? You are given 3 place accuracy for the M of the HCl. You can't get better accuracy than the standards you are working with. If you can obtain the (digital) data, then the graph would be less preferable, but converting points on a graph to a table is unlikely to improve your accuracy (as compared to precision), especially when the data appears to be (suspiciously) well behaved, as in this case. (In other words, if there were more noise in the data, then converting to a table then doing least squares regression would likely help). I assume you understand the difference between precision and accuracy (as well as the difference between efficiency and effectiveness).