Wednesday, December 30, 2020

Comparing the effects of a mindfulness versus relaxation intervention on romantic relationship wellbeing: No significant differences

Comparing the effects of a mindfulness versus relaxation intervention on romantic relationship wellbeing. Johan C. Karremans, Gesa Kappen, Melanie Schellekens & Dominik Schoebi. Scientific Reports volume 10, Article number: 21696 (2020). Dec 10 2020.

Abstract: There is increasing scientific interest in the potential association between mindfulness and romantic relationship wellbeing. To date, however, experimental studies using active control groups and testing dyadic effects (i.e. examining both actor and partner effects) are lacking. In the current study, romantically involved individuals engaged for 2 weeks daily in either guided mindfulness exercises, or guided relaxation exercises. Participants, and their partners, completed measures of relationship wellbeing at pre- and post-intervention, and at 1-month follow up. The mindfulness intervention significantly promoted relationship wellbeing, for both participants (i.e. actor effects) and their partners (i.e. partner effects). However, these findings did not significantly differ from changes in relationship wellbeing in the relaxation condition. Theoretical implications of these findings for understanding the association between mindfulness and romantic relationship wellbeing are discussed. Moreover, the findings are discussed in light of recent debates about the relative lack of proper control groups in mindfulness research.

General discussion

Does mindfulness promote the wellbeing of romantic relationships? The possible causal effect of mindfulness training on romantic relationships has received very little empirical attention so far. After 2 weeks of daily guided mindfulness practice, participants in the mindfulness intervention group reported significantly higher levels of relationship satisfaction, lower relationship distress, felt more connected to the partner, and were more accepting towards the partner (but not more excited about their relationship). These effects were maintained 1 month after the intervention. Their partners, who did not engage in mindfulness practice, also reported higher relationship satisfaction, less distress, and felt more accepted by their partners (but not more connected; and they reported actually less relationship excitement, a finding we further discuss below). In general, these findings may be considered as promising for the effectiveness of mindfulness intervention in promoting relationship wellbeing. A similar pattern of findings, however, was found in the active control condition. Participants who received a daily guided relaxation intervention for 2 weeks showed similar relationship benefits (in fact, they showed significantly more positive change in partner acceptance, and positive change in relationship excitement), as did their partners. Thus, while the current findings suggest that the daily practice of mindfulness generally can lead to various beneficial relationship outcomes, relaxation practice on a daily basis yielded comparable outcomes.

What is an adequate interpretation for these findings? First, these results might simply mean that any intervention, similar in structure, would positively affect self-reports of relationship wellbeing (sometimes referred to as a trial effect58). For example, the intervention may have prompted participants to reflect on their relationship, which could have promoted positive feelings and thoughts about the relationship (although in theory it might also do the opposite, of course). Also, engaging in an intervention could have led participants to report desired rather than actual outcomes. The fact that intervention participants showed similar changes in both conditions, across most relationship outcome measures, and that their partners showed similar changes (except regarding relationship excitement), makes this a possible interpretation of the current findings.

An alternative interpretation might be that the relationship benefits may be genuine, both in the mindfulness as well as relaxation intervention. In the introduction, we discussed several theoretical reasons why mindfulness may promote relationship wellbeing (for more details10), and the results in the mindfulness group may reflect such theorized effects. Similarly, the results obtained in the relaxation group may reflect true relationship benefits of daily relaxation exercises. For example, daily relaxation might reduce overall psychological and physiological stress levels, which in turn might positively affect how people behave and respond to their partners18,59. Accordingly, both mindfulness and relaxation intervention may have promoted relationship wellbeing, but possibly through different mechanisms. The obtained effects in both groups, however, also may have been caused by a similar ‘relaxation mechanism.’ While mindfulness is theoretically distinguishable from relaxation, in reality it is possible that the mindfulness exercises in our study increased feelings of relaxation to a similar extent as the relaxation exercises. As noted previously60, the historical development of (mindfulness) meditation and relaxation techniques as therapeutic strategies in the past century show substantial overlap, and indeed one of the challenges of studying the effects of mindfulness is to distinguish them from ‘mere’ relaxation effects61.

It could be theorized that longer mindfulness training is required to promote relationship wellbeing above and beyond effects of relaxation. Participants practiced mindfulness for 2-weeks, engaging in a relatively short guided mindfulness meditation each day. While such relatively short mindfulness interventions have revealed significant effects on various outcome measures42,43,44 (but rarely have been compared with active control groups), effects of mindfulness practice may occur and generalize to real life outcomes only after more extensive training, when paying non-judgmental attention becomes a more or less automatic manner of relating to one’s experiences62,63. How people cope with and respond to their experiences is the result of a lifelong process, and therefore difficult to ‘re-program’64,65. Likewise, particularly in long-term romantic relationships, interaction patterns between partners and appraisals about the relationship tend to become habitual, and are therefore difficult to change in the short run66. Thus, what the ‘dosage’ of mindfulness practice should be to potentially promote relationship wellbeing, and how much for whom, requires more research.

The results regarding relationship excitement revealed a somewhat different pattern than the other relationship outcome measures. There was no significant change in relationship excitement in the mindfulness group among intervention participants. This finding seems inconsistent with previous findings by Carson and colleagues, who found increases in relationship excitement after a mindfulness intervention35. The present results may suggest that it was not mindfulness per se leading to such changes in their study, but the fact that both partners engaged in their intervention may have explained an increase in relationship excitement. Interestingly, in the current study the partners of the mindfulness intervention participants actually reported significant declines in relationship excitement. We have previously speculated that there may be theoretical reasons to predict such declines13. For example, mindfulness may be associated with reductions in impulsivity, a potentially important source of relationship excitement67. This finding awaits future research and replication.

The results of the current study speak to the broader issue of the need for proper active control conditions in mindfulness research. A substantial percentage of research on mindfulness is lacking active control conditions, using waiting list controls, or examining changes in a variable of interest from pre- and post-mindfulness intervention68,69. Such studies can be informative to see whether mindfulness intervention is associated with any changes on variables of interest, or for example to study moderators of any effects of mindfulness intervention (e.g. for whom it does and does not ‘work’). Although the inclusion of an active control condition made the current findings perhaps more ambiguous and difficult to interpret, it highlights the importance of studying whether changes associated with mindfulness intervention can be attributed uniquely to mindfulness. In many previous articles on mindfulness research, the conclusion that mindfulness positively affects a certain outcome of interest is often unwarranted when an active control group is lacking32,40. The current findings raised a number of issues (as discussed above) that may remain unaddressed when control conditions are lacking, underscoring the need for active control groups to get a more nuanced and more comprehensive understanding of mindfulness and its effects, both for the individual and the relationship.

Similarly, the current findings have implications for couple intervention research more broadly. In the past decades, researchers have examined effects of various prevention and intervention couple programs. Some programs have obtained significant relationship benefits in the short- and longer-term70,71, others obtained negative outcomes (e.g. increased awareness that one is lacking relationship skills72, but often waiting list control groups are used to compare intervention outcomes. Relatively few studies used comparisons with active control groups, and if they did, similar effects have been found between target and control intervention groups73. Thus, there is a need for proper control interventions in order to examine relationship interventions more rigorously.

Before closing, some limitations should be noted. First, all outcome measures were self-reports, and future studies should include more objective measures of relationship functioning and wellbeing, such as observational coding of partner interactions, and/or physiological assessments of (relationship) distress. Second, as noted already, the interventions were relatively short. For example, examining the effects of longer protocolized interventions65, such as the mindfulness-based stress-reduction program on relationship outcomes, while including an active control intervention, would be a logical and important next step. Similarly, examining associations between length and frequency of mindfulness practice and relationship outcomes would be a valuable and complementary approach. Finally, a potential limitation of the current study is that the intervention in the current research targeted one member of the couple. While it is a theoretically interesting question whether the cultivation of mindfulness in one individual transfers to the relationship partner, mindfulness intervention might be more effective when both partners engage in the intervention.

The present study was the first to examine the causal impact of mindfulness training on romantic relationship outcomes using an active control group and testing both actor and partner effects, thereby extending previous research that has linked romantic relationship wellbeing mainly to self-report measures of mindfulness. The current study does not give definitive answers to the question whether or not mindfulness can causally affect relationship wellbeing, but does provide a compelling example of why research on mindfulness interventions would benefit from a wider use of active control groups, hopefully offering a springboard for future research. 

The vast majority students in a college sample believe that students can simply be bad test-takers; the majority also believe that they themselves are bad test-takers, a perspective which is maladaptive in light of relevant research

The Bad Test-Taker Identity. Jeffrey D. Holmes. Teaching of Psychology, December 29, 2020.

Abstract: There is widespread belief that test-taking ability is an influential component of academic success distinct from domain knowledge and comprehension. Most of today’s college students took many more tests over the course of their primary and secondary education than students of previous generations, and also participated in regular training to strengthen their test-taking skills. Although such training and experience should equalize students on any isolated test-taking ability, the present study reveals that the vast majority students in a college sample believe that students can simply be bad test-takers. Moreover, the majority of students believe that they themselves are bad test-takers, a perspective which is maladaptive in light of relevant research. Accordingly, the data show that students who identify in this way also tend to possess other maladaptive academic attitudes.

Keywords: test-taking beliefs, test-taking self-efficacy, bad test-taking

Taller individuals were less supportive of government wealth redistribution overall, especially if they were wealthier; but effects were equally strong in males & females, inconsistent with current evolutionary theories

Height is associated with more self-serving beliefs about wealth redistribution. Thomas Richardson. Evolution and Human Behavior, Dec 30 2020.

Abstract: People vary widely in their attitudes towards how much their government should redistribute wealth. Evolutionary theory may shed light on why this variation occurs. Numerous studies have established an association between upper body strength and attitudes towards equality and wealth redistribution in males, showing that physically stronger men are more likely to hold self-serving beliefs on these issues. This effect is typically weaker or absent in women. A question that has received little attention is whether there are similar associations between other aspects of formidability and attitudes towards wealth redistribution. One such aspect is height. I tested this prediction using data from the European Social Survey, in a sample of 27031 people from 20 European countries. Results show that taller people are more likely to have self-serving attitudes towards government redistribution of wealth. The result was robust to numerous control variables and alternative model specifications, but the direct effects of height were small. Taller individuals were less supportive of government wealth redistribution overall, but were especially averse if they were also wealthier. Post-hoc analyses suggested that for lower income deciles, the association was reversed. For these people, there was a positive association between height and support for wealth redistribution. However, effects were equally strong in males and females, and so are not fully consistent with current evolutionary psychological theories of resource distribution.

Keywords: HeightEvolutionary political psychologyFormidability

4. Discussion

In a large sample of over 27000 people from 20 European countries, I find that taller people are slightly less supportive of government wealth redistribution, and that being tall exacerbates the negative effect of income on attitudes towards wealth redistribution. Equal support for was found for both a negative main effect of formidability (Price et al., 2011) and a formidability x income interaction (Petersen et al., 2013) as assessed by the AIC of the final models. These effects remained when controlling for several possible mediators such as age, education, overall political orientation or whether a respondent has a position of authority at work. These results suggest that taller individuals are more likely to endorse negative and/or self-serving, attitudes towards government wealth redistribution. Furthermore, I found a similar interaction between height and income on conservatism, but only for men.

The effects found were not as large as previous studies that investigated muscularity. Indeed, the effects were quite small. When the effect was examined for each of the 20 countries separately, few countries showed significant effects, though nearly all were in the predicted direction, and the overall result was not driven by the countries that showed the largest effects. One reason for this is the use of single item measures for egalitarianism and political orientation, as well as the use of self-reported height, all of which increase measurement error. This may be why the effects were often non-significant when broken down by country, as fig. 2 show large uncertainty around the coefficients. Previous studies have used detailed measures of support for inequality, such as the social dominance orientation questionnaire (e.g. Price et al., 2011), which might have reduced the standard errors around the effects in this study.

A smaller effect than previous studies that measured muscularity is also consistent with a formidability-based explanation. This is because muscularity is more strongly associated with strength and conflict success than height is (Sell, Tooby, & Cosmides, 2009von Rueden et al., 2008). The beta coefficients indicated that the effects of height, as well as the height x income interaction, were similar in magnitude to the effects of 1 standard deviation increase in age (17.3 years) or the direct effect of being male in the same model. It should also be noted that the effects reported are the direct effects of height and the height income interaction effect on redistribution attitudes. I controlled for a larger number of possible mediators than is typical for studies of this type, which reduces the unique effects of height compared to previous studies. Height will also have an influence on attitudes indirectly through increasing occupational success, educational attainment and income (Judge & Cable, 2004Meyer & Selmer, 1999Stulp et al., 2013).

The results are consistent with previous findings that formidability negatively predicts support for redistribution and egalitarianism (Petersen et al., 2013Price et al., 2017) but also that it interacts with wealth (Petersen & Laustsen, 2019Price et al., 2011). The relationship is significant even when controlling for education and authority at work, indicating that the effects are not driven by the increased social standing and success that is associated with height. While height and a height x income interaction were found to predict greater conservatism, the effect on attitudes towards wealth redistribution remains when controlling for overall political orientation (as in Petersen et al., 2013). I also found that the relationship between conservatism and aversion to government redistribution of wealth is weak and highly variable between countries. Taken together, these suggests that there the effects of height are specific to wealth redistribution attitudes and not just a by-product of changes in general conservatism.

The lack of a sex-specific effect is not consistent with most previous work that suggests the association between formidability and attitudes is only seen in men (Petersen et al., 2013; Petersen and Laustsen, 2018; but see Kerry & Murray, 2019). In fact, when the analyses were split by sex the effect was significant in women but not men, though the effect in women was not significantly larger than the effect in men. This result is not necessarily predicted by the evolutionary psychological theories used to explain formidability effects. In the environments where we evolved, being more formidable would not have provided women with more status and resources as they did not engage in significant amounts of violent contest competition. Similarly, increased female height is not consistently associated with authority and status in modern societies (Bielicki & Charzewski, 1983Case & Paxson, 2008Gawley et al., 2009Hamstra, 2014). Given the large and diverse sample size of the present study, it is highly unlikely the lack of result is due to low statistical power.

The lack of sex difference in effects may be due to the use of self-reported height rather than measured height. Both Petersen and Laustsen (2019) and Kerry and Murray (2019) note that subjective, but not objective measures of upper body strength are negatively associated with support for redistribution in women and conservatism in women. Directly measured height may not have shown significant effects for females. Another issue with self-reported height is that, if anti-egalitarianism causes respondents to overestimate their height, this could introduce a causal path from attitudes to height in this study. As ESS data is collected through in-person interviews, large biases in self-reported height are likely to be minimal because they would raise suspicion. Nonetheless, this only predicts that biases in height would be small, and the effects found were small, so reverse causality cannot be conclusively ruled out as an alternative explanation.

Another possible explanation for the lack of sex difference may be that taller women, while not benefitting directly from their height, benefit indirectly when it comes to wealth and status. First, as height is highly heritable (Yang et al., 2010), tall females will be born to taller families, and as status and resources are shared within families, tall females may gain from the success of their male relatives. There is also the possibility that taller females learn their attitudes towards wealth redistribution from their taller fathers. Additionally, as humans mate assortatively for height (Stulp, Simons, Grasman, & Pollet, 2017), taller women may seek and attract taller men, so may accumulate status and resources through their tall, formidable husbands. This latter point would also explain the difference with studies of muscularity, as humans do not seem to mate assortatively for muscularity.

A link between height and opposition to wealth redistribution may have been selected for in males and show up in females as a by-product, a process known as sexually antagonistic selection. Sexually antagonistic selection can result in both sexes showing a trait that is adaptive for one sex even if it is maladaptive to the other sex (Rice, 1992). There is evidence that sexually antagonistic selection might have occurred in humans for some traits (e.g. Camperio-Ciani, Corna, & Capiluppi, 2004Garver-Apgar, Eaton, Tybur, & Thompson, 2011Lee et al., 2014). This speculative explanation does assume that the fitness advantage conferred to ancestral males was large enough to offset any disadvantages to females, an assumption that may not be plausible considering the small effect size found in this study.

The differences between previous work and the current study could be explained by the differences in measures. The most obvious difference is that the present study tested height, whereas previous studies almost all tested upper body strength and muscularity. Both height and muscularity are thought to be key components of formidability (Blaker & van Vugt, 2014), and were both intrasexually selected to be higher in males throughout our evolutionary history (Hill et al., 2017Puts, 2010). However upper body strength is far more sexually dimorphic than height is, with some studies estimating that male strength is on average around 3 standard deviations higher than women's (Lassek & Gaulin, 2009). In the present study, the average male height was 1.75 SD above the average female height. Smaller sexual dimorphism in height than upper body strength might imply smaller sex differences in effects. It is worth noting that men show reduced support for redistribution compared to women, which is what we'd predict as men are taller, though the difference was small. That said, sexual dimorphism in height is still substantial by most standards, so it is unclear why it produces no detectable sex differences in its effects on attitudes.

The results found here build on an emerging literature that shows our political psychology is influenced by variables that are arguably irrelevant to modern politics, in this case our height. Other literature has found that facial and vocal characteristics of political candidates can impact voting decisions (; Laustsen, Petersen, & Klofstad, 2015). If our political beliefs are affected by factors that have little relevance to political processes (such as our own height) it can threaten the very effectiveness of democracy. For this reason it is important to study these factors more, so that people can be made aware of their potential biases, and they can be addressed if necessary and possible.

One avenue of further research is to establish the mechanism by which height is associated with attitudes. If it is evolved, genetic factors may be primarily responsible. For example, genes that predispose greater height may also predispose aversion to wealth redistribution. Another possible mechanism is reactive heritability (Tooby & Cosmides, 1990), where individuals may be evolved to calibrate their behaviours or attitudes to their phenotype (in this case, their height). Finally, there may be more sociocultural factors that account for the effect that I could not test here. Given the small effect size, implications of this study are only likely to be seen at large scales. Further, this study shows that height differences between individuals are associated with differences in attitudes. Researchers should be careful not to assume that the same effect would be found within individuals. It does not necessarily follow from these findings that increasing the height of a given person (such as through better nutrition during development) will lower their adult levels of support for wealth redistribution. Further research is required to confirm this.

In conclusion, in a large sample of Europeans I find that taller people are less supportive of government redistribution of wealth than shorter people, especially when they have a high income. This relationship is independent of a large range of possible covariates and is found equally in males and females. This is partially consistent with evolutionary psychological theories of resource distribution, but the lack of a sex-specific effect remains to be conclusively explained.