated by test-retest, alternate-form and internal-consistency coefficients, the last from split halves, Kuder-Richardson formula twenty or coefficient alpha, and by interrater agreement across scorers, as when two teachers give the same essays nearly the same scores. More items, clear directions and clear scoring criteria raise it.

Error in every score.

EDP 1225 Error in Every Score

No score is perfectly precise: the standard error of measurement estimates how much a score would vary on retesting, and a relatively large one indicates relatively low reliability, so a reported score is best read as a band around the number. Systematic errors, pushing every score one way, reduce validity, not reliability.

Validity is about use.

EDP 1226 Validity Is About Use

Validity is the adequacy of the interpretations and uses of scores, not a property of the test; evidence comes from content, construct and criterion relationships, predictive when the criterion comes later, so a selection test calls for predictive evidence. A valid measure must be reliable, but a reliable one need not be valid.

Choosing a score interpretation.

EDP 1227 Choosing a Score Interpretation

A score report can be read in several ways, and each answers a different question. A criterion-referenced reading says what a student can do against a standard. A norm-referenced reading ranks the student in a norm group, usually as a percentile rank. A standard score, z, T or stanine, gives the distance from the mean in standard deviations. A grade equivalent estimates the grade level whose typical student would earn that raw score. A fourth grader at the eighty-fourth percentile, with a T of sixty and a grade equivalent of six point oh, who has met the fluency standard, is one standard deviation above the mean and did very well on fourth-grade work. Only the standard tells what the student can do.

Scaled scores.

EDP 1228 Scaled Scores

A scaled score converts a raw score, the number correct, onto a reporting scale that means the same across test forms. On CLEP exams each correct answer earns one point, wrong or skipped answers lose nothing, and a scaled score of fifty is equivalent to a course grade of C, so answering every question can only help.

High stakes and accountability.

EDP 1229 High Stakes and Accountability

High-stakes tests carry serious consequences. NCLB required disaggregated results and adequate yearly progress toward full proficiency by twenty thirteen to twenty fourteen, until ESSA replaced it in twenty fifteen, and growth models judge improvement where status models judge proficiency, so a school far behind may show strong growth. Test anxiety runs higher when stakes are high.

The first payoff: reading a score report.

EDP 1230 Reading a Score Report

Here is the first payoff. Every test score raises four questions the course has answered one at a time. How consistent is it? Reliability is necessary for any use, and the standard error of measurement makes every score a band. What does it compare the student with? A standard, a norm group, the normal curve, or a grade-equivalent estimate that is often misread. What use is it valid for? Content evidence for judging what was learned, construct evidence for an ability, predictive evidence for selection. What else could explain it? Language, anxiety, biased content, and contested effects such as stereotype threat. A parent brings a report showing her son at the forty-fifth percentile, a scaled score slightly below the proficiency cut, and an A in class math. The teacher explains the norm group, that the score sits within its standard error of the cut, and that grades and tests each tell something the other misses, then checks the criterion-referenced results for skills to practice. A single score is never the whole picture.

Five entries on conditioning, restating unit four: classical and operant conditioning, the quadrants, schedules, and extinction.

Classical conditioning.

EDP 1231 Classical Conditioning

In classical conditioning, a neutral stimulus paired with an unconditioned stimulus, which triggers an unconditioned response, becomes a conditioned stimulus that triggers a conditioned response, as Pavlov's bell, paired with food, came to make dogs salivate on its own. Classical conditioning pairs two stimuli before a response; operant conditioning links a behavior to what follows.

Operant conditioning and the four quadrants.

EDP 1232 Operant Conditioning and the Four Quadrants

In operant conditioning, behavior is shaped by its consequences; positive means adding and negative means removing, so reinforcement, which increases behavior, and punishment, which decreases it, each come in positive and negative forms: praise, a stopped seat-belt chime, a reprimand, and a toy taken away. Negative reinforcement is not punishment.

Naming the quadrant.

EDP 1233 Naming the Quadrant

Naming the operant quadrant takes two questions in order. Did the behavior afterward become more or less frequent? More means reinforcement; less means punishment. Was something added or taken away? Added means positive; taken away, negative. A student who calls out is scolded, and calling out increases: something was added and the behavior rose, so the scolding was positive reinforcement, because attention is what the student wanted. A student excused from homework for finishing early finishes early more often: negative reinforcement. A student who loses recess for pushing pushes less: negative punishment. The labels describe what happened, not what was meant.

Schedules of reinforcement.

EDP 1234 Schedules of Reinforcement

Partial schedules are ratio, depending on the number of responses, or interval, depending on time, and each is fixed or variable; ratio schedules produce faster responding, and variable schedules stronger responding, as unscheduled quizzes put studying on a variable-interval schedule. Continuous reinforcement brings fast learning but rapid extinction.

Extinction and intermittent reinforcement.

EDP 1235 Extinction and Intermittent Reinforcement

Extinction is the fading of a behavior once it is no longer reinforced; intermittent reinforcement makes a behavior slower to learn but also slower to extinguish, so clowning that is only sometimes rewarded with laughter is hard to extinguish. Removing one source of reinforcement may not be enough when classmates supply another.

Five entries on behavior in the classroom: rewards, the ABCs, consequences by function, fading, and the final payoff.

Rewards and intrinsic motivation.

EDP 1236 Rewards and Intrinsic Motivation

An expected extrinsic reward for an activity a person already enjoys can reduce intrinsic motivation, as when students paid for interesting puzzles later worked on them less; tangible and expected rewards harm more than intangible and unexpected ones, so a surprise great work differs from money promised in advance. How much rewards harm is contested.

The ABCs in the classroom.

EDP 1237 The ABCs in the Classroom

Applied behavior analysis records the antecedent, the behavior and the consequence, defines behavior so it can be observed, and states the replacement behavior positively, by the dead person's rule, so stay quiet fails and raise your hand and wait passes. Behavior is measured at baseline and afterward, so change can be seen.

Choosing a consequence by function.

EDP 1238 Choosing a Consequence by Function

Choosing a consequence starts with the function of the misbehavior, found by an ABC analysis. If attention maintains it, withdraw the attention, which is extinction, and a time-out from classmates' attention can work. If escape from a task maintains it, removal rewards the behavior, so block the escape, make the task finishable and reinforce each step. One student hums to make classmates laugh, and a brief time-out removes the audience; another tears up worksheets whenever fractions begin, so the teacher breaks fractions into short steps and keeps the student in the room. Punishment remains a last resort, and time-out is never for behavior maintained by escape.

Shaping, prompts and fading.

EDP 1239 Shaping, Prompts and Fading

Shaping reinforces successive approximations of a target behavior; prompts, verbal, gestural and physical, help the right response occur; and both prompts and artificial reinforcers are faded until natural reinforcers maintain the behavior, as a student who once practiced piano for points now practices because playing well is satisfying. Fade neither too slowly nor too fast.

The final payoff: the behavioral arc in one classroom.

EDP 1240 The Behavioral Arc in One Classroom

Here is the final payoff, and it closes the course. Behavioral principles run from the laboratory to the classroom as one connected set of tools. Classical conditioning explains how neutral stimuli come to trigger responses; operant conditioning explains how consequences shape behavior; schedules decide how durable a behavior becomes. A reward is not a reinforcer until it increases the behavior, and expected, tangible rewards for enjoyed activities can reduce intrinsic motivation. Applied behavior analysis brings these together: define the behavior, measure a baseline, find the function, choose the consequence by that function, shape and prompt, and fade. A teacher who wants prompt starts after lunch counts a week's baseline, states the target positively, praises every prompt start and then only now and then, gives a student who delays to escape a shorter first problem instead of sending him out, and fades the praise until finishing the warm-up is its own reward. Measuring, reading the function and fading keep behavioral methods effective and compatible with intrinsic motivation.

Look at what you can now do. You can describe a class's scores by center and spread, and see how an outlier moves the mean but not the median. You can choose the reading of a score report that answers the question being asked. You can answer the four questions every test score raises: how consistent, compared with what, valid for what use, and what else could explain it. You can name the operant quadrant by what happened to the behavior. And you can build a behavior plan that measures, reads the function and fades. That completes the course: educational psychology, the scientific study of learning, put to work in the classroom.