RBT Graphing Practice Questions

The graph's y-axis shows percentage correct, so the raw tally must be converted before it is plotted. Dividing 12 correct responses by 16 trials gives 0.75, and multiplying by 100 gives 75 percent, which is the value to plot. Plotting 12 enters the raw count on a percentage scale, which would make Isaac's performance look much lower than it was and would not be comparable with sessions that had a different number of trials. Plotting 25 reports the percentage of incorrect trials, the opposite of what the axis displays. Plotting 133 comes from dividing trials by correct responses instead of correct responses by trials. Checking the axis label before entering data helps the RBT catch conversion errors.

A cumulative graph adds each session's count to the running total, so the line can never go down. With zero mands on Wednesday, the total stays at 46, and plotting the point level with Tuesday creates a flat horizontal line, which is how a cumulative record shows zero occurrences. Plotting at 0 treats the graph like a regular line graph of session counts and wrongly erases the earlier total. Placing the point below Tuesday's suggests the total decreased, which is impossible on a cumulative record. Leaving Wednesday blank would look like missing data, as though no session occurred, rather than a session in which the behavior did not occur. The flat segment accurately tells the supervisor that zero mands were recorded.

A trend is the overall direction of a data path, and at least three data points within a condition are generally required before one can be identified. With only two sessions graphed, Zoe's data show a change between two values, not yet a trend, so the RBT keeps collecting baseline data. Two points can always be joined by a straight line, so they cannot show whether the change will continue or was a one-time fluctuation. Six and ten are not minimums; more data points make a trend more convincing, and the supervisor may want a longer baseline, but three is the generally accepted minimum. Waiting for that third point before reporting a trend prevents a premature conclusion about elopement.

A zero trend is a flat data path with no meaningful upward or downward direction. In a skill acquisition program, a flat path at about 40 percent across six intervention sessions means Hana is not currently learning to identify coins, and the RBT should make sure the supervisor sees this graph. Steady acquisition would appear as an ascending trend, with percentage correct rising across sessions. Meeting mastery requires performance at a high, predetermined criterion, and 40 percent falls well short of any typical mastery level. Gradually losing the skill would appear as a descending trend, but her data are not moving downward. Recognizing a flat path early helps the supervisor decide whether the teaching procedures need to change.

A multiple baseline graph stacks two or more tiers, one for each setting, behavior, or person, and all tiers share a common horizontal axis of sessions or days. Stacking Eli's home, clinic, and school data this way lets the supervisor see whether greetings change in each setting only when intervention starts there, because the staggered phase changes line up against the same time scale. Plotting all three settings as separate lines on one panel resembles an alternating treatments graph and makes the staggered starts hard to compare. Side-by-side graphs with separate time axes break the shared time scale the design depends on. A bar graph of averages hides the session-by-session changes needed to evaluate a multiple baseline.

This graph tracks Ben's break-card exchanges, the functional replacement behavior taught in place of throwing. An ascending trend on this graph shows that he is using the break card more and more often, which is the desired direction for a replacement behavior. Saying problem behavior is getting worse misreads the graph; a rising path is a concern only when the graphed behavior is the problem behavior, and throwing is not what this graph displays. Deciding that Ben is ready to have the break card faded is a treatment decision for the supervisor, and fading would be premature while the skill is still increasing. Describing his use as unchanged ignores the steady rise. The RBT should always check which behavior a graph displays before describing its trend.

The equal-interval line graph is the most common graph format in applied behavior analysis. Equal distances on each axis represent equal amounts, sessions run along the horizontal axis, and connected data points make level, trend, and variability easy to see, so it is the format an RBT will update most often. Bar graphs are useful for summarizing or comparing totals, but they hide session-by-session changes and are used less often. A cumulative record shows a running total of responses and is valuable for viewing rate, but it is not the standard format for most programs. A semilogarithmic celeration chart displays proportional change and is used mainly in precision teaching, so most RBTs update it far less often than a line graph.

Two features of a trend are its direction and its slope. Slope is the pitch or steepness of the data path relative to the horizontal x-axis, so Theo's sharply rising path has a steeper slope than Nia's gradually rising one. Direction is the same on both graphs, since both paths are ascending, so it cannot be the property that differs. Variability describes how much data points bounce around the trend, and both paths are described as smooth. Immediacy refers to how quickly behavior changes after a phase change, which compares two phases rather than describing a single trend. Noting both direction and slope gives the supervisor a clearer picture of how quickly each client's behavior is changing.

When observation lengths differ, counts cannot be compared directly, so the RBT converts each day's count to a rate before graphing. Monday's rate is 10 divided by 20, or 0.5 per minute, and Tuesday's is 18 divided by 60, or 0.3 per minute, so stereotypy actually occurred more often on Monday even though Tuesday's count was higher. Plotting total counts would wrongly suggest the behavior increased on Tuesday. Writing session lengths beside count data still leaves the graphed values misleading, because the data path itself shows counts. Plotting only Tuesday's count throws away valid data; converting to rate is what makes sessions of different lengths comparable. Graphing rate gives the supervisor an accurate picture of Jayden's behavior.

In visual analysis, trend direction is categorized as ascending, descending, or zero. Chloe's requests rise from 3 to 10 across six sessions, and despite the session-to-session bounce, the overall direction of the data path is upward, so the trend is ascending. Descending would describe values that decrease overall, which is not the case here. Zero would describe a flat data path with no overall change, but her values clearly increase from the first session to the last. Variable describes variability, meaning how much the data bounce around, which is a separate feature of graphed data rather than one of the trend direction categories. The RBT can accurately report both: an ascending trend with some variability.