I am not aware of a systematic analysis of this phenomenon but it's pretty easy to get a feel for this by perusing the results of the Dutch elections. For example, looking the results of the CDA list for the 2006 election from Wikipedia, we can see that 79.5% of the voters voted for the first person on the list (which is equivalent to voting for the list as a whole on the ballot, IIRC). That leaves 20% who do make use of the open list to express a preference and influence the rankings of the candidates.
At the same time, this did not have any effect on who was effectively elected. Specifically, six candidates got enough personal votes to be elected directly but five of them were already in the top five spots of the list (not necessarily in the same order). Annie Schreijer-Pierik got more preference votes than all but three other candidates, effectively jumping from number 11 to number 4 on the list. But since her party got 41 seats in total, this had no practical effect.
Further down the list, you will see people getting one or two thousand (and in one case even ten thousands) votes but that's insufficient to be elected and does not affect the final result. In fact, this is a desired effect! Dutch political parties occasionally put what they call “lijstduwers” (or list-pushers) in unelectable positions in the list. Those are well-known personalities who are not seeking a mandate but might get some personal votes. Since they are not expected to reach the threshold, those votes only increase the party's total number of seats, to be filled from the top.
Still, in the last election, three candidates were elected based on preference votes, displacing someone who was higher up on the list than they were.