There is no reason to quarrel about the meaning of 'significance' in statistical testing vs. in natural language. Just look at figs. 4 and 5 in the paper, and s

There is no reason to quarrel about the meaning of 'significance' in statistical testing vs. in natural language. Just look at figs. 4 and 5 in the paper, and see that mean times spent for the scanner task were: 5.17 hours (dynamical typing) 7.71 hours (static typing) And that the difference in statistically significant (p=0.04, Mann-Whitney U-test). Whether 5 vs. 8 hours is significant in the natural language sense, everyone can decide for themselves.