I picked your first example and googled it. The first result is a study demonstrating a positive correlation between divorce likelihood and IQ for people born in the 1940s in the Netherlands, a negative correlation between divorce likelihood and IQ for people born in the 1950s in the Netherlands, and a positive correlation again for people born in the USA in the 60s. The authors conclude that social factors (the difficulty of divorce and the negative social consequences) are important for this correlation.
It doesn't change that there's a correlation, but anybody using divorce rates as an argument for the effectiveness of IQ tests would have to be quite subtle.
Many iq tests actually measure multiple types of intelligences then combine them.