Open main menu
Home
Random
Recent changes
Special pages
Community portal
Preferences
About Wikipedia
Disclaimers
Incubator escapee wiki
Search
User menu
Talk
Dark mode
Contributions
Create account
Log in
Editing
G factor (psychometrics)
(section)
Warning:
You are not logged in. Your IP address will be publicly visible if you make any edits. If you
log in
or
create an account
, your edits will be attributed to your username, along with other benefits.
Anti-spam check. Do
not
fill this in!
==Practical validity== The extent of the practical validity of ''g'' as a predictor of educational, economic, and social outcomes is the subject of ongoing debate.<ref>{{Cite journal|last1=Kell|first1=Harrison J.|last2=Lang|first2=Jonas W. B.|date=September 2018|title=The Great Debate: General Ability and Specific Abilities in the Prediction of Important Outcomes|journal=Journal of Intelligence|volume=6|issue=3|pages=39|doi=10.3390/jintelligence6030039|pmid=31162466|pmc=6480721|doi-access=free}}</ref> Some researchers have argued that it is more far-ranging and universal than any other known psychological variable,<ref>{{cite journal | last1 = Neubauer | first1 = Aljoscha C. | last2 = Opriessnig | first2 = Sylvia | date=January 2014 |journal=Talent Development & Excellence | title = The Development of Talent and Excellence - Do Not Dismiss Psychometric Intelligence, the (Potentially) Most Powerful Predictor | volume = 6 |issue = 2 |pages = 1β15| url=https://www.researchgate.net/publication/290319779}}</ref> and that the validity of ''g'' increases as the complexity of the measured task increases.<ref name="Jensen 1998, 270">Jensen 1998, 270</ref><ref>Gottfredson 2002</ref> Others have argued that tests of specific abilities outperform ''g'' factor in analyses fitted to certain real-world situations.<ref>{{Cite journal|last=Coyle|first=Thomas R.|date=September 2018|title=Non-g Factors Predict Educational and Occupational Criteria: More than g|journal=Journal of Intelligence|volume=6|issue=3|pages=43|doi=10.3390/jintelligence6030043|pmid=31162470|pmc=6480787|doi-access=free}}</ref><ref>{{Cite journal|last1=Ziegler|first1=Matthias|last2=Peikert|first2=Aaron|date=September 2018|title=How Specific Abilities Might Throw 'g' a Curve: An Idea on How to Capitalize on the Predictive Validity of Specific Cognitive Abilities|journal=Journal of Intelligence|volume=6|issue=3|pages=41|doi=10.3390/jintelligence6030041|pmid=31162468|pmc=6480727|doi-access=free}}</ref><ref>{{Cite journal|last1=Kell|first1=Harrison J.|last2=Lang|first2=Jonas W. B.|date=April 2017|title=Specific Abilities in the Workplace: More Important Than g?|journal=Journal of Intelligence|volume=5|issue=2|pages=13|doi=10.3390/jintelligence5020013|pmid=31162404|pmc=6526462|doi-access=free}}</ref> A test's practical validity is measured by its correlation with performance on some criterion external to the test, such as college grade-point average, or a rating of job performance. The correlation between test scores and a measure of some criterion is called the ''validity coefficient''. One way to interpret a validity coefficient is to square it to obtain the [[Coefficient of determination|variance accounted]] by the test. For example, a validity coefficient of .30 corresponds to 9 percent of variance explained. This approach has, however, been criticized as misleading and uninformative, and several alternatives have been proposed. One arguably more interpretable approach is to look at the percentage of test takers in each test score [[wikt:quintile|quintile]] who meet some agreed-upon standard of success. For example, if the correlation between test scores and performance is .30, the expectation is that 67 percent of those in the top quintile will be above-average performers, compared to 33 percent of those in the bottom quintile.<ref name="sackett2008">Sackett et al. 2008</ref><ref>Jensen 1998, 272, 301</ref> ===Academic achievement=== The predictive validity of ''g'' is most conspicuous in the domain of scholastic performance. This is apparently because ''g'' is closely linked to the ability to learn novel material and understand concepts and meanings.<ref name="Jensen 1998, 270"/> In elementary school, the correlation between IQ and grades and achievement scores is between .60 and .70. At more advanced educational levels, more students from the lower end of the IQ distribution drop out, which restricts the range of IQs and results in lower validity coefficients. In high school, college, and graduate school the validity coefficients are .50β.60, .40β.50, and .30β.40, respectively. The ''g'' loadings of IQ scores are high, but it is possible that some of the validity of IQ in predicting scholastic achievement is attributable to factors measured by IQ independent of ''g''. According to research by [[Robert L. Thorndike]], 80 to 90 percent of the ''predictable'' variance in scholastic performance is due to ''g'', with the rest attributed to non-''g'' factors measured by IQ and other tests.<ref>Jensen 1998, 279β280</ref> Achievement test scores are more highly correlated with IQ than school grades. This may be because grades are more influenced by the teacher's idiosyncratic perceptions of the student.<ref>Jensen 1998, 279</ref> In a longitudinal English study, ''g'' scores measured at age 11 correlated with all the 25 subject tests of the national [[GCSE]] examination taken at age 16. The correlations ranged from .77 for the mathematics test to .42 for the art test. The correlation between ''g'' and a general educational factor computed from the GCSE tests was .81.<ref name="brody2006">Brody 2006</ref> Research suggests that the [[SAT]], widely used in college admissions, is primarily a measure of ''g''. A correlation of .82 has been found between ''g'' scores computed from an IQ test battery and SAT scores. In a study of 165,000 students at 41 U.S. colleges, SAT scores were found to be correlated at .47 with first-year college grade-point average after correcting for range restriction in SAT scores (the correlation rises to .55 when course difficulty is held constant, i.e., if all students attended the same set of classes).<ref name="sackett2008"/><ref>Frey & Detterman 2004</ref> ===Job attainment=== There is a high correlation of .90 to .95 between the prestige rankings of occupations, as rated by the general population, and the ''average'' general intelligence scores of people employed in each occupation. At the level of individual employees, the association between job prestige and ''g'' is lower β one large U.S. study reported a correlation of .65 (.72 [[Correction for attenuation|corrected for attenuation]]). Mean level of ''g'' thus increases with perceived job prestige. It has also been found that the [[Statistical dispersion|dispersion]] of general intelligence scores is smaller in more prestigious occupations than in lower level occupations, suggesting that higher level occupations have minimum ''g'' requirements.<ref name="s&h">Schmidt & Hunter 2004</ref><ref>Jensen 1998, 292β293</ref> === Job performance === Research indicates that tests of ''g'' are the best single predictors of job performance, with an average validity coefficient of .55 across several meta-analyses of studies based on supervisor ratings and job samples. The average meta-analytic validity coefficient for performance in job ''training'' is .63.<ref>Schmidt & Hunter 2004. These validity coefficients have been corrected for measurement error in the dependent variable (i.e., job or training performance) and for range restriction but not for measurement error in the independent variable (i.e., measures of ''g'').</ref> The validity of ''g'' in the highest complexity jobs (professional, scientific, and upper management jobs) has been found to be greater than in the lowest complexity jobs, but ''g'' has predictive validity even for the simplest jobs. Research also shows that specific aptitude tests tailored for each job provide little or no increase in predictive validity over tests of general intelligence. It is believed that ''g'' affects job performance mainly by facilitating the acquisition of job-related knowledge. The predictive validity of ''g'' is greater than that of work experience, and increased experience on the job does not decrease the validity of ''g''.<ref name="Jensen 1998, 270"/><ref name="s&h"/> In a 2011 meta-analysis, researchers found that general cognitive ability (GCA) predicted job performance better than personality ([[Five factor model]]) and three streams of [[emotional intelligence]]. They examined the relative importance of these constructs on predicting job performance and found that cognitive ability explained most of the variance in job performance.<ref>{{Cite journal|last1=O'Boyle Jr.|first1=E. H.|last2=Humphrey|first2=R. H.|last3=Pollack|first3=J. M.|last4=Hawver|first4=T. H.|last5=Story|first5=P. A.|date=2011|title=The relation between emotional intelligence and job performance: A meta-analysis|journal=Journal of Organizational Behavior|volume=32|issue=5|pages=788β818|doi=10.1002/job.714|s2cid=6010387|doi-access=free}}</ref> Other studies suggested that GCA and [[emotional intelligence]] have a linear independent and complementary contribution to job performance. CΓ΄tΓ© and Miners (2015)<ref>{{Cite journal|last1=CΓ΄tΓ©|first1=StΓ©phane|last2=Miners|first2=Christopher|date=2006|title=Emotional Intelligence, Cognitive Intelligence and Job Performance|journal=Administrative Science Quarterly|volume=51|pages=1β28|doi=10.2189/asqu.51.1.1|s2cid=142971341}}</ref> found that these constructs are interrelated when assessing their relationship with two aspects of job performance: [[Organizational citizenship behavior|organisational citizenship behaviour]] (OCB) and task performance. [[Emotional intelligence]] is a better predictor of task performance and OCB when GCA is low and vice versa. For instance, an employee with low GCA will compensate his/her task performance and OCB, if [[emotional intelligence]] is high. Although these compensatory effects favour [[emotional intelligence]], GCA still remains as the best predictor of job performance. Several researchers have studied the correlation between GCA and job performance among different job positions. For instance, Ghiselli (1973)<ref name=":0">{{Cite journal|last=Ghiselli|first=E. E.|date=1973|title=The validity of aptitude tests in personnel selection|journal=Personnel Psychology|volume=26|issue=4|pages=461β477|doi=10.1111/j.1744-6570.1973.tb01150.x}}</ref> found that salespersons had a higher correlation than sales clerk. The former obtained a correlation of 0.61 for GCA, 0.40 for perceptual ability and 0.29 for psychomotor abilities; whereas sales clerk obtained a correlation of 0.27 for GCA, 0.22 for perceptual ability and 0.17 for psychomotor abilities.<ref name=":1">{{Cite journal|last1=Vinchur|first1=Andrew J.|last2=Schippmann|first2=Jeffery S.|last3=S.|first3=Fred|last4=Switzer|first4=III|last5=Roth|first5=Philip L.|title=A meta-analytic review of predictors of job performance for salespeople.|journal=Journal of Applied Psychology|volume=83|issue=4|pages=586β597|doi=10.1037/0021-9010.83.4.586|year=1998|s2cid=19093290}}</ref> Other studies compared GCA β job performance correlation between jobs of different complexity. Hunter and Hunter (1984)<ref>{{Cite journal|last1=Hunter|first1=John E.|last2=Hunter|first2=Ronda F.|title=Validity and utility of alternative predictors of job performance.|journal=Psychological Bulletin|language=en-US|volume=96|issue=1|pages=72β98|doi=10.1037/0033-2909.96.1.72|year=1984|s2cid=26858912}}</ref> developed a meta-analysis with over 400 studies and found that this correlation was higher for jobs of high complexity (0.57). Followed by jobs of medium complexity (0.51) and low complexity (0.38). Job performance is measured by objective rating performance and subjective ratings. Although the former is better than subjective ratings, most of studies in job performance and GCA have been based on supervisor performance ratings. This rating criterion is considered problematic and unreliable, mainly because of its difficulty to define what is a good and bad performance. Rating of supervisors tends to be subjective and inconsistent among employees.<ref>{{Cite journal|last=Gottfredson|first=L. S.|date=1991|title=The evaluation of alternative measures of job performance|journal=Performance Assessment for the Workplace|pages=75β126}}</ref> Additionally, supervisor rating of job performance is influenced by different factors, such as [[halo effect]],<ref>{{Cite journal|last1=Murphy|first1=Kevin R.|last2=Balzer|first2=William K.|title=Systematic distortions in memory-based behavior ratings and performance evaluations: Consequences for rating accuracy.|journal=Journal of Applied Psychology|volume=71|issue=1|pages=39β44|doi=10.1037/0021-9010.71.1.39|year=1986}}</ref> [[facial attractiveness]],<ref>{{Cite journal|last1=Hosoda|first1=Megumi|last2=Stone-Romero|first2=Eugene F.|last3=Coats|first3=Gwen|date=1 June 2003|title=The Effects of Physical Attractiveness on Job-Related Outcomes: A Meta-Analysis of Experimental Studies|journal=Personnel Psychology|language=en|volume=56|issue=2|pages=431β462|doi=10.1111/j.1744-6570.2003.tb00157.x|issn=1744-6570}}</ref> racial or ethnic bias, and height of employees.<ref>{{Cite journal|last1=Stauffer|first1=Joseph M.|last2=Buckley|first2=M. Ronald|title=The Existence and Nature of Racial Bias in Supervisory Ratings.|journal=Journal of Applied Psychology|volume=90|issue=3|pages=586β591|doi=10.1037/0021-9010.90.3.586|pmid=15910152|year=2005}}</ref> However, Vinchur, Schippmann, Switzer and Roth (1998)<ref name=":1" /> found in their study with sales employees that objective sales performance had a correlation of 0.04 with GCA, while supervisor performance rating got a correlation of 0.40. These findings were surprising, considering that the main criterion for assessing these employees would be the objective sales. In understanding how GCA is associated job performance, several researchers concluded that GCA affects acquisition of job knowledge, which in turn improves [[job performance]]. In other words, people high in GCA are capable to learn faster and acquire more job knowledge easily, which allow them to perform better. Conversely, lack of ability to acquire job knowledge will directly affect job performance. This is due to low levels of GCA. Also, GCA has a direct effect on job performance. In a daily basis, employees are exposed constantly to challenges and problem solving tasks, which success depends solely on their GCA. These findings are discouraging for governmental entities in charge of protecting rights of workers.<ref>{{Cite journal|last=Schmidt|first=Frank L.|date=1 April 2002|title=The Role of General Cognitive Ability and Job Performance: Why There Cannot Be a Debate|journal=Human Performance|volume=15|issue=1β2|pages=187β210|doi=10.1080/08959285.2002.9668091|s2cid=214650608|issn=0895-9285}}</ref> Because of the high correlation of GCA on job performance, companies are hiring employees based on GCA tests scores. Inevitably, this practice is denying the opportunity to work to many people with low GCA.<ref>{{Cite journal|last1=Schmidt|first1=Frank L.|last2=Hunter|first2=John E.|title=The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings.|journal=Psychological Bulletin|volume=124|issue=2|pages=262β274|doi=10.1037/0033-2909.124.2.262|year=1998|citeseerx=10.1.1.172.1733|s2cid=16429503 }}</ref> Previous researchers have found significant differences in GCA between race / ethnicity groups. For instance, there is a debate whether studies were biased against Afro-Americans, who scored significantly lower than white Americans in GCA tests.<ref>{{Cite journal|last1=Roth|first1=Philip L.|last2=Bevier|first2=Craig A.|last3=Bobko|first3=Philip|last4=Switzer|first4=Fred S.|last5=Tyler|first5=Peggy|date=1 June 2001|title=Ethnic Group Differences in Cognitive Ability in Employment and Educational Settings: A Meta-Analysis|journal=Personnel Psychology|language=en|volume=54|issue=2|pages=297β330|doi=10.1111/j.1744-6570.2001.tb00094.x|issn=1744-6570|citeseerx=10.1.1.372.6092}}</ref> However, findings on GCA-job performance correlation must be taken carefully. Some researchers have warned the existence of [[Artifact (error)|statistical artifacts]] related to measures of job performance and GCA test scores. For example, Viswesvaran, Ones and Schmidt (1996)<ref>{{Cite journal|last1=Viswesvaran|first1=Chockalingam|last2=Ones|first2=Deniz S.|last3=Schmidt|first3=Frank L.|title=Comparative analysis of the reliability of job performance ratings.|journal=Journal of Applied Psychology|volume=81|issue=5|pages=557β574|doi=10.1037/0021-9010.81.5.557|year=1996}}</ref> argued that is quite impossible to obtain perfect measures of job performance without incurring in any methodological error. Moreover, studies on GCA and job performance are always susceptible to range restriction, because data is gathered mostly from current employees, neglecting those that were not hired. Hence, sample comes from employees who successfully passed hiring process, including measures of GCA.<ref>{{Cite journal|last1=Hunter|first1=J. E.|last2=Schmidt|first2=F. L.|last3=Le|first3=H|date=2006|title=Implications of direct and indirect range restriction for meta-analysis methods and findings|journal=Journal of Applied Psychology|volume=91|issue=3|pages=594β612|doi=10.1037/0021-9010.91.3.594|pmid=16737357|s2cid=14897081}}</ref> ===Income=== The correlation between income and ''g'', as measured by IQ scores, averages about .40 across studies. The correlation is higher at higher levels of education and it increases with age, stabilizing when people reach their highest career potential in middle age. Even when education, occupation and socioeconomic background are held constant, the correlation does not vanish.<ref>Jensen 1998, 568</ref> ===Other correlates=== {{See also|Evolution of human intelligence#Social exchange theory|Evolutionary aesthetics|Evolutionary linguistics|Evolutionary musicology|Sexual selection in humans|Social selection|Wason selection task}} The ''g'' factor is reflected in many social outcomes. Many social behavior problems, such as dropping out of school, chronic welfare dependency, accident proneness, and crime, are negatively correlated with ''g'' independent of social class of origin.<ref>Jensen 1998, 271</ref> Health and mortality outcomes are also linked to ''g'', with higher childhood test scores predicting better health and mortality outcomes in adulthood (see [[Cognitive epidemiology]]).<ref>Gottfredson 2007</ref> In 2004, psychologist [[Satoshi Kanazawa]] argued that ''g'' was a [[Domain specificity|domain-specific]], [[Species-typical behavior|species-typical]], [[Information processing (psychology)|information processing]] [[psychological adaptation]],<ref>{{cite journal|last=Kanazawa|first=Satoshi|author-link=Satoshi Kanazawa|year=2004|title=General Intelligence as a Domain-Specific Adaptation|journal=[[Psychological Review]]|publisher=[[American Psychological Association]]|volume=111|issue=2|pages=512β523|doi=10.1037/0033-295X.111.2.512|pmid=15065920|url=https://psycnet.apa.org/record/2004-12248-010}}</ref> and in 2010, Kanazawa argued that ''g'' correlated only with performance on [[Evolution of human intelligence|evolutionarily unfamiliar rather than evolutionarily familiar]] problems, proposing what he termed the "Savanna-IQ interaction hypothesis".<ref name="Kanazawa 2010a">{{Cite journal|last=Kanazawa|first=Satoshi|date=2010-02-16|title=Why Liberals and Atheists Are More Intelligent|journal=Social Psychology Quarterly|volume=73|issue=1|pages=33β57|citeseerx=10.1.1.395.4490|doi=10.1177/0190272510361602|issn=0190-2725|s2cid=2642312}}</ref><ref name="Kanazawa 2010b">{{cite journal|last1=Kanazawa|first1=Satoshi|date=MayβJune 2010|title=Evolutionary Psychology and Intelligence Research|url=http://personal.lse.ac.uk/Kanazawa/pdfs/AP2010.pdf|journal=[[American Psychologist]]|volume=65|issue=4|pages=279β289|doi=10.1037/a0019378|pmid=20455621|access-date=February 16, 2018|author-link1=Satoshi Kanazawa}}</ref> In 2006, ''[[Psychological Review]]'' published a comment reviewing Kanazawa's 2004 article by psychologists [[Denny Borsboom]] and [[Conor Dolan]] that argued that Kanazawa's conception of ''g'' was empirically unsupported and purely hypothetical and that an evolutionary account of ''g'' must address it as a source of [[Differential psychology|individual differences]],<ref>{{cite journal|last1=Borsboom|first1=Denny|author-link1=Denny Borsboom|last2=Dolan|first2=Conor V.|author-link2=Conor Dolan|year=2006|title=Why ''g'' is not an adaptation: a comment on Kanazawa (2004)|journal=[[Psychological Review]]|volume=113|issue=2|pages=433β437|pmid=16637768|doi=10.1037/0033-295X.113.2.433}}</ref> and in response to Kanazawa's 2010 article, psychologists [[Scott Barry Kaufman]], [[Colin G. DeYoung]], Deirdre Reis, and Jeremy R. Gray published a study in 2011 in ''[[Intelligence (journal)|Intelligence]]'' of 112 subjects taking a 70-item computer version of the [[Wason selection task]] (a [[logic puzzle]]) in a [[social relation]]s context as proposed by [[Evolutionary psychology|evolutionary psychologists]] [[Leda Cosmides]] and [[John Tooby]] in ''[[The Adapted Mind]]'',<ref name="Cosmides & Tooby 1992">{{cite book|last1=Cosmides|first1=Leda|title=The Adapted Mind: Evolutionary Psychology and the Generation of Culture|title-link=The Adapted Mind|last2=Tooby|first2=John|publisher=[[Oxford University Press]]|year=1995|isbn=978-0195101072|editor-last1=Barkow|editor-first1=Jerome H.|editor-link1=Jerome H. Barkow|place=New York|pages=179β206|chapter=3. Cognitive Adaptations for Social Exchange|author-link1=Leda Cosmides|author-link2=John Tooby|orig-date=1992|editor-last2=Cosmides|editor-first2=Leda|editor-last3=Tooby|editor-first3=John}}</ref> and found instead that "performance on non-arbitrary, evolutionarily familiar problems is more strongly related to general intelligence than performance on arbitrary, evolutionarily novel problems".<ref name="Kaufman et al. 2010">{{cite journal|last1=Kaufman|first1=Scott Barry|last2=DeYoung|first2=Colin G.|author-link2=Colin G. DeYoung|last3=Reis|first3=Deidre L.|last4=Gray|first4=Jeremy R.|date=MayβJune 2010|title=General intelligence predicts reasoning ability even for evolutionarily familiar content|url=https://scottbarrykaufman.com/wp-content/uploads/2011/08/Kaufman-DeYoung-Reis-Gray-2011.pdf|journal=[[Intelligence (journal)|Intelligence]]|volume=39|issue=5|pages=311β322|doi=10.1016/j.intell.2011.05.002|access-date=February 16, 2018|author-link1=Scott Barry Kaufman}}</ref><ref>{{cite journal|last1=Kaufman|first1=Scott Barry|date=July 2, 2011|title=Is General Intelligence Compatible with Evolutionary Psychology?|url=https://www.psychologytoday.com/blog/beautiful-minds/201107/is-general-intelligence-compatible-evolutionary-psychology|journal=[[Psychology Today]]|publisher=Sussex Publishers|access-date=February 16, 2018|author-link1=Scott Barry Kaufman}}</ref>
Edit summary
(Briefly describe your changes)
By publishing changes, you agree to the
Terms of Use
, and you irrevocably agree to release your contribution under the
CC BY-SA 4.0 License
and the
GFDL
. You agree that a hyperlink or URL is sufficient attribution under the Creative Commons license.
Cancel
Editing help
(opens in new window)