A Decade Review of Two Potential Analysis Altering Variables in Graph Construction

Area of Science:

• Behavioral Science
• Research Methodology
• Autism Research

Background:

• Single-case designs (SCDs) use time-series graphs to establish functional relations.
• Graph scaling, including vertical axis scaling and data points per x- to y-axis ratio (DPPXYR), may influence visual analysis.
• Previous research suggests inconsistencies in graph scaling practices.

Purpose of the Study:

• To systematically review time-series graph scaling practices in autism research published over the last decade.
• To evaluate adherence to recommendations for vertical axis scaling and DPPXYR in SCDs.
• To identify variations in graph construction impacting visual interpretation.

Main Methods:

• Systematic literature review of 348 articles from four premier autism journals.
• Analysis of 2675 time-series graphs for vertical axis scaling methods and DPPXYR.
• Assessment of graph dimensions (y:x ratio) against current recommendations.

Main Results:

• Significant variation observed in graph scaling (y:x ratio and DPPXYR) across and within SCD types.
• Few studies adhered to recommended axis scaling practices.
• A majority of graphs employed appropriate vertical axis scaling methods, minimizing Type I error inflation.

Conclusions:

• Current practices in time-series graph construction for SCDs in autism research show considerable heterogeneity.
• Adherence to recommended axis ratios and DPPXYR is low, potentially impacting the reliability of visual analysis.
• While vertical axis scaling is often appropriate, further standardization is needed to enhance the precision of functional relation determination in SCDs.