The original measurement of \(G\) was performed by Henry Cavendish in 1798 using a torsion balance. Cavendish's measurement was very accurate for its time, and it remained the most accurate measurement of \(G\) for over a century. In the 19th and 20th centuries, several other experiments were performed to measure \(G\), and these experiments gradually improved the accuracy of the measured value.
Today, the most precise measurements of \(G\) are performed using torsion balances and atom interferometers. These experiments have achieved uncertainties of less than 1 part in 100,000, which is an incredibly high level of precision.
Therefore, it is not true that measurements of \(G\) have varied much, rather, they have converged to an increasingly accurate value over time.