Tipping points of change: Everyday fluctuations in oneself and the social world create ambiguities about when people will diagnose lasting, qualitative change (and therefore act)

When Small Signs of Change Add Up: The Psychology of Tipping Points. Ed O’Brien. Current Directions in Psychological Science, December 12, 2019. https://doi.org/10.1177/0963721419884313

Abstract: Things change, but the exact point at which they do is often unknown. After how many loveless nights is a relationship “officially” in trouble? After how many happy days has one’s depression “officially” passed? When do recurring patterns in the climate or economy “officially” warrant a response? When is a person’s identity “officially” accepted? Everyday fluctuations in oneself and the social world create ambiguities about when people will diagnose lasting, qualitative change (and therefore act). Recent research documents these tipping points of change as a psychological process, shaped by individual and situational forces. People judge tipping points asymmetrically across valence and asymmetrically across time. Here, I review discoveries and outline future directions in tipping-points research.

Keywords: tipping points, change perception, self/others over time, evaluative judgment, qualitative and categorical shifts

Road Map for Future Research

Downstream behavior

Tipping points imply points when people become more likely to intervene or surrender. Future research should scale to higher-stakes contexts (e.g., changes in health, climate change action, decisions to change jobs or partners). The valence asymmetry suggests uphill battles for appreciating improvement. The temporal asymmetry suggests conflict between parties who experience evidence from different perspectives (e.g., policymakers may predetermine thresholds for reward or punishment that notoriously prove too high for constituents, who demand action at the first salient strike). Indeed, naive realism in change perceptions may stir conflict over identical evidence (Campbell, O’Brien, Van Boven, Schwarz, & Ubel, 2014). Other research should assess intrapersonal costs (e.g., consumers may overpay for lengthy product trials, assuming they will evaluate more than they actually will before drawing conclusions).

Motivated and nonmotivated mechanisms

If the basic process underlying tipping points is responding to evidence salience, there must be motivated sources of salience that interact with tipping points. Alcoholics may view themselves as more “cured” after their first week of sobriety than friends view them, CEOs may quickly view increases in revenue as signals whereas investors view them as noise, voters may dismiss a few days of poor stock returns or rising unemployment if they support the incumbent administration, and a person who goes on one date with an attractive partner may conclude that he or she is “the one.” More research should unpack potential self/other differences, as agents of change likely want to diagnose change. However, this may also reflect nonmotivated differences in accessibility (Klein & O’Brien, 2017; O’Brien, 2013). Only the alcoholic actor knows how effortful that first week felt; he or she actually has a more diagnostic signal. Differences across explicit and implicit change perceptions (Ferguson et al., 2019) may be more informative.

Other boundaries

Beyond self/other differences, testing still other factors that reverse the asymmetries is critical. When do people tip more quickly in response to improvement? Future research should assess additional domain differences (e.g., changes in identity-central features; Strohminger & Nichols, 2014) and individual differences (e.g., trait optimists may flip the valence asymmetry, assuming they reject entropy beliefs). When do people tip more slowly than they think? Extremely emotional events are often rationalized in ways hidden to intuition (Wilson & Gilbert, 2005), and thus may flip the temporal asymmetry; people may assume one horrible fight will forever render a friend a foe, but in reality, friends work to stay friends. For complex stimuli, reacting quickly to initial evidence may itself be mistaken; one may assume that a single reading of a book was enough to form a conclusion, but in reality, rereads may continually reveal new interpretations (Kardas & O’Brien, 2018; O’Brien, 2019). Regardless, the phenomenon appears not easily intuited; future research should assess other ways in which expectations diverge from experiences.

Evidence presentation

Future research should introduce more variance into observations. Variance likely will not affect asymmetries across conditions if it is similarly distributed (e.g., random draws of grades that slowly transition to C+s vs. A+s at equal rates), but extreme draws likely matter; one big shock may disrupt small compounding change. Future research should also integrate the full time course of tipping points. As retrospection and prospection rely on shared lay beliefs (O’Brien, Ellsworth, & Schwarz, 2012; Schwarz, 2012), the temporal asymmetry may stubbornly persist when looking back; people may predict being patient, then quickly make up their minds, yet then later recall being just as patient as imagined. However, other stereotypes about past and future selves (such as past selves seeming emotional and future selves seeming rational: O’Brien, 2015) may interact with tipping-point perceptions over time.

External benchmarks

Some changes are truly instantiated, which can be misperceived because of other attentional demands (Simons & Ambinder, 2005), miscalibrated beliefs (Davidai & Gilovich, 2015; Ross, 1989), and shifting reference points (Levari et al., 2018). An open question is whether tipping-point thresholds can be objectively quantified. Misperceiving genuine tipping points would bear on many real-world outcomes, from doctors who must anticipate when illnesses will manifest to investors who must anticipate when bear markets will return. One could gain traction on this question by comparing perceptions to other benchmarks, such as normative thresholds (e.g., feverish people may think their temperature has crossed 100.4° F before it does) and mathematical probabilities (e.g., testing how quickly people believe drawn outcomes have shifted from pool A to pool B against Bayesian standards; Massey & Wu, 2005). More research is needed, from all approaches, on categorical change perception in the self and others.
A broad study of tipping points is promising. The point when things change may be fiction, but hopefully this article encourages initial change toward these exciting directions.