The effect of data points per x- to y-axis ratio on visual analysts evaluation of single-case graphs.

Abstract

Research based on single-case designs (SCD) are frequently utilized in educational settings to evaluate the effect of an intervention on student behavior. Visual analysis is the primary method of evaluation of SCD, despite research noting concerns regarding reliability of the procedure. Recent research suggests that characteristics of the graphic display may contribute to poor reliability and overestimation of intervention effects. This study investigated the effect of increasing or decreasing the data points per x- to y-axis ratio (DPPXYR) on rater evaluations of functional relation and effect size in SCD data sets. Twenty-nine individuals (58.6% male) with experience in SCD were asked to evaluate 40 multiple baseline data sets. Two data sets reporting null, small, moderate, and large intervention effects (8 total) were modified by manipulating the ratio of the x- to y-axis (5 variations), resulting in 40 total graphs. Results indicate that raters scored effects as larger as the DPPXYR decreased. Additionally, a 2-way within-subjects analysis of variance (ANOVA) revealed a significant main effect of DPPXYR manipulation on effect size rating, F(2.11, 58.98) = 58.05, p < .001, η2 = .675, and an interaction between DPPXYR manipulation and magnitude of effect, F(6.71, 187.78) = 11.45, p < .001, η2 = .29. Overall, results of the study indicate researchers and practitioners should maintain a DPPXYR of .14 or larger in the interest of more conservative effect size judgments. (PsycINFO Database Record

A Decade Review of Single‐Case Graph Construction in the Field of Learning Disabilities

“…First, the DPPXYR recommendations are preliminary and are based on one empirical study; thus, researchers may not see the justification for this metric until additional data support the current guidelines. Second, the initial empirical study (Radley et al., 2018) and subsequent guideline paper (Dart & Radley, 2018) were published in the field of school psychology. Authors may not yet be aware of these recommendations, particularly considering that our parameters for this search focused on four journals in the field of special education.…”

“…It is worth noting that graphical elements that are classified as either aesthetic‐altering or analysis‐altering can be re‐categorized based on additional empirical evidence. Currently, there is preliminary evidence to suggest that two graphical elements are analysis‐altering: (1) the process used to scale the ordinate (Dart & Radley, 2017), and (2) the proportions of the abscissa (i.e., x ‐axis) length to the ordinate length (Radley et al., 2018). Thus, these elements are the focus of this project.…”

“…Radley et al. (2018) expanded on prior recommendations by suggesting that the number of data points that could be presented along the abscissa is a critical element when considering the axis proportions. Thus, the team created the data points per x ‐ to y ‐axis ratio (DPPXYR), (length of x ‐axis/length of y ‐axis) divided by the total possible number of data points to be plotted along the x ‐axis.…”

The effects of visualization on judgment and decision-making: a systematic literature review

“…Several studies investigate the prevalence of biases by searching for distinct patterns of deviations in judgment and decision accuracy with largely mixed results. Through a total of seven cognitive psychology experiments, Sun et al (2010Sun et al ( , 2016 and Radley et al (2018) find that varying scale proportions in graphs change the resulting decision-making since data points are evaluated in a cognitively biased manner based on their distance to other chart elements. Furthermore, Padilla et al (2015) demonstrate that uncertainty is understood to a disparate extent when it is encoded through spatial glyphs, color, or brightness.…”

Three Alternatives for Graphing Behavioral Data: A Comparison of Usability and Acceptability

“…To address the aforementioned limitations with using equal-interval graphs, some researchers have developed systematic protocols to aid evaluation of graphs (e.g., Wolfe et al, 2019), others have developed additional statistics or graphic aides (e.g., trend lines) to assist visual interpretation of data (e.g., Fisher et al, 2003;Lane & Gast, 2014;Lane & Sándor, 2009;Manolov & Vannest, 2019), and some suggest more intensive training (e.g., Blair et al, 2019;Nelson et al, 2017;O'Grady et al, 2018;Young & Daly, 2016). Additionally, many researchers have suggested that data should be graphed using basic principles of standardization in the construction of visual displays (e.g., Dart & Radley, 2017;Datchuk & Kubina, 2011;Kubina et al, 2017;Lindsley, 1992;Pennypacker et al, 2003;Radley et al, 2018). Put simply, standards applied to a graphic display allow less room for distortion and misinterpretation due to faulty graph construction (Cleveland, 1994, Dart & Radley, 2017Huff, 1954;Kubina et al, 2017;Tufte, 2001).…”