Test Content Outline Domain C SAFMEDS Materials

TEST DOMAIN C SAFMEDS (FLASHCARDS)

What components should be included in an operational definition?

Objective: describe observable, measurable behavior.

Clear: use unambiguous wording.

Complete: state what counts and what does not.

Describe the difference between direct and indirect measures

Direct measurement measures the target behavior itself through observation or a relevant performance task. Indirect measurement uses recollections, reports, and other proxies to infer the target behavior.

Define permanent-product measurement and give an example.

Measure behavior after it occurs by examining an environmental effect that lasts long enough to measure

Example: count completed worksheet problems afterward

Each target response must produce the defined product, and that product must result only from the target behavior of the person being measured

Define rate of response and provide an example.

Rate = count ÷ time

Example: 12 hand raises in 30 minutes = 0.4 hand raises per minute

Describe percentage measurement and when it is appropriate versus not appropriate.

A part-to-whole ratio using the same units: part ÷ whole x 100

Appropriate for correct responses out of opportunities, intervals scored, or time engaged.

Example: 8 correct out of 10 = 80%

Percentage alone does not measure rate or fluency; report the denominator. Small denominators can mislead, and a higher percentage need not mean fewer errors.

Define duration and provide an example.

How long a behavior lasts, measured from its beginning to its end.

Example: a student is out of their seat for 2 minutes

Total duration adds the duration of all episodes in the observation period.

Define latency and provide an example.

The elapsed time from a relevant stimulus or instruction to the beginning of the response.

Example: the teacher says “Begin,” and the student starts writing 8 seconds later. Latency = 8 seconds.

Define interresponse time (IRT) and provide an example.

The elapsed time from the end of one response to the beginning of the next response of the same class.

Example: one call-out ends and the next begins 15 seconds later; IRT = 15 seconds.

Unlike latency, the starting reference is the previous response, not an instruction.

Describe the difference between continuous and discontinuous measurement.

Continuous measurement captures each occurrence or full duration during the observation period.

Discontinuous measurement scores intervals or selected moments and may miss occurrences or details.

Example: count every hand raise versus sample engagement at timer signals.

Watching throughout whole-interval recording does not make it continuous measurement.

Describe the difference between partial-interval and whole-interval recording.

Partial interval: YES if behavior occurs at ANY time in the interval.

Whole interval: YES only if it occurs for the ENTIRE interval.

Example: 15 of 45 seconds = partial YES, whole NO; all 45 seconds = YES under either rule.

Choose the procedure before observing; it does not change with the score.

Describe momentary time sampling and provide an example.

Record whether behavior is occurring at a specified instant, usually the end of each interval.

Example: at each one-minute timer signal, record whether the student is engaged at that exact moment—not whether engagement occurred earlier in the minute.

Define trials to criterion.

The number of response opportunities needed to reach a specified performance criterion.

Example: a learner reaches a criterion of five consecutive independent correct responses on trial 20; trials to criterion = 20.

Define cost-benefit analysis.

Compare the expected benefits of behavior change with its costs when choosing a target or procedure. Consider money, staff time, materials, the learner’s time, and other valuable learning opportunities displaced by the program. Cost is not just dollars.

Describe the training-duration measure.

The total instructional time required to reach a predetermined performance criterion.

Example: six 10-minute teaching sessions to criterion = 60 minutes of training.

It complements trials-to-criterion when comparing instructional efficiency; it measures training time, not just the number of trials.

Describe the difference between measurement validity and reliability.

Validity: data are relevant to the target behavior and the reason for measuring it.

Reliability: repeated measurement of the SAME event produces consistent values.

A measure can be consistent without being valid or accurate.

Two independent observers agree on 95% of intervals. What concept does this demonstrate, and why is it important?

Interobserver agreement (IOA): their scores agree in 95% of the compared intervals.

Agreement supports confidence in consistent scoring across observers, but does not by itself establish validity or accuracy; both observers could make the same error.

Give an example of data collection that is reliable but invalid.

An observer repeatedly counts the same 100 words in an essay but uses that count to assess spelling accuracy. The count is consistent, but it does not measure whether the words are spelled correctly.

Define interobserver agreement (IOA).

The degree to which two or more independent observers obtain the same observed values when measuring the same events. Use the same definitions and measurement system, and collect observations independently.

Why is IOA important to measure?

It helps assess observer competence, detect possible observer drift, evaluate the clarity and usability of definitions and measurement procedures, and increase confidence that reported changes reflect behavior rather than observer differences. It does not prove accuracy.

Define total count IOA.

Compare the observers’ total response counts for the same observation period.

Formula: smaller count ÷ larger count × 100.

Example: 18 ÷ 20 × 100 = 90%.

Equal totals do not guarantee agreement about individual responses.

Define mean count-per-interval IOA.

Calculate smaller count ÷ larger count × 100 for each interval, then average those percentages.

Example: interval agreements of 100%, 50%, and 100% yield 83.33%.

This is not a comparison of grand totals.

Define exact count-per-interval IOA.

The percentage of intervals in which observers recorded exactly the same count.

Formula: intervals with identical counts ÷ total intervals × 100.

Example: 3 exact matches in 5 intervals = 60%.

Partial count agreement earns no credit for that interval.

Define trial-by-trial IOA.

Compare observers’ scores for each individual trial or response opportunity.

Formula: trials with agreement ÷ total trials × 100.

Example: agreement on 9 of 10 trials = 90%.

Define total duration IOA.

Compare the total durations recorded by two observers for the same observation period.

Formula: shorter total duration ÷ longer total duration × 100.

Example: 80 seconds ÷ 100 seconds × 100 = 80%.

Define mean duration-per-occurrence IOA.

For each matched occurrence, calculate shorter duration ÷ longer duration × 100. Then average the occurrence percentages.

Example: agreements of 80% and 100% yield 90%.

Unlike total duration IOA, each occurrence is compared separately.

Define interval-by-interval (total interval) IOA.

The percentage of intervals in which observers agree on occurrence OR nonoccurrence.

Formula: agreed intervals ÷ total intervals × 100.

Example: 8 agreements in 10 intervals = 80%.

Define scored-interval IOA.

Compare only intervals in which at least one observer scored occurrence.

Formula: intervals both scored occurrence ÷ intervals at least one scored occurrence × 100.

Exclude no/no intervals. Particularly useful when behavior occurs infrequently.

Define unscored-interval IOA.

Compare only intervals in which at least one observer scored nonoccurrence.

Formula: intervals both scored nonoccurrence ÷ intervals at least one scored nonoccurrence × 100.

Exclude yes/yes intervals. Particularly useful when behavior occurs frequently.

What variables matter when selecting a measurement system?

Match the measure to the target dimension and purpose. Consider observability, countability, response rate and duration, goals and expected change, observation timing and setting, observer availability and skills, equipment, and planned data use. Balance resources against the information needed for timely decisions.

How do equal-interval graphs differ from graphs with a logarithmic scale?

On an equal-interval axis, equal spacing represents equal numerical differences (0, 5, 10, 15).

On a logarithmic axis, equal spacing represents equal ratios (1, 2, 4, 8).

In an ABA semilogarithmic ratio chart, time is linear on the horizontal axis and the vertical axis is logarithmic.

Define bar graphs and describe their potential weakness.

Bar graphs compare values using bar heights or lengths across categories or conditions.

Summarized totals, means, or medians can hide the order of observations, trends, and variability.

Example: equal weekly averages can conceal very different daily patterns.

Define cumulative records and give an example.

A graph of a running total: add each observation period’s count to the previous total.

Example: session counts of 3, 2, and 0 produce cumulative values of 3, 5, and 5.

Between designated resets, the record rises or stays flat; it does not decrease.

Which graph best displays changes over sessions in a single-case design, and why?

An equal-interval line graph is generally the standard choice. It displays each measure in session order, allowing comparison of level, trend, and variability within and between conditions. Condition-change lines show where procedures change.

On a cumulative record, what do a steep slope and a flat line indicate?

With time on the horizontal axis and consistent scales, a steeper upward slope means a higher response rate.

A flat segment means no new responses were recorded during that observed period—not that behavior continued at a steady, nonzero rate.

In graphical analysis, define level.

The value or range on the vertical axis around which the data cluster within a condition. A mean or median can summarize level, but may hide distinct levels or substantial variability.

In graphical analysis, define trend.

The overall direction of a data path—increasing, decreasing, or zero trend—and its degree or steepness. A trend may have little or considerable variability around it.

In graphical analysis, define variability.

How often and how much repeated measures differ within a condition. It is the spread or fluctuation in the data, including bounce around a level or trend.

Think: how much do the values vary?

In graphical analysis, define immediacy of effect.

How soon behavior changes after a condition change. Examine data immediately before and after the phase-change line. Distinguish a change at the transition from one that appears only after several sessions.

In graphical analysis, define overlap.

The extent to which data values in different conditions share the same vertical range.

Example: baseline 5–8 and intervention 2–6 overlap from 5–6.

Less overlap shows greater separation. Interpret it alongside level, trend, variability, and the experimental design.

In graphical analysis, define consistency.

Similarity of level, trend, and variability when the same or similar condition is repeated.

Example: both baseline phases show high responding and both intervention phases show low responding.

This concerns repeated condition patterns, not agreement between observers.

Define procedural integrity (treatment fidelity).

The extent to which procedures are implemented as planned.

Treatment integrity concerns implementation of the intervention; procedural fidelity also applies to baseline and comparison conditions.

It measures implementation, not whether the learner improved.

What does accuracy mean in procedural integrity?

How correctly the implementer carries out the prescribed steps.

Percentage accuracy = correctly implemented step opportunities ÷ total applicable step opportunities × 100.

Example: 9 of 10 applicable opportunities completed correctly = 90%.

Missed required steps count as errors, not as not-applicable.

What does dosage mean in procedural integrity?

The amount of intervention actually delivered compared with the planned amount, such as session frequency, minutes, or teaching opportunities.

Example: five 15-minute sessions were planned, but only three were delivered: 45 of 75 planned minutes.

List ways to measure procedural integrity.

Use live or recorded observation to score operationally defined steps with a checklist.

Examine reliable permanent products, such as time-stamped records or photos, that document implementation. Self-monitoring can supplement verification.

Record correct and incorrect implementation by applicable step or opportunity; learner progress alone is not a fidelity measure.