Fig. 2From: Defining and measuring completion and assessment biases with respect to English language and development status: not all MOOCs are equalComparison of the no-effect and the sample log odds ratio distributions by development and controlled for language (full = sample, dotted = no-effect)Back to article page