# Non-g Factors Predict Educational and Occupational Criteria: More than g

## Abstract

**:**

## 1. Introduction

## 2. g and Non-g Factors: The Primacy of g

_{non-g}< 0.10, [7], pp. 283–285; see also, [9,11]).

## 3. A Foundational Study by Coyle and Pillow [6]: Non-g Residuals Predict College GPA

## 4. Non-g Residuals of the SAT and ACT Predict Specific Abilities and GPAs

_{β}= 0.29) and negatively with verbal ability (M

_{β}= −0.32). In contrast, the verbal residuals of the SAT and ACT correlated positively with verbal ability (M

_{β}= 0.29) and negatively with math ability (M

_{β}= −0.25) (The non-g residuals of the SAT and ACT correlated negligibly with the ASVAB shop and speed abilities, demonstrating discriminant validity).

_{β}| ≈ 0.17) was smaller than the mean absolute effect for the ASVAB abilities (|M

_{β}| ≈ 0.29) (cf. [15]). (The smaller effect could be attributed to the use of GPAs, which are less reliable than standardized test scores.) The results confirm the predictive power of non-g residuals and are inconsistent with the primacy of g hypothesis, which assumes that non-g factors have negligible predictive power. In addition, the results are consistent with investment theories. SAT math residuals presumably reflect investment in math, which boosts STEM GPAs but retards humanities GPAs. In contrast, SAT verbal residuals presumably reflect investment in verbal areas, which yields the opposite pattern of effects.

## 5. Ability Tilt Predicts Diverse Criteria

_{β}| ≈ 0.28). (Math and verbal tilt correlated negligibly with the non-academic shop and speed abilities, demonstrating divergent validity.) In addition, math tilt predicted STEM majors and jobs, whereas verbal tilt predicted humanities majors and jobs (|M

_{β}| ≈ 0.35). The results confirm the predictive power of non-g factors and are inconsistent with the primacy of g hypothesis, which assumes that non-g factors have negligible predictive validity. In addition, the results are consistent with investment theories ([16], pp. 138–146). Ability tilt presumably reflects investment in math or verbal abilities, which boost similar abilities and preferences (e.g., math tilt and STEM) and inhibit competing abilities and preferences (e.g., math tilt and humanities).

#### A Non-g Nexus Involving Non-g Group Factor Residuals

_{β}| = 0.51) [14]. (The shop and speed residuals generally correlated negligibly with all criteria, providing divergent validity.) The results were interpreted in terms of a non-g nexus involving non-g residuals of group factors and diverse criteria. The non-g nexus complements Jensen’s ([7], pp. 544–583) notion of a “g nexus” involving g and diverse criteria. Like the tilt effects, the non-g nexus suggests trade-offs, with investment in a specific ability (reflected by non-g residuals) boosting similar abilities (e.g., math) but inhibiting competing abilities (e.g., verbal).

## 6. Standing on the Shoulders of Giants: Other Research on Non-g Factors

_{r}| = 0.31, range = −0.21 to 0.40) ([29], p. 427). Moreover, the effects were based on a large and representative sample of participants and tests, inspiring confidence in the results.

_{r}= −0.55), which predict math/STEM criteria (e.g., [25,31]). The residual correlations of the VPR verbal and spatial abilities are analogous to the residual correlations of the ASVAB verbal and math abilities. Both sets of correlations are negative, which suggests a tradeoff between competing abilities (e.g., verbal-spatial or verbal-math). The tradeoff is consistent with investment theories, which predict that investment in one ability (e.g., verbal) comes at the expense of investment in competing abilities (e.g., spatial), yielding negative effects.

## 7. Future Directions: There is Nothing More Practical than a Good Theory

_{β}≈ −0.12), indicating that strong non-academic abilities were associated with weak academic abilities. The results suggest a tradeoff in investment in non-academic abilities (shop) and academic abilities (math and verbal), yielding negative effects. Further research is needed to substantiate non-g effects with other non-academic abilities (e.g., technical tilt) and to examine whether the effects vary with ability specialization factors (e.g., life history and ability level). In addition, future research could examine other non-academic traits such as social intelligence and Big Five personality traits. Possible candidates include emotional intelligence, agreeableness, and theory of mind, which may predict economic and social criteria (e.g., wealth, trust, prosocial norms) beyond g [39].

## 8. Conclusions

## Acknowledgments

## Conflicts of Interest

## References

- Coyle, T.R. Relations among general intelligence (g) aptitude tests, and GPA: Linear effects dominate. Intelligence
**2015**, 53, 16–22. [Google Scholar] [CrossRef] - Frey, M.C.; Detterman, D.K. Scholastic assessment or g? The relationship between the scholastic assessment test and general cognitive ability. Psychol. Sci.
**2004**, 15, 373–378. [Google Scholar] [CrossRef] [PubMed] - Koenig, K.A.; Frey, M.C.; Detterman, D.K. ACT and general cognitive ability. Intelligence
**2008**, 36, 153–160. [Google Scholar] [CrossRef] - Coyle, T.R. Predictive validity of non-g residuals of tests: More than g. J. Intell.
**2014**, 2, 21–25. [Google Scholar] [CrossRef] - Reeve, C.L.; Charles, J.E. Survey of opinions on the primacy of g and social consequences of ability testing: A comparison of expert and non-expert views. Intelligence
**2008**, 36, 681–688. [Google Scholar] [CrossRef] - Coyle, T.R.; Pillow, D.R. SAT and ACT predict college GPA after removing g. Intelligence
**2008**, 36, 719–729. [Google Scholar] [CrossRef] - Jensen, A.R. The g Factor: The Science of Mental Ability; Praeger: Westport, CT, USA, 1998. [Google Scholar]
- Roth, B.; Becker, N.; Romeyke, S.; Schäfer, S.; Domnick, F.; Spinath, F.M. Intelligence and school grades: A meta-analysis. Intelligence
**2015**, 53, 118–137. [Google Scholar] [CrossRef] - Schmidt, F.L.; Hunter, J.E. The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychol. Bull.
**1998**, 124, 262–274. [Google Scholar] [CrossRef] - Thorndike, R.L. Intelligence and Information Processing: The Mind and the Computer; Center on Evaluation, Development, and Research: Bloomington, IN, USA, 1984. [Google Scholar]
- Ree, M.J.; Earles, J.A.; Teachout, M.S. Predicting job performance: Not much more than g. J. Appl. Psychol.
**1994**, 79, 518–524. [Google Scholar] [CrossRef] - Peterson, R.A.; Brown, S.P. On the use of beta coefficients in meta-analysis. J. Appl. Psychol.
**2005**, 90, 175–181. [Google Scholar] [CrossRef] [PubMed] - Roth, P.L.; Le, H.; Oh, I.-S.; Van Iddekinge, C.H.; Bobko, P. Using beta coefficients to impute missing correlations in meta-analysis research: Reasons for caution. J. Appl. Psychol.
**2018**, 103, 644–658. [Google Scholar] [CrossRef] [PubMed] - Cohen, J. Statistical Power Analysis for the Behavioral Sciences, 2nd ed.; Lawrence Erlbaum Associates: Hillsdale, NJ, USA, 1988. [Google Scholar]
- Coyle, T.R.; Purcell, J.M.; Snyder, A.C.; Kochunov, P. Non-g residuals of the SAT and ACT predict specific abilities. Intelligence
**2013**, 41, 114–120. [Google Scholar] [CrossRef] - Cattell, R.B. Intelligence: Its Structure, Growth and Action; North-Holland: New York, NY, USA, 1987. [Google Scholar]
- Coyle, T.R.; Snyder, A.C.; Richmond, M.C.; Little, M. SAT non-g residuals predict course specific GPAs: Support for investment theory. Intelligence
**2015**, 51, 57–66. [Google Scholar] [CrossRef] - Lubinski, D. Exceptional cognitive ability: The phenotype. Behav. Genet.
**2009**, 39, 350–358. [Google Scholar] [CrossRef] [PubMed] - Achter, J.A.; Lubinski, D.; Benbow, C.P.; Eftekhari-Sanjani, H. Assessing vocational preferences among gifted adolescents adds incremental validity to abilities: A discriminant analysis of educational outcomes over a 10-year interval. J. Educ. Psychol.
**1999**, 91, 777–786. [Google Scholar] [CrossRef] - Lubinski, D.; Webb, R.M.; Morelock, M.J.; Benbow, C.P. Top 1 in 10,000: A 10-year follow-up of the profoundly gifted. J. Appl. Psychol.
**2001**, 86, 718–729. [Google Scholar] [CrossRef] [PubMed] - Park, G.; Lubinski, D.; Benbow, C.P. Contrasting intellectual patterns predict creativity in the arts and sciences: Tracking intellectually precocious youth over 25 years. Psychol. Sci.
**2007**, 18, 948–952. [Google Scholar] [CrossRef] [PubMed] - Wai, J.; Lubinski, D.; Benbow, C.P. Creativity and occupational accomplishments among intellectually precocious youths: An age 13 to age 33 longitudinal study. J. Educ. Psychol.
**2005**, 97, 484–492. [Google Scholar] [CrossRef] - Deary, I.J.; Egan, V.; Gibson, G.J.; Brand, C.R.; Austin, E.; Kellaghan, T. Intelligence and the differentiation hypothesis. Intelligence
**1996**, 23, 105–132. [Google Scholar] [CrossRef] - Coyle, T.R.; Purcell, J.M.; Snyder, A.C.; Richmond, M.C. Ability tilt on the SAT and ACT predicts specific abilities and college majors. Intelligence
**2014**, 46, 18–24. [Google Scholar] [CrossRef] - Coyle, T.R.; Snyder, A.C.; Richmond, M.C. Sex differences in ability tilt: Support for investment theory. Intelligence
**2015**, 50, 209–220. [Google Scholar] [CrossRef] - Coyle, T.R. Ability tilt for whites and blacks: Support for differentiation and investment theories. Intelligence
**2016**, 56, 28–34. [Google Scholar] [CrossRef] - Coyle, T.R. Non-g residuals of group factors predict ability tilt, college majors, and jobs: A non-g nexus. Intelligence
**2018**, 67, 19–25. [Google Scholar] [CrossRef] - Newton, I. Newton to Hooke. In The Correspondence of Isaac Newton; Turnbull, H.W., Ed.; (Original Work Published in 1676); Cambridge University Press: Cambridge, UK, 1959; Volume 1, p. 416. [Google Scholar]
- Calvin, D.M.; Fernandes, C.; Smith, P.; Visscher, P.M.; Deary, I.J. Sex, intelligence and educational achievement in a national cohort of over 175,000 11-year-old schoolchildren in England. Intelligence
**2010**, 38, 424–432. [Google Scholar] [CrossRef] - Johnson, W.; Bouchard, T.J. Sex differences in mental abilities: g masks the dimensions on which they lie. Intelligence
**2007**, 35, 23–39. [Google Scholar] [CrossRef] - Wai, J.; Lubinski, D.; Benbow, C.P. Spatial ability for STEM domains: Aligning over 50 years of cumulative psychological knowledge solidifies its importance. J. Educ. Psychol.
**2009**, 101, 817–835. [Google Scholar] [CrossRef] - Lewin, K. Problems of research in social psychology. In Field Theory in Social Science: Selected Theoretical Papers; Cartwright, D., Ed.; Harper & Row: New York, NY, USA, 1951; pp. 155–169. [Google Scholar]
- Spearman, C. The Abilities of Man: Their Nature and Measurement; AMS Press: New York, NY, USA, 1932. [Google Scholar]
- Blum, D.; Holling, H. Spearman’s law of diminishing returns. A meta-analysis. Intelligence
**2017**, 65, 60–66. [Google Scholar] [CrossRef] - Woodley, M.A. The cognitive differentiation-integration effort hypothesis: A synthesis between the fitness indicators and life history models of human intelligence. Rev. Gen. Psychol.
**2011**, 13, 228–245. [Google Scholar] [CrossRef] - Woodley, M.A.; Figueredo, A.J.; Ross, K.C.; Brown, S.D. Four successful tests of the cognitive differentiation-integration effort hypothesis. Intelligence
**2013**, 41, 832–842. [Google Scholar] [CrossRef] - Scarr, S.; McCartney, K. How people make their own environments: A theory of genotype→environment effects. Child Dev.
**1983**, 54, 424–435. [Google Scholar] [PubMed] - Bouchard, T.J. Experience producing drive theory: How genes drive experience and shape personality. Acta Paediatr. Suppl.
**1997**, 422, 60–64. [Google Scholar] [CrossRef] [PubMed] - Freeman, J.; Coyle, T.R.; Baggio, J.A. The functional intelligences proposition. Pers. Individ. Differ.
**2016**, 99, 46–55. [Google Scholar] [CrossRef] - Schmidt, F.L. Beyond questionable research methods: The role of omitted relevant research in the credibility of research. Arch. Sci. Psychol.
**2017**, 5, 32–41. [Google Scholar] [CrossRef] - Brown, K.G.; Le, H.; Schmidt, F.L. Specific aptitude theory revisited: Is there incremental validity for training performance? Int. J. Sel. Assess.
**2006**, 14, 87–100. [Google Scholar] [CrossRef] - Viswesvaran, C.; Ones, D.S.; Schmidt, F.L.; Le, H.; Oh, I.-S. Measurement error obfuscates scientific knowledge: Path to cumulative knowledge requires corrections for unreliability and psychometric meta-analyses. Ind. Organ. Psychol. Perspect. Sci. Pract.
**2014**, 7, 507–518. [Google Scholar] [CrossRef] - Wherry, R.J. A new formula for predicting the shrinkage of the coefficient of multiple correlation. Ann. Math. Stat.
**1931**, 2, 440–457. [Google Scholar] [CrossRef]

1 | Peterson and Brown ([12], p. 180) show that the relation between β and r is independent of sample size and number of predictors and that the imputation of r (given β) yields an estimate similar to the population statistic (ρ) (for a criticism of Peterson and Brown’s [12] approach, see [13]). Given the robust relationship between β and r, βs of 0.10, 0.30, and 0.50 could be described as small, medium, and large, respectively, using Cohen’s [14] criteria for correlations. |

2 | The main analyses analyzed SAT and ACT composite scores, which were the sum of the math and verbal subtest scores. The results replicated in separate analyses of SAT and ACT subtest scores. |

3 | Preliminary support for SLODR comes from Coyle’s [26] study of tilt effects for whites and blacks, two groups that show an average differences in g (favoring whites) of about 1 SD. In general, tilt levels were higher, and tilt relations with specific abilities were stronger, for whites than for blacks (e.g., [26], p. 32). Such a pattern is consistent with SLODR, which assumes that non-g effects (e.g., tilt effects) should be stronger for higher ability groups than for lower ability groups. |

4 | The predictions of the magnification model should be tested after correcting for measurement error, which can increase the predictive power of g relative to non-g factors ([40]; see also, [41]). In addition, corrections for shrinkage should be used to avoid capitalization on chance (e.g., [42], p. 515; see also, [43]), and corrections for range restriction should be used to avoid variance compression, which can reduce effects sizes. |

**Figure 1.**Model of g with the SAT, ASVAB tests (T1–T12), and college GPA. A parallel model (not shown) analyzed the ACT. The symbol “u13” represents the non-g residuals of SAT composite scores (math + verbal), obtained after removing g. The u13→GPA path estimates the relation of the SAT non-g residuals with GPA (β = 0.29). Figure adapted from Coyle and Pillow [6].

**Figure 2.**Model of g with the SAT subtests, ACT subtests, ASVAB abilities. The symbol “u16” represents the SAT math non-g residuals (based on the math subtest), obtained after removing g. The u16→Verbal path estimates the relation of the SAT math non-g residuals with ASVAB verbal ability (β = −0.34). Figure adapted from Coyle et al. [15].

**Figure 3.**Model of g with STEM and humanities GPA factors. g was based on an SAT factor, estimated using SAT scores; a STEM factor, estimated using STEM GPAs, and a humanities factor, estimated using humanities GPAs. The non-g residuals of the SAT subtests, obtained after removing g, were correlated with the STEM and humanities factors. The model shows the relation of the SAT math non-g residuals with the humanities factor (β = −0.19). Figure adapted from Coyle, Snyder, Richmond, and Little [17].

**Figure 4.**Model of g with ASVAB abilities (math, verbal, speed, shop). The symbol “R1” represents the ASVAB verbal non-g residuals, obtained after removing g. The R1→SAT math path estimates the relation of the ASVAB verbal non-g residuals with the SAT math subtest (β = −0.32). Figure adapted from Coyle [27].

**Figure 5.**Magnification model of non-g factors. Non-g effects are predicted to strengthen nonlinearly with ability specialization factors (e.g., ability level, life history, education).

© 2018 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Coyle, T.R.
Non-*g* Factors Predict Educational and Occupational Criteria: More than *g*. *J. Intell.* **2018**, *6*, 43.
https://doi.org/10.3390/jintelligence6030043

**AMA Style**

Coyle TR.
Non-*g* Factors Predict Educational and Occupational Criteria: More than *g*. *Journal of Intelligence*. 2018; 6(3):43.
https://doi.org/10.3390/jintelligence6030043

**Chicago/Turabian Style**

Coyle, Thomas R.
2018. "Non-*g* Factors Predict Educational and Occupational Criteria: More than *g*" *Journal of Intelligence* 6, no. 3: 43.
https://doi.org/10.3390/jintelligence6030043