From an analysis of the single study they all cite:

Now it is clear that the very oldest respondents have a higher rate of left-handedness than those who are somewhat younger, an effect which is significant (McManus et al ., in press a), and is utterly at odds with explanations due either to differential mortality or greater social pressure to be right-handed.

Or in other words, social pressure does not explain this trend. Either something directly caused it, or much more likely, the study was flawed. Perhaps even...made up.