There appears to be some - note that this, of course, is highly controversial - evidence pointing at the direction that intelligence declined in the US and UK during the last century. There is a relatively new review by Michael Lynch (2016) in which he suggests that modern humans are accumulating slightly deleterious mutations with increasing rate due to relaxed purifying selection (a consequence of technology, medicine, ...), resulting in a 'future genetic load'. Let me just sum up some of the major points he makes that are relevant to assess current evolution of human intelligence:
Heritable mutations that lead to a reduction in fitness of 1% take about 100 generations to be eliminated by selection. Many of these mutations have a direct or indirect effect on the mutation rate (mostly increasing) and relaxed purifying selection on them might cause a positive feedback loop that causes even more mutations. Additionally, parental age increases the number of mutations passed on to the offspring. Most mutations will be slightly deleterious (keep in mind that the majority of new mutations is neutral, the majority of the remainder is deleterious and only very few mutations are beneficial. Therefore the really deleterious ones will still be selected against, but the slightly deleterious ones will not be removed by purifying selection). Accordingly, relaxed purifying selection increases the amount of slightly deleterious selection and decreases the efficiency of getting rid of them.
Lynch argues that the brain might be particularly sensitive to mutations as brain function depends on the finely tuned expression of thousands of genes which for him leads to the proposition that the effect of germline mutation rate on psychological disorders could be higher than expected by chance. In support, he cites Iossifov et al. (2015) who have suggested that autism spectrum disorders might be caused by de novo mutations in at least 30% of the cases and also looked at mutations interfering with IQ measurements.
Finally, he cites two papers that are of interest here as they deal with the change of intellectual abilities of humans over time (Crabtree (2013) and Woodley (2015)). The latter is a meta-analytic study based in the US and UK and Lynch already points out that there there are the usual issues with entangling genetic and environmental factors. However, in that study a slow decline of general intelligence of about 0.39 points per decade due to selection and about 0.84 points per decade due to mutational load is suggested. Interestingly, Woodley points out that there are two major hypothesis for this decline: (i) Since the industrial revolution people with lower general intelligence have higher average number of offspring (that might be the 'differential procreation' you have been hinting at). (ii) Mutation accumulation due to relaxed purifying selection against deleterious variants. Woodley concludes that the decline is a combination of both.
Of course, these points are strongly debatable. One of the major conceptual criticisms on Lynch's view is that slightly deleterious mutations (that Lynch claim to be accumulated as a consequence of relaxed purifying selection) are not slightly deleterious but totally neutral if they do not have a fitness effect. In terms of evolution it does not make a difference if an allele is neutral or negative selection is not acting on it because of technological interference. Another objection - more relevant to your question - is that measuring intelligence in general, or the IQ in particular, is a controversial thing to do and not at all easy to standardize which causes great difficulty in obtaining comparable data. Moreover, tracking down underlying genetics, estimating heritability of intelligence (a trait we even have difficulties to properly define and that is known to be strongly influenced by the environment), and consequently putting together a comprehensive assessment of the evolution of intelligence in humans is really tricky, set aside the ethical problems that may arise which we have not even touched yet.
A last note: In your edit you write:
[...] a summary of fecundity and IQ would potentially answer my question but only if it accounted for the current selection process we see going on in modern technophile societies.
This is problematic as to date it is really difficult, if not impossible, to detect recent selection. This is mainly due to the fact that most methods depend on detecting changes in allele frequencies over time and this needs a considerable number of generations to integrate over.
There have been some efforts to develop more accurate methods, but looking at selection in the last 100 years (which you would need to be able to when looking at the technologically modern populations) is not possible so far. The best attempt I am aware of so far is Field et al. (2016, preprint) who looked, based on genome-wide association study data, at recent selection in the last 2000 years. However, they do not mention intelligence.
Update: This paper was a pre-print. It is now published in Science: Field et al. (2016).
