While it's not an absolute answer, I have to say I would unsurprised to see some variation in expression between two identical transgenic constructs inserted in differing positions even within the same gene locus. Again these variant expression levels are/will be almost certainly due to the inserts position in relation to the regulatory units associated with the gene (or potentially if your unlucky) neighbouring gene(s) and with respect to the surrounding gene 'architecture'. I have previously performed a P-element replacement of an extant insert with a newer construct and seen overt expression differences between two identical P-elements, inserted at the identical loci but in opposing orientations. Again variant promoter fusion constructs containing different restricted parts (i.e. differing upstream sequence, core sequence, downstream intronic sequences) of a gene of interest have demonstrated differing restricted (some over-lapping, some not) expression patterns.
While the PhiC3 system is meant to abrogate these insertional effects by reiterating a transgenic insertion to a defined integrant site, even here with transgenic lines using the identical insertional site in the same genetic background and the exact same construct I still have seen variation in expression levels/patterns (i.e. one line being homozygous inviable or manifesting an overt phenotype while another appears WT or is viable), which I can only ascribe to minor alterations at the loci or impinging on the construct itself that occur during integration into the genome.
So the simple answer is, if you see differences that aren't alleviated by isogenising the lines then yes this is almost certainly an insertional effect.