Humans Prefer to See and Imagine Drawing Curved Objects

Abstract

Lines contribute to the visual experience of drawings. People show a higher preference for curved than sharp angled lines. We studied preference for curvature using drawings of commonly-used objects drawn by design students. We also investigated the relationship of that preference with drawing preference. Experiments 1 and 2 revealed preference for the curved drawings in the laboratory and web-based contexts, respectively. Experiment 3 showed that the curved drawings were also preferred to draw than the sharp-angled ones. However, this effect only appeared when the drawings were made by hand, but not when they were made by computer. We found a moderate positive correlation between liking and drawing preference. This relationship was mainly explained by the hand-made drawings. Sex, art experience and openness to experience did not influence preference for curvature. Altogether, our findings support the curvature effect and the hypothesis that people prefer to draw what they like to see.

How universal is preference for visual curvature? A systematic review and meta‐analysis

“…For example, researchers have long suspected that relevant domain expertise might impact one's sensitivity to and preference for curvature. However, whereas some studies have evaluated the influence of expertise by recruiting expert and nonexpert participants, 13,34,37,65,79,128 others have assessed expertise as a continuous variable/trait via self‐reported questionnaires 14,15,32,36,50,51,81,90,129 . Furthermore, when recruiting experts, some studies have recruited true experts (e.g., Ref.…”

“…These measures provide independent mean values for each one of the groups compared (e.g., curvilinear vs. angular). In addition, a smaller set of studies investigated the effect using dependent dichotomous measures 21,30,32,38,[80][81][82][83][84][85][86] . These measures provide complementary preference values for each one of the groups compared.…”

“…In addition, a smaller set of studies investigated the effect using dependent dichotomous measures. 21 , 30 , 32 , 40 , 47 , 48 , 49 , 50 , 51 , 52 , 53 These measures provide complementary preference values for each one of the groups compared. That is, in these studies, a curvilinear stimulus and an angular stimulus are presented simultaneously, and participants choose one of the two stimuli.…”

“…We focused only on studies employing continuous measures (i.e., rating scales or two-alternative responses), and were not able to compute the effect sizes from all the studies meeting eligibility criteria because of insufficient data for their calculation. When we computed the average effect size of the smaller set of studies employing dependent dichotomous measures 21,30,32,38,[80][81][82][83][84][85][86] , results indicated an Odds Ratio of 2.13 -consistent with a small effect size 87 . Nevertheless, this meta-analysis comprises a relevant set of studies that investigated the effect of visual curvature under various conditions, providing an estimate of the true effect of preference for visual curvature.…”

“…For example, researchers have long suspected that relevant domain expertise might impact one's sensitivity to and preference for curvature. However, whereas some studies have evaluated the influence of expertise by recruiting expert and non-expert participants 13,34,36,47,89,90 , others have assessed expertise as a continuous variable/trait via self-reported questionnaires 14,15,33,35,49,81,82,91,92 . Furthermore, when recruiting experts, some studies have recruited true experts (e.g., 34 ) whereas others have relied on quasiexperts (e.g., 36 )-despite well-established differences between true experts (e.g., professionals with formal training working in a field) and quasi-experts (e.g., apprentices or graduate students in a field) in relative familiarity with a domain.…”