Method
Recruitment
Participants were recruited via posters placed in local community buildings, university buildings, and on local websites, as well as through word of mouth. Volunteers were requested who were aged 18 years of age or over, spoke fluent English, and had little experience of meditation (currently meditating less than once per week). Individuals interested in participating in the experiment were provided with an information sheet about the study. Those who remained interested having read this further information were invited to a testing session. In total, 55 participants attended this session and completed the experimental procedure. A further five were recruited but reported high levels of depression or anxiety during initial screening for eligibility within the experimental session and so were excluded before completing all measures. All participants received a small prize for participating in the study (e.g., confectionary). The study was approved by Oxford University Research Ethics Committee.
Materials and Measures
Beck Depression Inventory II (BDI-II; Beck et al. 1996)
The Beck Depression Inventory is a widely used 21-item self-report instrument for measuring severity of depressive symptoms. The BDI-II has high internal consistency and high 1-week test–retest reliability (Beck et al. 1996). Individuals scoring 20 or more on the BDI-II were excluded.
Beck Anxiety Inventory (BAI)
The Beck Anxiety Inventory is a 21-item questionnaire that measures the severity of symptoms of anxiety. Individuals scoring 19 or more on the BAI were excluded.
FFMQ (Baer et al. 2006)
See description in “Method” of “Study 1”.
MEPGAP—Adapted (Vincent et al. 2004; Hadley and MacLeod 2010)
The Measure to Elicit Positive Future Goals and Plans was utilized as described for study 1. However, in the current study, participants were randomly allocated to one of two conditions. In condition A, they completed ratings of goals 1, 3, and 5 before the meditation or rest period and goals 2, 4, 6 after the mediation or rest period, whereas in condition B, they did the reverse.
Goal Adjustment Questionnaire (Wrosch et al. 2003)
The Goal Adjustment Questionnaire is a 10-item self-report questionnaire, containing two subscales which explore the ability to disengage from unattainable goals (four items), and reengage with new goals when existing goals become unattainable (six items). Goal re-engagement, the ability to find new sources of meaning in life when existing goals are challenged, has the most relevance for conditional goal setting, since it appears to directly contrast with the belief that happiness, self-worth, and fulfillment can only be attained through achievement of particular, existing goals. Items on the re-engagement scale include “If I have to stop pursuing an important goal in my life… I convince myself that I have other meaningful goals to pursue (I seek other meaningful goals/I put effort toward other meaningful goals).” The goal re-engagement scale had excellent internal consistency in the current sample, α = 0.88.
Manipulations
Meditations
The two guided meditations utilized in the study were both approximately 15 min in length and were equated with regard to the amount of time during which the participants were receiving auditory instructions and the amount of time left silent for participants to practice on their own. In the guided breathing meditation, participants were invited to focus their attention on the sensations of their breathing, before broadening their field of awareness to bodily sensations as a whole. They were instructed that if they noticed that their mind had wandered during the meditation, they should notice where their attention had gone and then gently return their attention to the breath or body, without judging themselves.
The loving-kindness meditation began by asking participants to think about a good thing that they had done, or a good quality that they had, or if nothing came to mind to gently focus on their own wish to be happy. After this initial period, they were encouraged to direct feelings of loving-kindness towards themselves by silently repeating four phrases, “may I be safe,” “may I be happy,” “may I be healthy,” and “may I live with ease.” Once they had established loving kindness for themselves, they were asked to use the same phrases to cultivate loving kindness first for someone close, then to someone neutral, then to someone they did not know firsthand, and finally to all people, everywhere. These instructions were based on loving kindness, or metta, meditation as described by Salzberg (2002) and have been used in our previous research (see Barnhofer et al. 2010).
Rest Condition
Participants in the rest condition were instructed that they should find a position that was comfortable and then sit and relax, but try to remain awake. The period of rest lasted 15 min. The experimenter remained quiet and did not disturb the participant during the rest period.
Allocation of Participants to Groups
Participants were randomly allocated to one of six groups defined by allocation to either breathing, loving kindness, or rest and by completion of CGS ratings for goals 1, 3, and 5 preintervention and 2, 4, and 6 postintervention (A/B), or the reverse (B/A). Randomization to groups was conducted through the use of pre-prepared envelopes containing allocations, with an allocation picked at random from the envelope for each consecutive participant. Nineteen participants were randomly allocated to the breathing meditation group (10 females, nine males; nine A/B), 19 to the loving-kindness meditation group (12 females, seven males; 10 A/B), and 17 to the rest group (nine females, eight males; nine A/B). No main effects or interactions were observed when CGS order (A/B or B/A) was entered into the analyses, and so analyses were repeated (and are reported here) without this as a factor.
Procedure
Testing took place individually in a quiet room. At the start of the session, participants were given the information sheet to read and the opportunity to ask any questions. Written informed consent was then obtained. Participants began by filling in a series of questionnaires which gathered information on basic demographic characteristics, mood state (the BDI and BAI), and prior experience of meditation. Following review of this information, those participants who remained eligible (little experience of meditation, BDI and BAI < 20) completed the MEPGAP interview and made conditional goal setting ratings for three of their six most important goals. The procedures for the meditation or rest period were explained. The experimenter remained unobtrusive during the period of meditation or rest that followed to prevent distraction. After the meditation, the participant completed the second set of CGS ratings. Following the completion of these, they were fully debriefed and thanked for their time.
Data Analysis
Inspection of key variables indicated that CGS scores did not contain any outlying values. The pre-meditation/rest CGS score departed moderately from normality, while the post-meditation/rest CGS score was normally distributed. Since transformation of the pre-CGS score did not significantly improve the distribution, the untransformed data were retained.
Results
Participant Characteristics
Participants ranged in age from 20 to 63 years, with a mean of 27 years (SD = 9.62). The majority of the participants were white (76%); 11% were Asian, 7% were of mixed ethnicity, 2% were black, 2% were Hispanic, and 2% did not identify with any of the previous categories. Twenty-nine percent of the sample had completed only compulsory schooling, 62% had obtained an undergraduate degree, and 9% had obtained a postgraduate degree. Regarding marital status, 55% of the participants were single, 29% were in a relationship, 9% were married or cohabiting, and 7% were either separated, divorced, or widowed.
BDI-II and BAI Scores
The mean BAI score for the sample as a whole was 3.96 (SD = 3.13, range 0–13), out of a maximum total of 63, and the mean BDI score was 4.92 (SD = 4.15, range 0–14), also out of a maximum total of 63. Thus, the sample had low mean levels of depression and anxiety.
Conditional Goal Setting, Goal Re-engagement, and FFMQ
The mean total CGS score, across items measuring happiness, self-worth, and fulfillment for those goals rated before the meditation or rest period was 28.05 (SD = 11.77). This corresponds to a score of 3.06 (SD = 3.12) on each 1 to 8 scale, indicating that happiness, self-worth, and fulfillment were not regarded as highly contingent on successful attainment of important goals in this sample. In order to explore whether there was a significant association between CGS and dispositional mindfulness in this healthy community sample, baseline CGS scores were correlated with total FFMQ score. This revealed a significant negative correlation, r = −0.36, p = 0.007 (Spearman’s rho = −0.35, p = 0.008), suggesting that the previously found association between low dispositional mindfulness and high conditional goal setting is not confined to clinical groups (see Fig. 2).
The mean goal re-engagement score was 23.23 (SD = 3.96). Pearson’s correlation coefficients indicated that goal re-engagement was significantly correlated with total FFMQ score, r = 0.37, p = 0.005. Thus, people who had higher levels of dispositional mindfulness are more able to identify new rewarding goals when existing goals become unattainable. However, baseline levels of conditional goal setting and goal re-engagement were not significantly correlated with one another, r = −0.22, p = 0.11.
Effects of Meditation on Ratings of Conditional Goal Setting
In order to explore the effects of brief periods of meditation on ratings of CGS, we conducted a two (time: pre and post) by three (meditation type: breathing, loving kindness, and rest) repeated-measures ANOVA with CGS rating as the dependent variable. This revealed a main effect of time, F (1, 52) = 6.58, p = 0.013, but no significant main effect of group and no significant time × group interaction. Comparison of the means across the sample as a whole indicated that, contrary to expectation, levels of CGS increased slightly from pre- (M = 28.05, SD = 11.77) to post-meditation (M = 32.15, SD = 13.81) or rest across the sample as a whole.
In order to explore whether goal adjustment ability influenced reactions to meditation, we conducted a repeated-measures ANOVA, entering goal re-engagement ability as a second between-subjects factor. Participants were divided into three equal groups according to whether they reported low (eight breathing, six LK, and six rest), moderate (five breathing, eight LK, and four rest), or high (six breathing, five LK, and seven rest) levels of goal re-engagement. Changes in CGS were analyzed in a two (time: pre and post) by three (group: breathing, loving kindness, and rest) by three (re-engagement: low, moderate, and high) mixed ANOVA. As before, this revealed a main effect of time, F (1, 46) = 12.46, p = 0.001. There was also a significant two-way time × re-engagement interaction, F (2, 46) = 2.49, p = 0.02, qualified by a significant three-way time × re-engagement × group interaction F (4, 46) = 5.55, p = 0.001.
The three-way interaction was broken down into separate time × group ANOVAs for participants low, moderate, and high in goal re-engagement ability. The ANOVA for participants low in goal re-engagement ability revealed a significant main effect of time, F (1, 17) = 37.85, p < 0.001, and a significant time × group interaction, F (2, 17) = 16.50, p < 0.001. Post hoc tests indicated a significant increase in CGS ratings from pre- to post-meditation in participants allocated to the loving-kindness meditation condition (pre: M = 26.67, SD = 9.48; post: M = 49.50, SD = 7.87; p < 0.001), but no significant change in those allocated to breathing (pre: M = 31.25, SD = 12.98; post: M = 34.75, SD = 13.23, p = 0.18) or rest M = 33.67, SD = 13.23; post: M = 36.50, SD = 15.04, p = 0.34). In those low in goal re-engagement ability, allocated to loving-kindness meditation CGS scores increased from an average rating of approximately 3, to a rating of 5.5 on an eight-point scale. As shown in Fig. 3, this increase appeared to be highly consistent across the participants. In contrast, the ANOVAs for participants moderate or high in goal re-engagement ability indicated no significant main effects or interactions.
Discussion
The two studies presented in this paper explored the association between mindfulness and conditional goal setting and the impact of meditation training on these variables. Study 1 examined changes in dispositional mindfulness and conditional goal setting in a sample randomized to either immediate or delayed treatment with MBCT. Although there was no evidence of significant differential change in the two groups, patterns of change in mindfulness and CGS favored those allocated to MBCT, and across the sample as a whole increases in dispositional mindfulness over the 3–4-month period were associated with significant decreases in CGS. Study 2 confirmed the cross-sectional association between dispositional mindfulness and CGS in a sample with low levels of depression and anxiety. These findings together support the suggestion that a high level of dispositional mindfulness is associated with low conditional goal setting, and that where dispositional mindfulness changes significantly, levels of CGS are likely to also shift in parallel.
The results of study 2 indicated that participants who reported low goal re-engagement responded to a (brief and unfamiliar) loving-kindness meditation with marked increases in conditional goal setting, while those moderate or high in goal re-engagement, and those allocated to rest or breathing meditation, showed no significant change. The finding that people low in goal re-engagement ability responded negatively on initial exposure to loving-kindness meditation is consistent with the findings of previous research. It is recognized that people differ in their initial reactions to the practice (e.g., Salzberg 2002; Johnson et al. 2009; Barnhofer et al. 2010) and that it may take time for benefits to accrue. For example, Fredrickson et al. (2008) found that differences between controls and individuals practicing loving-kindness meditation on a daily basis only began to emerge after 2 weeks. In loving-kindness practice, participants are asked to direct feelings of loving kindness towards themselves and others through the silent repetition of phrases such as “may I be happy” and “may I live with ease.” For people who find it difficult to identify new meaningful goals, initial exposure to loving-kindness meditation may only serve to increase the desire to be happy and hence the salience and importance of existing goals, without generating sufficient unconditional positive affect to enable alternative paths to fulfillment to become apparent. As such, for individuals low in goal re-engagement (those who are known to find it difficult to engage with new rewarding goals when existing goals are challenged), it may be necessary to introduce loving-kindness meditation following greater preparation in order to reduce the likelihood of negative reactions, and indeed, greater experience with loving-kindness meditation may be required by most individuals before beneficial effects on psychological functioning emerge.
The absence of an effect of a brief period of breathing meditation on conditional goal setting may relate to a number of issues, but most importantly, the fact that participants completed only a 15-min meditation session in the lab. Although previous studies using similarly brief mode of processing manipulations, regarded by some as providing an analog of meditation, have demonstrated effects on cognitive variables (e.g., Watkins and Teasdale 2001, 2004), it is likely that 15-min is simply insufficient to bring about changes of the magnitude that would be required in order to alter ratings of conditional goal setting. In order to be able to observe and reflect on one’s own attachment to particular goals or outcomes, more extended practice over weeks or indeed months may be required. However, the fact that individuals low in goal re-engagement did not respond negatively to breathing meditation does suggest that it is the nature and content of loving-kindness practice, rather than meditation per se, which produced the negative impact of loving kindness on CGS, and in this regard, retaining a comparison of two distinct forms of meditation practice, albeit over a greater number of practice sessions, may be helpful in a future study.
A number of limitations of the current studies must be considered when interpreting the results. First, the sample size employed in study 1, for comparisons between immediate and delayed treatment with MBCT, was small, and hence the study lacked power. Equally, although the overall sample size was adequate in study 2, groups defined on the basis of a goal re-engagement were again more restricted, and so the results, while they do not appear to be the result of outliers, should be interpreted cautiously. Second, in study 2, we employed a healthy sample whose levels of CGS were relatively low. This may have led to floor effects with little room for further reductions in CGS following meditation, and it would be interesting to explore the impact of both types of meditation on CGS in a sample likely to display stronger CGS tendencies at the outset. Finally, we did not explore whether the effects of meditation in study 2 were mediated by changes in state mindfulness, since the capacity to measure such changes is hampered by lack of a good measure, and indeed by consensus as to precisely what such a measure might assess.
Despite these limitations, the findings of the studies reported here are still of interest. They provide support for the proposed link between dispositional mindfulness and conditional goal setting, including evidence of both a cross-sectional association in a non-clinical sample and consistent patterns of change in the two variables in a clinical sample over a period of 3–4 months. Additionally, they add to existing evidence that initial exposure to meditation practice may have counterintuitive effects. This is important because our clinical trials have found that, when participants drop out from mindfulness-based interventions, they to do so early in treatment, after relatively little exposure to the practice of mindfulness. As meditation-based interventions becoming increasingly widespread, understanding the impact of individual differences on both initial and longer-term responses to meditation is likely to be of critical importance in improving the acceptability and effectiveness of treatment. Future research is now required to examine the relationship between initial reactions to meditation and longer-term response, as well as to clarify which forms of meditation practice are most accessible for individuals with particular cognitive profiles.