Today’s positive affect predicts tomorrow’s experience of meaningful coincidences: a cross-lagged multilevel analysis

ABSTRACT

The perception of meaningful patterns in random arrangements and unrelated events takes place in our everyday lives, coined apophenia, synchronicity, or the experience of meaningful coincidences. However, we do not know yet what predicts this phenomenon. To investigate this, we re-analyzed a combined data set of two daily diary studies with a total of N = 169 participants (mean age 29.95 years; 54 men). We investigated if positive or negative affect (PA, NA) predicts the number of meaningful coincidences on the following day (or vice versa). By means of a cross-lagged multilevel modelling approach (Bayesian estimation) we evaluated with which of two theoretical assumptions the data are more in line. First, if meaningful coincidences are facilitated by a broader and more flexible thinking style, PA should positively predict meaningful coincidences at the following day. However, if the experience of meaningful coincidences signifies a strategy to cope with negative feeling states, NA should predict the experience of meaningful coincidences during the following day. In favour of a more flexible thinking style, we found that PA predicted the number of perceived coincidences the following day. We did not find any effect for NA, and therefore, no evidence arguing for the coping mechanism hypothesis of meaningful coincidences.

KEYWORDS:

• Granger causality
• synchronicity
• apophenia
• serendipity
• seriality

Many of us might have experienced meaningful coincidences – a sudden perception of a significant connection between objectively unrelated events such as the moment when we recognise that a dream or a part of a dream that we had last night comes true (Diaconis & Mosteller, 1989; for a collection of synchronicity experiences see Kammerer, 1919). Although meaningful coincidences and synchronicity are phenomena that many people experience throughout their daily lives (see e.g. Beitman, 2011; Bressan, 2002; Roxburgh et al., 2016), we do not know much about the reasons why these experiences occur (for potential explanations see e.g. Beitman, 2011; Diaconis & Mosteller, 1989; Matthews & Blackmore, 1995). The perception of meaningful coincidences is mostly studied as a personality trait showing considerable interindividual differences (Bressan, 2002; Coleman & Beitman, 2009; Rominger et al., 2011). Research on these interindividual differences suggest that the propensity to experience meaningful coincidences is associated with paranormal beliefs and positive schizotypy (Bressan, 2002; Brugger et al., 1990; van Elk et al., 2016), delusions (Morrison & Murray, 2009), belief in conspiracy theories (see van Elk et al., 2016), reduced working memory capacities (Rominger et al., 2011; Rominger et al., 2019), depressive symptoms (Russo-Netzer & Icekson, 2022), altered brain functioning during sensory perceptions, working memory tasks, and during resting conditions (Rominger et al., 2018; Rominger et al., 2019; Rominger et al., 2023), as well as an altered brain structure (Unger et al., 2021). However, studies also reported associations with more benevolent traits. For example, people who perceive more meaningful coincidences reported a higher frequency of everyday creative activities, more creative achievements (Rominger et al., 2023), as well as a higher life satisfaction (Russo-Netzer & Icekson, 2022), which points to certain benefits in “relating the unrelated”.

However, studying the experience of meaningful coincidences (also labelled synchronicity, apophenia, serendipity, and seriality; see e.g. Beitman, 2011) from an intraindividual perspective is less established. This is astounding since the dynamic assessment within individuals might help to find more detailed answers when and why people experience meaningful coincidences in their everyday lives. This is even more important as we do not know how to experimentally induce the experience of meaningful coincidences or synchronicity. This leaves observational studies the only logical choice to investigate potential mechanisms. In a recent three-study approach we collected data of the experience of meaningful coincidences and affective states in an internet daily diary study lasting one week (Rominger et al., 2023). By means of these available data, we aimed to further evaluate whether dynamic changes in positive affect (PA) and negative affect (NA) might impact the perception of meaningful coincidences and synchronicity throughout the life as it is lived.

Our goal in this study was to evaluate two potential theories explaining why people experience (more or less) meaningful coincidences. First, the experience of meaningful coincidences and synchronicity might reflect a coping strategy relevant for achieving more life satisfaction in the long run (Nachman, 2009; Russo-Netzer & Icekson, 2022; Unger et al., 2021). In line with this, Beitman (2009) suggested that our brain aims to reduce negative emotions by finding patterns in a chaotic and unpredictable reality (i.e. emotion-focused coping, Folkman, 1984). Subjectively establishing or increasing order might reduce negative feelings and induce comfort and feelings of safety in a constantly changing and potentially adverse environment. In line with this, Coleman and Beitman (2009) reported a positive association between meaningful coincidences and NA as well as PA (see also Rominger et al., 2023). If synchronicity serves as a coping strategy, NA should impact the “need” to experience more meaning in meaningless noise (during the next day) in order to deal with prevalent negative emotions. Alternatively, we might also observe the consequences – that is that more meaningful coincidences associated with higher PA and lower NA during the following day. To sum up, if the coping strategy hypothesis has predictive value, we would expect that NA predicts meaningful coincidences or that meaningful coincidences predict an increase of PA and a decrease of NA. Any of these patterns of cross-lagged correlations would argue for the suggested coping function of meaningful coincidences (Beitman, 2009; Russo-Netzer & Icekson, 2022; Unger et al., 2021).

Second, we can theorise that the experience of meaningful coincidences might take place due to a broadened thinking style allowing a more flexible perception and connection of events and semantic concepts (Rominger et al., 2011). This more flexible style of thinking would allow to connect more loosely associated elements hidden from the view of others and to creatively make sense out of (randomly occurring) perceptions, experiences, events, and thoughts. In line with this, studies showed that meaningful coincidences, positive schizotypy, as well as creativity are linked to wider associations (e.g. Gianotti et al., 2001; Mohr et al., 2001; Rominger et al., 2022; Rossmann & Fink, 2010). PA can broaden the thinking style, an assumption deeply rooted in the broaden and build theory as well as in the dual pathway model of creativity (Fredrickson, 2004; Nijstad et al., 2010). Fredrickson (2004) suggested that PA broadens the thought-action repertoire and therefore, PA increases the chance to develop novel and creative ideas. de Dreu et al. (2008) suggested that PA can increase creativity via cognitive flexibility. Furthermore, PA is related with increased dopamine release in the brain (Ashby et al., 1999; Boot et al., 2017), which is associated with curiosity, exploration, and novelty seeking, ultimately founding an exploratory state of the mind (Ivancovsky et al., 2023; Jauk, 2019). If we consider the experience of meaningful coincidences as an active cognitive process, in terms of finding connections between randomly occurring events, PA should broaden the thinking style. This might take place due to an increase of exploration, curiosity-driven behaviour, as well as creativity fostering the sudden perception of coincidences in random arrangements. These cognitive processes should benefit from PA (Baas et al., 2008). If these assumptions are valid, PA, but not NA, should predict the number of experienced meaningful coincidences the next day, while meaningful coincidences should not predict PA the next day.

By applying an internet daily diary method in combination with cross-lagged multilevel modelling (i.e. random-intercept cross lagged panel model, Hamaker et al., 2015; Lucas, 2023; Usami, 2021), we investigated the direction of the association between the experience of meaningful coincidences and PA as well as NA. This allowed to investigate if meaningful coincidences are more likely a result of PA (associated with a broadened thinking style) or NA (arguing for a strategy to cope with adverse situations).

Methods

Participants

For the purpose of this study, we reanalyzed the aggregated data of a preregistered study (Rominger, 2023). We included all participants who provided answers throughout two consecutive days at least. This resulted in a sample of 169 participants (54 men, 1 divers) with a mean age of 29.95 years (SD = 14.63). All participants gave informed consent before participating in the online survey. The local ethics committee of the university of Graz approved this study (GZ. 39/82/63 ex 2014/15). Data and syntax can be found at https://osf.io/9zfsr/.

Daily diary method

Participants received a daily email with a link to the online survey, which included eight questions to the experience of meaningful coincidences and 20 items of the PANAS to assess affective well-being. The internet daily diary study started on Monday, where also further questionnaires were answered. The two studies were conducted during May and June 2022 and 2023. The approach of both studies was comparable as the second study served as a replication of the first study. As the only minor difference, the second daily diary study applied additional questionnaires such as daily questions asking for creative activities and for synchronicity experiences (for more details see Rominger et al., 2023).

Experience of meaningful coincidences

Participants answered how often they perceived meaningful coincidences during the day by means of 8 items, each capturing different categories (taken from the coincidence questionnaire; Bressan, 2002). Example items are “Unexpected solution of a problem (like meeting a friend who wants to sell his computer exactly when we were looking for one)” or “Spontaneous associations (like thinking of someone and running unexpectedly into that person soon afterwards)” (for all items see Bressan, 2002). When at least one meaningful coincidence was perceived (out of the 8 items), we counted this day as a day with a meaningful coincidence (coded as 1), otherwise the day was without a meaningful coincidence (coded as 0). In total, 262 days (39.5%) of the 664 days were without an experience of a meaningful coincidences. Twenty-one participants (12.4% of the total sample) perceived no single meaningful coincidence. ICC analyses indicated that 33.5% of the variance of meaningful coincidences was due to between-person, leaving up to 66.5% for within-person (plus error) variance.

Affect

Each day, participants answered 20 items of the German version of the Positive and Negative affect Schedule (PANAS; Krohne et al., 1996). The reliability of the PA and NA were good for both between person (PA: R_kR= .79, NA: R_kR= .82) and within person (PA: R_C= .90, NA: R_C= .80). The ICC showed that 38.2% and 43.8% of the sum variance were due to between-person differences for PA and NA.

Statistical analyses

We used multilevel cross-lagged panel models with random intercept for participants to assess bidirectional pathways linking affect (i.e. PA, NA) and the experience of meaningful coincidences across days (Schuurman et al., 2016; for similar applications see Glatz & Schwerdtfeger, 2022; Wekenborg et al., 2022). This allows the investigation of the directionality of associations by a simultaneous estimation of relations. Each variable at the first measurement time point (day n) predicts variables at the next measurement time point (day n + 1; cross-lagged component). These relations are controlled for longitudinal stability (autoregressive component). The results reflect the effects of the constructs at the first time point on changes in other constructs relative to the baseline level, thereby controlling for the other pathways. We used the R package brms (Version: 2.18.0; Bürkner, 2017; non-informative priors, 4 chains, 20,000 iterations, and 50% warm-up) fitting Bayesian models using the probabilistic programming language Stan to estimate the parameters in the multilevel cross-lagged models for PA and NA, respectively. We used a Bernoulli distribution, because we had to deal with the dichotomised variable of perceiving coincidences at the next day as outcome variable. We used Gaussian distribution when the outcome variable was the affective state of the next day (i.e. PA or NA). We considered effects as meaningful when the credible intervals (CIs) did not include zero.

Results

Testing a broadened thinking style as a facilitator for meaningful coincidences

The cross-lagged multilevel analysis showed that PA predicted the experience of meaningful coincidences of the next day n + 1 (b = 0.034; 95% CI [0.004, 0.066]), while meaningful coincidences did not predict PA of day n + 1 (b = − 0.690; 95% CI [− 1.900, 0.508]). This indicates a unidirectional association from PA to the frequency of meaningful coincidences on the following day (n + 1, see Figure 1).

Figure 1. Cross-lagged associations between meaningful coincidences and PA. The numbers indicate unstandardised effect estimates. *CI does not include zero. CIs are in parenthesis.

The results remained similar when adding NA on day n and day n + 1 to the cross-lagged model. The cross-lagged effect of PA on day n on coincidences on day n + 1 was even slightly more pronounced (b = 0.046; 95% CI [0.014, 0.081]).

Testing meaningful coincidences as a coping strategy

The cross-lagged multilevel analysis for NA showed no significant effect. Specifically, NA on day n did not predict meaningful coincidences on day n + 1 (b = 0.027; 95% CI [−0.020, 0.075]) and meaningful coincidences on day n did not predict NA on day n + 1 (b = 0.528; 95% CI [−0.225, 1.290]; for more details see Figure 2). The results remained non-significant when adding PA of day n and day n + 1 to the cross-lagged model.

Figure 2. Cross-lagged associations between meaningful coincidences and NA. The numbers indicate unstandardised effect estimates. *CI does not include zero. CIs are in parenthesis.

Discussion

In this combined re-analysis of two available internet daily diary studies (Rominger et al., 2023), we investigated by means of two multilevel cross-lagged panel models if PA or NA bidirectionally correlate with meaningful coincidences. By means of this procedure, we tested two assumptions why meaningful coincidences might occur. First, we might experience more meaningful coincidences in everyday life to cope with adverse situations (i.e. emotion-focused coping; Unger et al., 2021), thus suggesting a path from NA to the experience of meaningful coincidences. Second, we could assume that meaningful coincidences take place due to a more flexible and broadened thinking style (Rominger et al., 2011), thus suggesting a path from PA on day n to the experience of meaningful coincidences on day n + 1.

Importantly, we found that PA, but not NA, predicted meaningful coincidences. The experience of meaningful coincidences, in turn, did not predict affective states (neither PA nor NA). The unidirectional association of PA with coincidences clearly favours the model of a broadened thinking style as one explanation to perceive more meaningful patterns in everyday life. This is in line with the broaden and built theory suggesting that PA causes broadening in attention, cognition, and action (Fredrickson, 1998; Fredrickson & Branigan, 2005). This interpretation nicely corresponds with a study of Rominger et al. (2011), which found that people perceiving more meaningful coincidences also showed broader and wider associations to given word pairs. Furthermore, people who perceive more meaningful coincidences also seemed to perceive more meaning in random arrangements (Rominger et al., 2018), with a similar pattern also found for self-rated creativity (Rominger et al., 2022).

Furthermore, these findings are in accordance with research showing connections between creativity and the experience of meaningful coincidences (Rominger et al., 2023), positive schizotypy (Acar & Sen, 2013), as well as pareidolia (Diana et al., 2021). In this respect, curiosity, exploration behaviour, and novelty seeking might be driving mechanisms of meaningful coincidences, which are associated with dopamine concentration in the brain (Ivancovsky et al., 2023; Jauk, 2019). PA might increase cognitive flexibility via dopamine increases (Ashby et al., 1999; Boot et al., 2017; Dreisbach & Goschke, 2004). The more dopamine, the more curious we might be about our environment, which might benefit creative thoughts (Boot et al., 2017). This might further go along with a higher sensitivity to elaborate and detect novel stimuli, which is in line with the assumption that the experience of meaningful coincidences and apophenia are associated with a lower threshold to detect signals in noisy surroundings (Blain et al., 2020; Brugger & Graves, 1997) and to make connections between mental patterns and outer events (Brugger, 2001).

This study’s findings do not support the assumption that the experience of meaningful coincidences may serve as a coping mechanism increasing life-satisfaction (Russo-Netzer & Icekson, 2022; Unger et al., 2021). NA did not predict the number of coincidences, and meaningful coincidences did not predict PA (Coleman & Beitman, 2009). This null finding does not mean that these pathways do not exist at all, however, it shows that NA might not that strongly (and consistently) drive the day-to-day variation of the experience of meaningful coincidences, which in turn, does not strongly affect PA and NA (Cristofaro, 2021). This pattern might be different when investigating less frequently occurring but highly significant life events – such as the death of a beloved one, divorce, life threatening illnesses, marriage and births (Beitman, 2011; Greyson, 2011). Furthermore, this first evidence of a within-person pattern of findings does not argue against the observations that synchronicity is associated with the trait variables of NA, PA, and life satisfaction (see Coleman & Beitman, 2009; Russo-Netzer & Icekson, 2022) as well as other relevant third variables such as depressive symptoms (Russo-Netzer & Icekson, 2022).

The directional association between PA and the experience of meaningful coincidences underlines the validity of the assessment and extends the trait perspective of a link between PA and the perception of meaning in everyday life (Coleman & Beitman, 2009) to a state perspective. This finding indicates that PA may shape the experience of meaningful coincidences (probably also in the long run; see Beitman, 2011). Not only PA states fluctuate from day-to-day, but also the experience of meaningful coincidences, and both seem to covary at this within-person level of assessment with a cross-lag of one day. The observed strong autoregressive term indicates temporal stability of the experience of meaningful coincidences (at least in the NA cross-lagged model; see Hamaker et al., 2015), which should not be equalised with the personality trait perspective (Bressan, 2002; Russo-Netzer & Icekson, 2022; Unger et al., 2021). People, who perceive more meaning today will more likely perceive more meaning in random events tomorrow (as well as people, who perceive less today will perceive less tomorrow), irrespective of their trait level. However, the dynamic approach of the present study allows a deeper look into the mechanisms explaining why and when the chance to perceive meaning in randomly occurring phenomena in our environment might rise (or drop), controlling for trait level and temporal stability. Investigating the within-person changes in combination with sophisticated multilevel cross-lagged models might help us to better understand the driving forces behind the perception of meaning coincidences in noisy surroundings more deeply.

A limitation of this study is the short observation of one week, which only allowed to calculate 6 cross-lags at the best. Future studies should therefore target to collect diary data of more than a week to replicate and generalise these findings. Furthermore, we did not assess control beliefs, which might affect the directional link between NA and the experience of meaningful coincidences as a need to establish order when uncertainty prevails (Beitman, 2009). As a further limitation, we should note that the applied approach does not allow to conclude strict causality (which only controlled experiments can reach; Hamaker et al., 2015); however, it allows first insights into potentially relevant mechanisms (Schuurman et al., 2016). Furthermore, the present study constitutes an indirect test of the mechanisms proposed to foster the perception of meaningful coincidences. We did not directly observe negative life events or changes in cognitive flexibility, but observed changes in PA and NA. It should also be considered that a time-lag of one day might be too large to observe the hypothesised coping mechanism of meaningful coincidences, which might be a more time-contingent effect. A similar criticism is valid for evaluating the broaden and build effect of PA on the experience of meaningful coincidences. Although we found a significant unidirectional link, the time-lag of one day can only capture more long-lasting effects of PA. However, previous research was able to indicate effects of PA with a one-day delay (see e.g. Conner & Silvia, 2015; Fredrickson et al., 2021; Karwowski et al., 2017).

Still, we embedded our conclusions in well-founded theories assuming connections between PA and cognitive functioning as well as NA and adverse life events with emotion-focused coping (Folkman, 1984). Future studies should target to measure the broadened and more flexible thinking style more directly by means of cognitive tasks, association tasks, and creative ideation tasks as well as dopamine concentrations in parallel with PA states, to capture the involved mechanisms more precisely. Furthermore, these studies should also target to increase the sampling rate from a daily to a momentary assessment, which would allow a more direct and time-sensitive observation of the effects of interest.

Conclusions

By applying an internet daily dairy approach across one week, we studied the dynamic of meaningful coincidences in daily life. This approach allowed to investigate the pathways involved when perceiving meaning in random arrangements of events. This study showed that today’s PA positively predicted tomorrow’s odds to perceive meaningful coincidences, while NA did not. This observation does not provide evidence for the emotion-focused coping mechanism idea of meaningful coincidences and is more in line with the hypothesis that a broader and more flexible thinking style might increase the frequency to experience meaningful coincidences.

Disclosure statement

No potential conflict of interest was reported by the author(s).