How many replicators does it take to achieve reliability? Investigating researcher variability in a crowdsourced replication