The findings (of papers I read) suggest that the unique “T+1 trading rule” in China may be a reason that incurs:
the lower opening price in the morning
and the higher closing price in the afternoon,
resulting in the statistically significant differences between the daytime and overnight returns.
So I get descriptive statistics of rd (daytime returns) and ro (overnight returns) (i.e. mean, median, correlation coefficient).
Also, I try to test whether they are statistic significant different from zero through t-test. In order to show that t-test is suitable, I plot figures of rd and ro to get their rough distribution intuitively.
However, we can see that during 16-04-2017 to 01-01-2018, rd and ro are not statistic significant different from zero .