THEORY-1345: In three cross-sectional studies of undergraduates, most of the variance in adult attachment security lay within persons, across relational partners (56–79%, error included), and relationship-specific satisfaction of the needs for autonomy, competence and relatedness predicted relationship-specific security in every relationship tested, also with the relatedness items removed; relatedness was the strongest unique predictor, then autonomy, then competence

Source

La Guardia et al. (2000), LIT-1979, pp. 371–380; read in NOTE-1523.

What was actually shown

• Within-person variance 56% (Study 1), 79% (Study 2), about two thirds (Study 3).
• HLM slopes of security on within-person need satisfaction, all p < .001, in all three studies, and with the relatedness items removed (Table 1); residual regressions significant in every specific relationship (Table 3).
• Relatedness strongest in 15 of 18 simultaneous regressions; autonomy stronger than competence in 15 of 18 without relatedness (pp. 375–376).

These could have come out otherwise: within-person variance could have been mostly error unexplained by need satisfaction, or the effect could have vanished without the relatedness items. Neither happened.

What this does not say

• That need satisfaction produces security. Cross-sectional; the authors say the relation is bidirectional and could reflect third variables (p. 381).
• That the measures separate the needs. Patrick et al.’s corrected correlations among them run to .96 (NOTE-1515).