Source
Patrick, Knee, Canevello and Lonsbary (2007), LIT-1971, JPSP 92, 434–457; read in NOTE-1515.
What was actually shown
Association (Study 1, Table 2, p. 440). Eight samples, N = 1,918, 81% women, corrected and pooled: overall need fulfilment with satisfaction r = .72, commitment .57, perceived conflict −.61, defensiveness −.51, avoidant attachment −.56, self-esteem .42.
Relatedness first (Table 3, p. 440; p. 449). Partial r of relatedness with satisfaction .65 and commitment .59, against autonomy .11 and .07 and competence .09 and .05; in the diary study relatedness alone predicts post-disagreement satisfaction (β = .39) and commitment (β = .46).
Partner effects (Study 2, Table 5, p. 444). In 66 couples, the partner’s overall fulfilment predicts one’s own satisfaction (β = .31), perceived conflict (−.17) and defensiveness (−.14) beyond one’s own; relatedness fulfilment interacts across partners (satisfaction t(62) = 3.19).
Motivation as a carrier (Study 3, Table 7, p. 451). Fulfilment predicts the relative autonomy of one’s reasons for being in the relationship (β = .62); with those reasons controlled, fulfilment’s β on post-disagreement satisfaction falls from .42 to .15 and on commitment from .41 to .02.
The needs are not separated (pp. 439, 441). Autonomy–relatedness .86, competence–relatedness .84, autonomy–competence .96 (CI .88 to 1.0) corrected, .68 uncorrected; autonomy’s α is as low as .59.
This could have come out otherwise: the three subscales could have been distinct, and each could have predicted relationship outcomes equally.
What this does not say
- That fulfilment causes relationship quality. “all of the studies were correlational, thus precluding causal inferences” (p. 454).
- That autonomous motivation mediates. The predictor and mediator are measured together, once; the diary outcomes follow both.
- That relatedness matters more than autonomy and competence as needs. With subscales this collinear, partial correlations split shared variance; the paper’s composite check (p. 441) addresses autonomy and competence, not relatedness against them.
- That item overlap is ruled out. It is argued (p. 441), not tested.
Connections
- QUESTION-369. Evidence of the kind the question says cannot settle it, and a further problem the question does not state: a relationship-level measure of three needs that behaves, in these data, as one or two.
- THEORY-054. Study 3 reports a simplex among relationship-motivation subscales and then scores one weighted index (p. 447); integration is not measured.