Karpinski, R. I., Kolb, A. M. K., Tetreault, N. A., & Borowski, T. B. (2017). High intelligence: A risk factor for psychological and physiological overexcitabilities. Intelligence, (in press). https://doi.org/10.1016/j.intell.2017.09.001
Although this research sounds interesting, here is my conclusion. A better title would be:
"High IQ adults, who both took an IQ test and decided to join MENSA, and who signed up for their research pool and who responded to our email invitation and who completed our questionnaire completely, report on a unvalidated* questionnaire a high prevalence of being diagnosed with some mental/physiological disorders, when compared to the National Survey data for the unselected population"
*Nowhere in the article is the reliability and validity of the questionnaire used explained. It is thus not clear that even if the questionnaire is reliable and valid whether it would provide the same prevalence ratios in an unselected sample of the population as the prevalence ratios you obtained from the national surveys.
Looking at the 'limitations' section, no attempt has been made to critically evaluate if Mensans are representative of the entire HIQ population, i.e. the sample used is far from unselected. There are 323000000 people in the US, of which thus 0.02 x 323000000 = 6460000 qualify for MENSA. Since MENSA America has about 55000 members, this means that 55000/6460000 * 100 = 0.85 % of all eligible become members. Also unaddressed in your paper is the issue whether Mensans who participate in research differ from those who don't. At the very least, a check whether they differ along certain dimensions (age, sex, SES, IQ) could have been performed.
The same check could also have been done for evaluating whether the respondent group different from the invited group, etc....
Furthermore, by not controlling for other potentially influencing factors that might influence the disease prevalence (such as SES for asthma and autism), your findings become even more unreliable. Using prevalence data that can be used to (at the very least) control for SES, and perhaps even sex and age will probably alter your outcomes drastically.
Finally, if your aim was to find a correlation between IQ and prevalence (or susceptibility or something) for certain diseases/ conditions, then why not, using the data you already have, split the respondent group over the median IQ, create matching pairs on SES, age and sex and then see if there is a difference in your dependent variables?
Or alternatively, use IQ as a continuous (well actually discrete, since it is measured in some number :D) independent variable, and see if a linear (or other types of) relation exists?
In its current form, this research cannot support the claims you are making, since the methodology you used is severely flawed. I sincerely hope this paper will be retracted.
Regarding the OverExitabilities issue, I would like to point you to a very well-designed study, who IQ tested a nationwide sample of third-graders with 3 intelligence tests. They thus found a 'gifted' group (+2 SD) and then matched each of them individually to a control child with average IQ (90-110) of the same gender, school, class, SES etc... Then 6 years later, they tested them again, using updated norms (to account for the Flynn effect). The final target group consisted of those who scored +2 SD in the first round of testing and minimally IQ 125 (using an SD of 15) six years later.
This group has been followed throughout their life. The study is known as the Marnburg Giftedness project. Since this is a totally unselected sample of HIQ individuals, a well-matched control group is available and study participants have not been informed about their IQ, this study provides a unique insight into the HIQ population. In:
Wirthwein, L., & Rost, D. H. (2011).Focussing on overexcitabilities: Studies with intellectually gifted and academically talented adults.Personality and Individual Differences, 51(3), 337-342
They found that the target group did not differ from the control group with regard to OE, using the OEQ-II questionnaire.
From the article:
"In four of the five overexcitability scales (emotional, imaginational, psychomotor, sensual) of the Overexcitability Questionnaire-Two (OEQII; Falk et al., 1999), there were only small and no statistically significant differences between gifted and nongifted adults. A statistically significant result only emerged in ''intellectual overexcitability'' in favor of the gifted (d = .42). Our findings are in accordance with the results of other related studies (see the summary of Mendaglio & Tillier, 2006: gifted individuals often display higher scores in ''intellectual overexcitability''). A closer look at the items of the ''intellectual overexcitability'' scale makes plain that the content of the items centers on a combination of self assessment of problem solving ability as well as on the affective facets of person's self-concept of general intellectual ability (e.g., ''I love to solve problems and develop new concepts''; ''I can take difficult concepts and translate them into something more understandable''). As a rule gifted adults frequently notice that they possess a higher intellectual capacity and thus develop a higher self concept of general intellectual ability. Contrary to other findings, the giftedness sample scored slightly lower in ''emotional overexcitability'' (d = .21) as well as in ''psychomotor overexcitability'' (d = .22)."
This will conclude my comments on your research, for a lengthiness of a level unconducive to critical comment might otherwise arise.