nucleation

THEORY-1316: In eight samples pooled to 3,692 respondents, the covariance of the BPNSFS items is organized more by item keying than by need: factors sharing a keying direction correlate .64 to .75 while satisfaction and frustration of the same need correlate .17 to .38, a higher-order model grouping the six scales by keying fits better than one grouping them by need, and the frustration items carry almost no thwarting content

3 min read

Properties4
StatusActive
Tagspsychometrics, motivation
SourceLIT-1978 — The Basic Psychological Need Satisfaction and Frustration Scales probably do not validly measure need frustration.
Bears onQUESTION-369 — What would show that an experience is a basic psychological need, something an organism's functioning requires, rather than something that reliably raises reported well-being, when self-determination theory identifies needs by their effects on well-being and treats well-being as the presence of need satisfaction?

Graph View

Source

Murphy, Watts, Baker, Don, Jolink and Algoe (2023), LIT-1978, Psychological Assessment 35(2); read in NOTE-1530 from the author manuscript, PMC11554446.

What was actually shown

The items. The twelve BPNSFS frustration items are all negatively keyed and the twelve satisfaction items all positively keyed. None of the frustration scales has “substantial active thwarting content”; two competence items are reverse-worded twins of satisfaction items (“I have serious doubts about whether I can do things well” / “I feel confident that I can do things well”).

The structure, pooled N = 3,692 (eight samples, 153 to 926 each; seven of university students; English).

  • Two-factor EFA: one factor of all positive items, one of all negative items. Three-factor EFA: the first two still pure keying (Table 1).
  • Six-factor EFA intercorrelations: frustration factors among themselves .64 to .70; satisfaction factors .66 to .75; satisfaction with frustration of the same need .17, .34 and .38 (frustration items reverse-scored).
  • Higher-order CFA with three need factors over the six scales: RMSEA .064, CFI .924, TLI .915, BIC −28,734, with need-factor correlations .94 to 1.11. With two keying factors: RMSEA .052, CFI .950, TLI .944, BIC −29,631.

External correlates sorted by keying (all items scored toward well-being). The frustration items correlate more with the negatively keyed items of external well-being scales than the satisfaction items do, in all five measures, with Steiger’s z of 6.50, 11.16 and 3.58 (self-compassion, La Guardia need satisfaction, perceived stress) and 1.63 and 1.90 (depression, self-esteem). The satisfaction items’ excess with positively keyed items appears in two of five measures.

This could have come out otherwise. Had the needs been the stronger grouping, the need-level higher-order model would have fitted better and the same-need correlations would have exceeded the same-keying ones. They did not.

What this does not say

  • That the frustration scales measure only keying. The authors say the confound between keying and construct is complete in the BPNSFS, so no factor analysis can decide it. That inference, and what would settle it, is THEORY-1311.
  • That need frustration does not exist or does not matter. The paper attacks a measure, not the construct, and points to the thwarting scale (LIT-1954) as closer to it.
  • Anything about other languages, versions or populations. Three of the scale’s fifty-odd versions were used.
  • Anything about the autonomy scales that the paper’s own table supports. The text says at one point that the Autonomy Frustration items load weakly on Autonomy; its Table 3 shows the Autonomy Satisfaction items doing so (.34 to .49) and the frustration items loading .62 to .76 (NOTE-1530, Limitations).

Backlinks