When the frequency distribution is normal

A normal (Gaussian) frequency distribution is symmetric and bell-shaped, so values pile up around a single central point. In such a distribution the mean, median and mode coincide at that center because the left and right tails mirror each other; if these three measures were unequal the distribution would be skewed rather than symmetric [1]. Therefore, when the frequency distribution is normal, the mean, median and mode are identical, making option 2 the correct choice. This equality is a defining property of the ideal normal curve and underlies many statistical procedures that assume normality [1].

Sources

1. [1] https://soc.utah.edu/sociology3112/normal-distribution.php