Shapiro time delay is one of the fundamental tests of general relativity. To date all measurements of time delay have been conducted on astronomical scales. It was asserted in 2010 that gravitational wave detectors on Earth could be used to measure Shapiro delay on a terrestrial scale via massive rotating systems. Building on that work, we consider how measurements of Shapiro delay can be made using next-generation gravitational wave detectors. We perform an analysis for measuring Shapiro delay with the next-generation gravitational wave detector Cosmic Explorer to determine how precisely the effect can be measured. Using a rotating mass unit design, we find that Cosmic Explorer can measure the Shapiro delay signal with an amplitude signal to noise ratio (SNR) up to $\sim37$ in 1 year of integration time. By measuring Shapiro delay with this technique, Cosmic Explorer will allow for terrestrial measurements of $\gamma$ in the paramaterized post-Newtonian formalism of gravity with precision of $<1\%$.

One of the proposed channels of binary black hole mergers involves dynamical interactions of three black holes. In such scenarios, it is possible that all three black holes merge in a so-called hierarchical merger chain, where two of the black holes merge first and then their remnant subsequently merges with the remaining single black hole. Depending on the dynamical environment, it is possible that both mergers will appear within the observable time window. Here we perform a search for such merger pairs in the public available LIGO and Virgo data from the O1/O2 runs. Using a frequentist p-value assignment statistics we do not find any significant merger pair candidates. Assuming no observed candidates in O3/O4, we derive upper limits on merger pairs to be $\sim11-110\ {\rm year^{-1}Gpc^{-3}}$, corresponding to a rate that relative to the total merger rate is $\sim 0.1-1.0$. From this we argue that both a detection and a non-detection within the next few years can be used to put useful constraints on some dynamical progenitor models.