The graph in the OP is from one study; reading carefully, I now realise that it's not actually clear that it is one of the 63 studies. Ars Technica brings it in to the discussion thus:
Is there a chance that higher intelligence makes people less religious? Two sets of large scale studies tried to answer this question.
The first are based on the Terman cohort of the gifted, started in 1921 by Lewis Terman, a psychologist at Stanford University.
...
The second set of studies is based on students of New York’s Hunter College Elementary School for the intellectually gifted.
...
Other studies on the topic have been ambiguous. A 2009 study, led by Richard Lynn of the University of Ulster, compared religious beliefs and average national IQs of 137 countries.
That last study is the one the graph is from. Ars Technica does not actually specify that it is one of the ones used in the meta-analysis.
The studies used for the meta-analysis are not just using IQ tests:
In the various studies being examined, analytic intelligence has been measured in many different ways, including GPA (grade point average), UEE (university entrance exams), Mensa membership, and Intelligence Quotient (IQ) tests among others.
As for the Yahoo headline in the other thread that you object to so much, I'll remind you of the opening of the paper's abstract: "a meta-analysis of 63 studies showed a significant negative association between intelligence and religiosity."
= new reply since forum marked as read