The bias becomes even clearer when we quantify inaccuracies. I score the
seriousness of each editing error by assigning it a value on a scale from fully
accurate (at 0) to wildly inaccurate (at 6), in the manner of the semantic
differential (Osgood et al., 1957). A news story fully accurately edited
scores 0. A story with one very serious editing inaccuracy, or several smaller
errors, scores 5-6, with two very serious inaccuracies about 10-12. 13 We can
then examine the data on parameters such as originating source, topic
classifications, country of origin, or the news factors mentioned above. I
present data on only one dimension here: country of origin.