Campbell makes a very interesting observation. He observes that when Paul (whom he judges to have written Colossians) wrote Colossians he seems to have supposed that the letter-carrier would reach that city before he or she reached nearby Laodicea. The reasoning is that in 4:16 Paul says that this letter is to be also read in the church at Laodicea. Campbell's argument is it would normally be the case that the letter would have arrived in Colossae, and then the Colossians would be expected to produce a copy for Laodicea. Campbell then argues that Paul must be writing from somewhere east of Colossae, and looking westwards it, as only in such a situation would the letter-carrier naturally be expected to reach Colossians first. Judging also (on the basis of information from Philemon, which probably was written and sent also to Colossae at the same time) that Paul must be located not far away, and thus suggests that Paul was writing Apamea, about 3.6 days journey east of Colossae.
I find this discussion interesting because it highlights what to me is a weakness in Campbell's approach. He has paid great attention to the data. He has noticed something that might in fact help greatly in discovering the locale from which Colossians was written. But because he has on programmatic grounds excluded Acts data from the discussion he has to posit that Paul was imprisoned (for we know he was in prison when he wrote these letters) in a locale that we otherwise have no evidence that he even visited, when in fact we have data that suggests he was imprisoned in Caesarea, from which one would just happen to travel west to get to Colossae and Laodicea. If I know that when Paul wrote these letters he was likely imprsioned somewhere to the east of Colossae and Laodicea, and if I have data indicating that Paul was imprisoned in a specific locale to the east of Colossae and Laodicea, why do I need to posit that he was imprisoned somewhere we don't even know that he visited?
Campbell, I think, gets into this difficulty because he converts a priority appropriate to the data into a priority appropriate to judgment. Yes, the Pauline data has a certain priority when making judgments about Paul's life. It does not follow that one should make all one's judgments about Paul's life from the Pauline data before even consulting the Acts data, as Campbell insists, nor does it follow that judgments made regarding the Acts data should not inform our judgments formed on the basis of the Pauline data. That this does not follow is highlighted by the fact that Campbell is quite willing to use data from Josephus to inform his judgments about Paul. If Josephan data can inform such judgments, why not Lukan? If it is licit to use Pauline and Josephan data from the off in the discussion of Pauline chronology, then I would suggest what is necessary is to consider a model of investigation that allows us to work with all the potentially relevant data, not excluding the Lukan.
I would suggest that we think about judgments in terms of Venn diagrams. There are three circles, one representing judgments that we can possibly infer from Pauline data, one representing judgments that we can possibly infer from Acts data, and one representing judgments that we can possibly infer from data that is neither Pauline nor Acts. What we are seeking is the place where these three circles overlap. These circles will overlap, and the reason they will overlap is because we are dealing with judgments, not data. The judgments will overlap, because even if we must judge in a given instance that Luke has made a mistake over and against something reported by Paul, that is still an overlap in possible judgments. And of course that always remains a possibility in principle (as does the possibility that Paul lied, or misreported the matter at hand, etc.).