Contribution
It is the primary study behind SDT’s reply that autonomy is not individualism. It measures autonomy as the relative internalization of one’s reasons for a practice, applies it to practices of each of Triandis’s four cultural orientations, and shows in four countries that more autonomous enactment of any type of practice, collectivist and vertical included, goes with more well-being (pp. 97–99, 104–106).
Key insight
Autonomy and independence are on different axes. Dependence versus independence concerns whether one relies on others; autonomy versus heteronomy concerns whether one endorses what one does. “One can be autonomously dependent” (p. 98), and one can conform to group norms volitionally if one endorses them, as one can obey a traffic officer volitionally if one values traffic laws. So a collectivist practice can be enacted autonomously, and the cultural critique of SDT’s autonomy need “equated SDT’s constructs of autonomy and self-determination with making choices independently from one’s reference group” (p. 97).
Assumptions
- Autonomy is relative internalization, measured per practice by the Self-Regulatory Questionnaire of Cultural Practices: “Why do you or would you do this?”, four reasons rated 1–5, combined as the relative autonomy index −2·external −1·introjected +1·identified +2·integrated (pp. 101– 103). The “integrated” reason reads in part “I feel free in choosing and doing it, and feel responsible for the outcomes” (p. 102).
- Intrinsic motivation is excluded as not pertinent to internalization of instrumental practices (p. 99).
- Culture is measured as perceived norms: participants rated how important and frequent each practice is among “most Americans [Koreans, Russians, Turks]” (p. 102).
- Well-being is a composite of standardized SWLS, Short Index of Self-Actualization, Rosenberg self-esteem and six CES-D items (reversed) (pp. 102, 104).
- Measurement comparability by MACS, accepting strong invariance at CFI ≈ .89–.90 by a difference-in-fit criterion (pp. 102–104).
Key results
Samples (Table 1, p. 101)
- USA 195, Russia 159, Turkey 94, South Korea 111; university students; 73% women in the US and Russian samples, 27% in Korea, 43% in Turkey. The Turkish sample was much poorer (48% “low–low” income; 68.5% of fathers without high school).
Perceived cultural norms (p. 103, Table 3)
- Turkish data on perceived practices were “noncomparable” and excluded from that comparison. Relative to the US (latent means), Koreans saw their culture as more horizontally (0.42) and vertically (0.75) collectivist and less horizontally (−0.71) and vertically (−0.22) individualist; Russians as more vertically collectivist (0.32) and less vertically individualist (−0.59). Russia “a mixed model”.
Internalization of practices (pp. 103–104, Table 4)
- Across samples, horizontal practices mean index 4.68 against vertical 1.60, t(557) = 27.20, p < .0001. Unpredicted: individualist above collectivist within each dimension (horizontal 5.08 vs 4.45; vertical 2.01 vs 1.19).
- US students internalized most practices least (raw vertical collectivism −0.71). Men internalized vertical practices more than women in three of four samples.
Autonomy and well-being (pp. 104–105, Tables 6–7)
- Composite well-being regressed on one internalization index, three country contrasts and their interactions, one model per practice type: β = .24 (horizontal individualism), .24 (horizontal collectivism), .27 (vertical individualism), .29 in Table 6 (vertical collectivism); R² ≈ .19–.20; none of the twelve country × internalization interactions significant (Table 6 betas −.08 to .00).
- Within each country, all sixteen internalization betas positive and significant: Korea .23–.28, Russia .17–.24, Turkey .34–.42, USA .15–.22 (Table 7).
- With overall internalization and the likelihood of performing individualist and collectivist practices: only internalization (β = .17) and collectivist practice (β = .09) were significant; no interactions (p. 105). No gender × internalization interactions.
- Country contrasts carried more well-being variance than autonomy (Contrast 1, US against the rest, β ≈ .37–.44).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Autonomy (volition) is conceptually distinct from independence and individualism | conceptual | pp. 97–99; definitional within SDT |
| C2 | More autonomous enactment of any type of cultural practice goes with more well-being | moderate | four regressions and sixteen within-country betas, pp. 104–106; cross-sectional |
| C3 | The autonomy–well-being relation does not differ by country or gender | weak–moderate | non-significant interactions; small samples (94–195) |
| C4 | Vertical practices are harder to internalize than horizontal ones | moderate as a mean difference | t(557) = 27.20; item sets chosen by the authors, who caution that other vertical practices might be internalized (p. 107) |
| C5 | Participants’ cultures differ in the directions Triandis’s scheme predicts | moderate | perceived-norm latent means, three countries; Turkey not comparable |
Concepts
- Autonomy — behaviour “experienced as willingly enacted”, fully endorsed; its opposite is heteronomy, not dependence (p. 98).
- Dependence / independence — reliance or non-reliance “on others for guidance, support, or needed supplies” (p. 98).
- Reflective vs reactive autonomy (Koestner) — endorsement against resistance to influence (p. 98).
- Horizontal / vertical; individualism / collectivism — Triandis’s equality versus hierarchy and individual versus group priority (pp. 99–100).
Connections
- Iyengar and Lepper 1999 (LIT-1949) is the paper’s chief target. It reports them as finding that experimenter-imposed choice undermined intrinsic motivation in both Asian and Anglo American children, and that choice by trusted others “uniquely enhanced intrinsic motivation for the Asian group”, with the interpretation focused on the latter (p. 97). This is Chirkov et al.’s account; the original is not held.
- Oishi 2000 and Miller 1997 are criticized for measuring autonomy as individualist values or as freedom from all influence (pp. 97–98).
- Sheldon et al. 2001 (LIT-1966), cited for event autonomy predicting well-being in Korea and the USA (p. 99).
- Reis et al. 2000 (LIT-1968) and Sheldon, Ryan and Reis 1996 are cited as the only diary studies of the needs, both American (p. 107).
Bearing on the record
- NOTE-441 holds this study second-hand (p. 1577 of Ryan and Deci 2006; its C8), as “Autonomous reasons predicted well-being with no moderation by culture in Russia, the United States, Turkey and South Korea”. The primary text bears that out. Two qualifications for anyone leaning on it: the samples are 94 to 195 students per country, and “no moderation” is twelve non-significant interaction terms, not a bound on their size.
- THEORY-047 and THEORY-054 (internalization and regulation): primary evidence that the internalization continuum, scored as a relative autonomy index, predicts well-being outside the US, for practices the participants did not choose.
- QUESTION-369. The paper is about whether autonomy’s link to well-being is universal, not about what makes autonomy a need. Its argument for need-hood is the universality of the link, which is one of the nine criteria; it still uses well-being as the criterion, and the autonomy measure’s top anchor includes “I feel free”, which is close to a well-being report.
- The definition carries the thesis. The claim that autonomy is not individualism is true by SDT’s definition; what is empirical is that the so-defined autonomy predicts well-being for collectivist practices too. The paper tests the second; the first is argued.
- Filed from it: THEORY-1325 (Proposed).
- No instruction for machine-learning practice.
Limitations
- University students; cross-sectional (“it is also conceivable that high well-being leads people to view their behaviors as more autonomous”, p. 107); four cultures only (the authors’).
- Measurement invariance was accepted below conventional cut-offs: for well-being, strong invariance RMSEA = .11, CFI = .89 (p. 104); for the autonomy index CFI = .90. Turkey’s perceived-norm data were not comparable.
- The four internalization indices are never entered together, and their betas are nearly identical (.24–.29), which suggests a general person-level tendency to give autonomous reasons more than a practice-specific effect; the paper does not report their intercorrelations.
- A reporting error. The text for vertical collectivism (p. 105) repeats the vertical individualism figures word for word (β = .27, F = 28.55 for the index; Contrast 1 .44, F = 82.50; Contrast 2 .30, F = 33.10), while Table 6 gives vertical collectivism .29, .43 and .25.
- The practices were rated hypothetically (“would you do”), so autonomy for practices one does not perform is included.
Open questions
- Would the relation hold with outcome measures that are not self-reports, and in non-student samples, in Africa and South America as the authors ask (pp. 107–108)?
- Does the relative autonomy index for cultural practices carry variance beyond a general autonomy orientation?
- Do longitudinal designs show autonomy preceding well-being across cultures?