As I continue to work through Campbell's excellent Framing Paul I have been thinking about the order of Acts. Like John Knox (the 20th-century historian of the New Testament era, not the Scottish Reformer), he emphasizes the fact that temporal succession is not necessarily the foremost structural concern in Acts, such that Luke is willing to place events out of temporal sequence in order to serve other, rhetorical and structural, purposes. Whilst we can grant this without difficulty, it nonetheless seems to me that Luke is not disinterested in temporal succession, and that we can suppose that as a general rule he understands that the situation of narrated events in the sequence of Acts is a general index to when they occurred relative to others. Generally speaking, if he narrates A before B that is, I suspect, because he understood that A occurred before B.
This suspicion is based in the data. Acts demonstrably unfolds over time. Its narrative opens in the early 30s, shortly after Jesus's death. In chapter 9 we learn of Paul's flight from Damascus, which on independent grounds we can date to 34 to 37 (Campbell says 36-37, but I think that the evidence he provides in fact allows the flight to have occurred as early as 34). In chapter 12 we learn of Herod Agrippa's death, which on independent grounds we can date to 44. In chapter 18 we learn that Paul is before Gallio in Corinth, and on independent grounds we know that this could only have occurred in 51/52. In chapter 24 we see Paul appearing before first Felix and then Festus in Judea, and we know that these were present in that province between, respectively, c. 52-59 and c. 59-62. Chapter 28 then rounds out the book with a spatio-temporal notice that Paul spent two years in Rome, which joins earlier notices about the length of time that he spent in places such as Corinth and Ephesus, such that on the basis of what we read in Acts 24 we can with a high degree of confidence state that the narrative ends in 62. In fact, as far as I can tell, Luke's probable error of placing Theudas before Judas is the only instance in which he demonstrably places events datable upon independent grounds out of temporal sequence without signposting (these italicized words are a crucial qualifier), and this occurs not in the narrative proper but rather in a dialogue attributed to Gamaliel (cf. Acts 5:34-39). As such, I am not unaware of an independently datable event in the narrative proper that is out of temporal sequence without signposting. Such temporal accuracy seems exceedingly improbable if temporal succession was largely a matter of indifference to Luke.
That said, Luke does narrate events out of strictly temporal sequence. We know that because he gives us signposts to that fact. Perhaps the most obvious is his statement in Acts 11:19-20: "Now those who were scattered because of the persecution that took place over Stephen traveled as far as Phoenicia, Cyprus, and Antioch, and they spoke the word to no one except Jews. 20 But among them were some men of Cyprus and Cyrene who, on coming to Antioch, spoke to the Hellenists also, proclaiming the Lord Jesus." Although the event is narrated almost immediately prior to the narration of the Agrippan persecution and Agrippa's death (cf. Acts 12), which probably occurred in the years 42-44, it explicitly connects back to events narrated in 8:1, and which cannot have occurred any later than 35 and perhaps as early as 31 (respectively, the latest and earliest dates at which Paul was likely converted, and thus the terminus ante quem for the Pauline persecution). We know that here Luke might well be narrating events that took place up to a decade or more prior to the time upon which the narrative is currently focused, but we only know this because Luke provided us with signposting. Such signposting speaks against the idea that Luke was unconcerned with temporal sequence, precisely because he evinces a concern to let us know when that sequence has been disrupted. As such, as best I can tell, with one exception, which lies outside his narrative proper, Luke's temporal sequence either is demonstrably accurate or he provides us with signposts that let us know that the sequence is disordered.
Consequent to the above, I would argue that our treatment of Acts needs to reckon seriously with two realities. One, that temporal succession is indeed an intentional structural feature of Luke's narrative; the order of independently datable events and presence of temporal signposting demonstrate that. Two, that Luke has a fairly decent grasp on the temporal succession of the events that he narrates; again, this is demonstrated by the order of independently datable events. Now, he might well make errors, as with Theudas and Judas, but he seems overall to aim for and achieve a fairly high degree of temporal accuracy. As such, contra Campbell and Knox, it seems not only reasonable to incorporate the data offered by the Acts sequence into our own historical work, but also potentially unreasonable to do otherwise.