@article{KuehlFehringerMuenzer2022,
  author    = {K{\"u}hl, Tim and Fehringer, Benedict C. O. F. and M{\"u}nzer, Stefan},
  title     = {Unifying the ability-as-compensator and ability-as-enhancer hypotheses},
  series = {Educational psychology review},
  volume    = {34},
  journal   = {Educational psychology review},
  number    = {2},
  publisher = {Springer},
  address   = {New York},
  issn      = {1040-726X},
  doi       = {10.1007/s10648-021-09650-5},
  pages     = {1063 -- 1095},
  year      = {2022},
  abstract  = {Spatial abilities have been found to interact with the design of visualizations in educational materials in different forms: (1) spatial abilities enhanced learning with optimized visual design (ability-as-enhancer) or (2) spatial abilities compensated for suboptimal visual design (ability-as-compensator). A brief review of pertinent studies suggests that these two forms are viewed as mutually exclusive. We propose a novel unifying conceptualization. This conceptualization suggests that the ability-as enhancer interaction will be found in the low-medium range of a broad ability continuum whereas the ability-as-compensator interaction will be found in the medium-high range. The largest difference in learning outcomes between visual design variations is expected for medium ability. A corresponding analytical approach is suggested that includes nonlinear quadratic interactions. The unifying conceptualization was confirmed in an experiment with a consistent visual-spatial task. In addition, the conceptualization was investigated with a reanalysis of pooled data from four multimedia learning experiments. Consistent with the conceptualization, quadratic interactions were found, meaning that interactions depended on ability range. The largest difference between visual design variations was obtained for medium ability, as expected. It is concluded that the unifying conceptualization is a useful theoretical and methodological approach to analyze and interpret aptitude-treatment interactions that go beyond linear interactions.},
  language  = {en}
}