%PDF-1.4 % 4 0 obj > endobj 7 0 obj (ABSTRACT) endobj 8 0 obj > endobj 11 0 obj (INTRODUCTION) endobj 12 0 obj > endobj 15 0 obj (GRAPHICAL PERCEPTION) endobj 16 0 obj > endobj 19 0 obj (WEB-BASED EXPERIMENTS AND MECHANICAL TURK) endobj 20 0 obj > endobj 23 0 obj (RESEARCH GOALS) endobj 24 0 obj > endobj 27 0 obj (EXPERIMENT 1A: PROPORTIONAL JUDGMENT) endobj 28 0 obj > endobj 31 0 obj (Method) endobj 32 0 obj > endobj 35 0 obj (Results) endobj 36 0 obj > endobj 39 0 obj (EXPERIMENT 1B: RECTANGULAR AREA JUDGMENTS) endobj 40 0 obj > endobj 43 0 obj (Method) endobj 44 0 obj > endobj 47 0 obj (Results) endobj 48 0 obj > endobj 51 0 obj (EXPERIMENT 2: GRIDLINE ALPHA CONTRAST) endobj 52 0 obj > endobj 55 0 obj (Method) endobj 56 0 obj > endobj 59 0 obj (Results) endobj 60 0 obj > endobj 63 0 obj (EXPERIMENT 3: CHART SIZE AND GRIDLINE SPACING) endobj 64 0 obj > endobj 67 0 obj (Method) endobj 68 0 obj > endobj 71 0 obj (Results) endobj 72 0 obj > endobj 75 0 obj (MECHANICAL TURK: PERFORMANCE AND COST) endobj 76 0 obj > endobj 79 0 obj (Turkers Overlap Across Studies) endobj 80 0 obj > endobj 83 0 obj (HIT Completion Rates: ``Samplers'' and ``Streakers'') endobj 84 0 obj > endobj 87 0 obj (With Qualification, Turkers Provide High-Quality Results) endobj 88 0 obj > endobj 91 0 obj (Standard HITs Frustrate Fine-Grained Timing) endobj 92 0 obj > endobj 95 0 obj (Reward Level Affects Study Completion Time) endobj 96 0 obj > endobj 99 0 obj (Crowdsourcing Reduces Money and Time Costs) endobj 100 0 obj > endobj 103 0 obj (FINDINGS AND FUTURE WORK) endobj 104 0 obj > endobj 107 0 obj (REFERENCES) endobj 108 0 obj > endobj 125 0 obj > stream xڵ[Y6~ϯ jD>L8ICIHx̎ ۵U"ht yW+'0S7~w$nWy_o~ssk?vC;Qz/.?_F6_]o,ws=P4ﯲčjn׹L_hl{;j8o=LE?8U7'-}\H kӫMin_s^+>fˉ70_2D]/afl_[wh|WjՔ3ߪZ5#A