3 ms·
The paper is behind a paywall and not open-access, so here are some pieces of information I thought were interesting: "Intelligence" was measured by an individ
by fwip 6y ago
The paper is behind a paywall and not open-access, so here are some pieces of information I thought were interesting:
"Intelligence" was measured by an individual's performance on the "Armed Forces Qualifying Test" (AFQT).
"Grit" is self-reported on the "Short Grit Scale" which asks participants to rate themselves from 1 to 5 on statements like "I am diligent."
"Success" was measured by level of education attained and hourly wage rate in 2015.
Control variables include: Gender, household income as a child and level of education by parents. Study participants were born in the early 1980's.
Top-level findings:
> Although in our representative sample of the American population grit does explain educational and job-market success
over and beyond other predictor variables such as socioeconomic status, intelligence, and the Big Five personality characteristics, its absolute effects are rather minimal. In particular, its effects are negligible in comparison to the effect of intelligence. Finally, the effects of grit were also considerably weaker than the effects of conscientiousness.
- puranjay 6y ago> Grit" is self-reported on the "Short Grit Scale" which asks participants to rate themselves from 1 to 5 on statements like "I am diligent." This is a pretty awful way to measure grit and makes me doubt the results
- tsumnia 6y agoThis is an issue with most self-report surveys.
- gwern 6y agoBut the measurement of Conscientiousness was much worse, and it still outpredicted Grit considerably: "Given the NLSY 2-item measure of conscientiousness and the 8-item measure of grit, this comparison is likely to overestimate the effect of grit relative to conscientiousness (Crede et al., 2012 for the consequences of using short measures of the Big Five personality traits)." Plus, going from an 8-item test to a longer one isn't going to close a gap with IQ that large, and as they point out, the reliability is decent enough and pro-Grit authors have no trouble making claims based on, so what's sauce for the goose is sauce for the gander: "The reliability of the scale in our data was .72, which is comparable to the reliability reported by Duckworth and Quinn (2009) and within the ranges reported in the meta-analysis of Crede et al. (2017)."
- thyrsus 6y agoThank you for this. What is the difference between their measure of grit and their measure of conscientiousness?
- fwip 6y agoConscientiousness was measured by the "The Ten-Item Personality Inventory (TIPI)" > a 10-item measure of the Big Five (or five-factor model) dimensions of personality: openness, conscientiousness, extroversion, agreeableness, and emotional stability