The adverse effects of stress on health and well-being have been widely established (e.g. Åkerstedt, 2006; Boyd, Bakker, Pignata, Winefield, Gillespie, & Stough, 2011; Härmä, 2006; Ronen & Baldwin, 2010; Stansfeld & Candy, 2006; Tytherleigh, Jacobs, Webb, Ricketts, & Cooper, 2007). It is therefore not surprising that numerous studies have investigated methods that can be employed to reduce stress (e.g. Richardson & Rothstein, 2008; Sanz-Vergel, Demerouti, Mayo, & Moreno-Jimenez, 2011). Among these are various meditation techniques which—although they may differ in their specific form—generally request the meditator to focus attention on one object or thought in a sustained fashion and to non-reactively monitor the content of experience from moment to moment (Cahn & Polich, 2006; Lutz, Slagter, Dunne, & Davidson, 2008; Winzelberg & Luskin, 1999). Previous research has clearly shown that meditation can reduce perceived stress and stress-related symptoms (e.g. Carmody & Baer, 2008; Chang, Palesh, Caldwell, Glasgow, Abramson, Luskin, Gill, Burke, & Koopman, 2004; Grossman, Schmidt, Niemann, & Walach, 2004; Janowiak & Hackman, 1994; Rausch, Gramling, & Auerbach, 2006). Unsurprisingly, meditation courses have become very popular and people increasingly practice and use meditation techniques to free themselves from their day-to-day hassles. This notwithstanding, previous research has not paid much attention to possible mediating and moderating mechanisms explaining meditation's effectiveness. The present study therefore aims to contribute to the knowledge on the recovery effects of meditation, by (1) addressing recovery experiences as possible mediating mechanisms that may explain its beneficial effects, and (2) investigating the meditator's intrinsic motivation as potential moderating variable affecting its effectiveness.
To this purpose we used an experimental design in which, after a stress-inducing speech preparation task, participants either participated in a 15-minute audio-taped guided imagery meditation exercise or listened to a recording of a radio interview on meditation for the same amount of time. Subjectively experienced stress and serenity (see Sonnentag, Binnewies, & Mojza, 2008a) were measured before and after meditating or listening to the radio in order to determine the effects these interventions had on recovery from stress.
REVIEW OF RELATED LITERATURE
Recovery from Stress and the Role of Meditation
In the course of a day, people may encounter demanding and stressful situations: They may perform strenuous physical labor, have an emotional conflict with their colleagues, and perform cognitively demanding tasks, such as preparing a presentation they have to deliver at an important meeting with their boss. Engagement in these stressful and/or demanding situations requires effort and draws on individuals' resources, resulting in, for example, negative affective states and stress-related physiological responses (e.g. Meijman & Mulder, 1998; Hobfoll, 1998). When stressors are no longer present or when stressful events are resolved, the psychophysiological systems that were activated during the stressful situation will return to their pre-demand levels (Meijman & Mulder, 1998; Geurts & Sonnentag, 2006), a process that is called “recovery”. For recovery to take place, it is essential that no further demands be placed on the specific psychological (e.g. performance anxiety) and physiological (e.g. neuroendocrine and cardiovascular) systems that were activated during the stressful situation. For example, people who suffered from a stressful conflict with their boss about their salary will not recover when following the conflict further demands are placed on the same system (e.g. when they have a heated debate with their spouse about how to raise the kids); however, recovery will occur if other systems are activated (e.g. physical activities, such as running), or if people engage in activities with little or no demands (e.g. relaxation, such as enjoying the sauna). Recovery is vital for the protection of health and well-being, because an accumulated lack of recovery will—due to prolonged activation of bodily stress systems (McEwen, 1998)—eventually result in manifest physical and mental impairment, such as burnout (Van Hooff, Geurts, Taris, Kompier, Dikkers, Houtman, & Van den Heuvel, 2005) and even cardiovascular death (Kivimäki, Leino-Arjas, Kaila-Kangas, Lukkonen, Vahtera, Elovainio, Härmä, & Kirjonen, 2006).
Previous research has revealed a number of activities and conditions that are beneficial to recovery from stress. These include sleep quality (e.g. Shirom, Nirel, & Vinokur, 2010), sports (e.g. Sonnentag & Jelden, 2009), low-effort activities (e.g. Sonnentag, 2001), and meditation (e.g. Chang et al., 2004), on which we focus in the current paper. Meditation may facilitate recovery, because it has been shown to evoke a “relaxation response” (Benson, 1976; Morse, Martin, Furst, & Dubin, 1977), which can be seen as the physiological opposite of stress, and that is characterised by a “wakeful hypometabolic physiological state and a balance between parasympathetic and sympathetic functions” (Perez-De-Albeniz & Holmes, 2000, p. 55). Previous research has indeed shown that meditation is effective in reducing stress and stress-related symptoms and is beneficial to recovery from stress. Based on a meta-analysis, Grossman et al. (2004) concluded that an 8- to 10-week mindfulness meditation course leads to enhanced coping with and recovery from distress within both clinical and non-clinical populations. Furthermore, Richardson and Rothstein (2008) showed meditation to be effective in reducing psychological indicators of stress in their meta-analysis of the effectiveness of work-related stress management programs.
In most of the studies that were included in these meta-analyses, participants took part in multiple meditation sessions to examine their effectiveness on recovery from stressful events. However, Rausch et al. (2006) also showed that a single meditation session that participants could engage in without previous practice was sufficient to reduce cognitive, somatic, and general state anxiety. Because we were interested in possible interventions that would lead to recovery from stress, but that are also short in duration and relatively easy to do (and therefore relatively easy to implement at home or in the workplace), in the present study we examined the effects of such a short single meditation session.
The meditation exercise in our study was based on the principles of guided imagery. Guided imagery is “a technique that utilizes stories or narratives to influence the images and patterns that the mind creates” (Hart, 2008, p. 295) and uses an individual's capacity to create mental images that bring about pleasant thoughts, positive emotions, and physical effects (e.g. reduction in stress, anxiety, or pain; Hart, 2008). Guided imagery techniques share a number of core features (e.g. Hart, 2008; Lutz et al., 2008). They ask meditators to calm the mind and sustain attention moment by moment on the (usually pleasant) images, objects, or scenes that are presented in the story. To sustain this focus, meditators are asked to non-reactively monitor the content of experience and attention. Meditators are also instructed that in case attention wanders away, they should recognise the wandering and restore attention. Guided imagery can be done by a live practitioner, an audiotape or by participants themselves (Van Kuiken, 2004). In the present study, using guided imagery meditation, the audio-taped voice of a meditation instructor guides participants through the session by telling them the specific images to create in their mind.
The beneficial effects of guided imagery on well-being and stress have already been demonstrated. For example, Tsai and Crockett (1993) showed that it leads to reduced levels of self-reported work-related stress and increases in self-reported pschophysiological health among Chinese nurses. In her meta-analysis of the effects of guided imagery in mainly clinical populations, Van Kuiken (2004) found it to be positively related to physiological and psychological indicators of health and well-being.
To investigate its effectiveness for recovery from stress in this study, the effects of the guided imagery meditation exercise were compared to those of a control condition, in which participants listened to a radio interview about meditation. Recovery is often operationalised by means of an increase in subjective well-being (Demerouti, Bakker, Geurts, & Taris, 2009). We also followed this approach in the present study, by including two subjective recovery indicators. First, because in the present study we aimed to experimentally induce elevated stress levels, we included (decreased) subjectively experienced stress as a proximal indicator of recovery. Second, as recovery results in restoration of decreased mood, we followed Sonnentag et al. (2008a) and included (increased) serenity as indicator of positive affect in this study. Serenity refers to a state of being calm, relaxed, and at ease and is characterised by positive feelings and low arousal (Sonnentag et al., 2008a; Watson, Clark, & Tellegen, 1988).
Both participating in the meditation exercise and listening to the radio interview can be considered “low-effort activities”, i.e. activities that do not place many demands on an individual's psycho(physio)logical and physical systems (see Sonnentag, 2001). These kinds of activities will facilitate participants' recovery from stress, because no further demands are placed on their psychological (e.g. performance anxiety) and physiological (e.g. neuroendocrine and cardiovascular) systems that were activated during the stressful situation, which allows these systems to return to their baseline levels (Geurts & Sonnentag, 2006; Meijman & Mulder, 1998). Therefore, we expect subjectively reported stress to decrease and serenity to increase both among participants who meditate and among participants who listen to the radio interview. We nonetheless expect these effects to be stronger in the former group of participants because—as opposed to participants listening to the radio interview—they are explicitly instructed to focus on releasing tension and on generating pleasant thoughts. Also, meditation has been shown to evoke a “relaxation response” (Benson, 1976) that will reduce experienced stress. Thus:
Hypothesis 1: Both participation in the meditation exercise and listening to the radio interview are positively associated with recovery from stress (i.e. decrease in subjectively reported stress and increase in serenity). These effects will be stronger for the meditation exercise.
Recovery Experiences as Mediators
It has been shown that on a physiological level, the “relaxation response” (Benson, 1976; see also Morse et al., 1977) plays a pivotal role in explaining meditation's effectiveness. Nevertheless, in several studies in which meditation did lead to decreases in subjectively reported stress, no effects were obtained for physiological indicators, such as decreased heart rate or lower levels of skin conductance (Eifert & Heffner, 2003; Erisman & Roemer, 2010). This suggests that physiological mechanisms are not sufficient to explain meditation's effectiveness and that other, psychological, mechanisms (also) play a role. Indeed, previous research (e.g. Sonnentag & Fritz, 2007; Van Hooff, Geurts, Beckers, & Kompier, 2011) has already shown that, to recover from stress, it is not only important which activities one engages in, but also how these activities are subjectively experienced. To our knowledge, however, no previous studies have addressed possible psychological mediating mechanisms. Uncovering these psychological mechanisms was another goal of the current study.
Sonnentag and Fritz (2007) distinguished between four types of experience (detachment, relaxation, mastery, and control) that are supposed to be positively related to recovery from stress, an assumption that has received substantial support in previous research (e.g. Sonnentag & Bayer, 2005; Sonnentag & Fritz, 2007; Von Thiele Schwarz, 2011). Detachment refers to the feeling of being away from the stressful situation (e.g, Sonnentag & Bayer, 2005; Sonnentag, Mojza, Binnewies, & Scholl, 2008b) and facilitates recovery because it implies that the person is no longer engaged in or thinking about the stressful situation (Sonnentag & Fritz, 2007). Relaxation, the second type of recovery experience, is characterised by a process in which activation and arousal decreases and positive affect increases (Stone, Kennedy-Moore, & Neale, 1995), and is presumed to facilitate recovery because the person experiences low demands (Meijman & Mulder, 1998) and because positive affective states help to replenish resources and cope with negative affective states, such as stress (Fredrickson, Mancuso, Branigan, & Tugade, 2000).
The explanation underlying the positive effects of mastery (i.e. “challenging . . . experiences that provide opportunities for learning and success”; Sonnentag et al., 2008a, p. 675) is twofold. First, experiencing mastery will lead to an increase in positive emotions (Parkinson & Totterdell, 1999), which are associated with the production of certain hormones in the brain's “pleasure reward” system (e.g. serotonin, dopamine) that may quickly down-regulate the stress response, thus promoting recovery (Esch & Stefano, 2004). Second, the recovery potential of mastery experiences can be explained on the basis of the Conservation of Resources Model (Hobfoll, 1998). According to this theory, stress implies a loss of resources (i.e. “objects, personal characteristics, and energies that are either themselves valued for survival, directly or indirectly, or that serve as a means of achieving these resources”; Hobfoll, 1998, p. 45), whereas recovery from stress will be facilitated by gaining new resources. As mastery experiences will help people to build new resources, such as skills, competencies, and self-efficacy (Bandura, 1997; Hobfoll, 1998), these will therefore facilitate recovery from stress. The fourth type of recovery experience, control (i.e. the degree to which people have the feeling that they can decide which activities to pursue and when and how to pursue them; Sonnentag & Fritz, 2007; Sonnentag et al., 2008a), is assumed to contribute to recovery because it may increase resources such as self-efficacy and feelings of competence. Furthermore, having control during leisure time allows people to choose the activities they like most, which may be beneficial for the recovery process (Sonnentag & Fritz, 2007; Van Hooff et al., 2011).1
We expect that—compared to those who listen to the radio interview—participants who meditate experience higher levels of detachment. That is, during the meditation session, participants are guided to disregard distracting thoughts (Winzelberg & Luskin, 1999) and to focus their attention on following the instructions given by the meditation trainer. This leaves no room for thoughts about the stressful speech preparation task. Participants in the radio condition, on the other hand, are not explicitly instructed to focus their attention, which makes it possible that they still worry about the presentation task. Furthermore, because participants doing the meditation exercise are explicitly guided to become relaxed and those listening to the radio program are not, we expect the experience of relaxation to be higher in the meditation condition as well.
Finally, for two reasons, we also expect that participants who meditate experience higher levels of mastery than those listening to the radio interview. First, in the meditation condition, participants are actively involved and will notice that they are capable of influencing their own subjectively experienced stress levels. This skill will not be experienced by passively listening to a radio interview about meditation. Second, because it is likely that at least some of the participants will have no or only limited previous experience with guided imagery meditation, they will consider this as a new learning opportunity that absorbs them. Participants in the radio condition may hear new information about meditation, but, as they are not as absorbed in the activity as those that are actively meditating, this will not (or will to a lesser extent) result in feelings of mastery.
Because of (1) these expected differences in recovery experiences between the two conditions, (2) our prediction that the meditation and radio condition are differentially related to recovery (see Hypothesis 1), (3) the fact that previous research has shown detachment, mastery, and relaxation to be positively related to recovery (Sonnentag & Fritz, 2007), all conditions for a mediating effect of recovery experiences are met (cf. Baron & Kenny, 1986), which leads to the following predictions:
Hypothesis 2a: Detachment will partially mediate the relationship between type of stress-reducing activity (meditation vs. radio) and recovery (i.e. reduction in subjectively reported stress levels, increase in serenity).
Hypothesis 2b: Mastery will partially mediate the relationship between type of stress-reducing activity (meditation vs. radio) and recovery (i.e. reduction in subjectively reported stress levels, increase in serenity).
Hypothesis 2c: Relaxation will partially mediate the relationship between type of stress-reducing activity (meditation vs. radio) and recovery (i.e. reduction in subjectively reported stress levels, increase in serenity).
Intrinsic Motivation as a Moderator
Generally, for a stress intervention to be effective, it is not only important that the theoretical effectiveness of the technique employed be established, but also that it be implemented and used to its full potential. As Kristensen (2005, p. 207) stated: “it does not help that the pill has effect if the patient does not take it”. One factor affecting whether or not the participants “take the pill” may be their intrinsic motivation for the specific stress-reducing activity.
Intrinsically motivating activities are activities that “individuals find interesting and would do in the absence of operationally separable consequences” (Deci & Ryan, 2000, p. 233). If people are intrinsically motivated to perform an activity, they do so simply because they consider the activity interesting and pleasurable (Ryan & Deci, 2000)—as a consequence, people are more engaged in the task and show stronger involvement (Deci & Ryan, 2000). Intrinsic motivation can be assessed by levels of interest and enjoyment (Ryan & Deci, 2000), implying that people who are more intrinsically motivated for an activity will experience more pleasure while engaged in this activity than those who are less intrinsically motivated. Previous research (Esch & Stefano, 2004) has shown that the experience of pleasure is associated with the production of certain hormones in the brain's “pleasure reward” system (e.g. serotonin, dopamine) that will down-regulate the stress response. Also, Fredrickson's (2001) Broaden-and-Build theory underlines the beneficial effects of positive emotions—such as pleasure (Parkinson, 1995; Parrot, 2001; Warr, 2005)—by stating that these can undo the effects of negative emotions such as stress. This undoing effect of positive emotions has indeed been confirmed in experimental studies (Fredrickson et al., 2000). Due to the strong connection between intrinsic motivation and pleasure, we expect participants who are highly intrinsically motivated to recover more than participants who are less intrinsically motivated.
However, we do expect intrinsic motivation to a play a greater role in the meditation condition than in the radio condition. As stated, during the meditation exercise (but not during listening to the radio interview), participants are specifically instructed to control attention, relax, release tension, and to generate pleasant thoughts. Because these elements are strongly associated with recovery experiences, such as detachment and relaxation, we expect that intrinsically motivated people who are more engaged in the meditation exercise benefit more from these experiences, and recover more quickly. Therefore, we expect this moderating effect of intrinsic motivation to be stronger for participants who meditate than for participants who listen to the radio interview.
Hypothesis 3: There will be a stronger decrease in subjectively experienced stress and increase in serenity for those participants who are highly intrinsically motivated compared to those who are less intrinsically motivated and this effect will be stronger in the meditation condition.
THE PRESENT STUDY: OVERVIEW AND HYPOTHESES
Meditation has been shown to lead to recovery from stress, but little is known about the underlying psychological mechanisms. To contribute to the knowledge on the recovery effects of meditation, we (1) invoked recovery experiences as possible mediating mechanisms that may explain its beneficial effects, and (2) examined the meditator's intrinsic motivation as potential moderating variable affecting its effectiveness. To this purpose we used an experimental design in which, after a stress-inducing speech preparation task, participants either engaged in a 15-minute audio-taped guided imagery meditation exercise or listened to a recording of a radio interview on meditation for the same amount of time. Subjectively experienced stress and serenity were measured before and after meditating or listening to the radio. We tested the predictions (1) that both participation in the meditation exercise and listening to the radio interview would lead to a decrease in subjectively reported stress and increase in serenity, but that effects would be more pronounced in the meditation condition (Hypothesis 1); (2) that the recovery experiences detachment (Hypothesis 2a), mastery (Hypothesis 2b), and relaxation (Hypothesis 2c) each would partially mediate this effect; and (3) that especially in the meditation condition, there would be a stronger decrease in subjectively experienced stress and increase in serenity for those participants who are highly intrinsically motivated compared to those who are less intrinsically motivated (Hypothesis 3).
Design and Participants
One hundred students at the University of Amsterdam took part in this study in exchange for 7 Euros or partial credit toward fulfilling a course requirement. All participants read and signed an informed consent form before starting the experiment. They were randomly assigned to one of two stress-reducing activities (guided meditation vs. radio fragment). Dependent variables were experienced stress and serenity, recovery experiences, and intrinsic motivation. Data from one participant were excluded as (s)he did not complete all stress-related questionnaires. Therefore, the final sample comprises 99 participants (30 male, Mage= 21.99; SDage= 5.38).
Subjective Units of Distress. Participants were asked to evaluate their immediately experienced stress on a subjective units of distress (SUD) scale, ranging from 0 (state of perfect calm) to 100 (worst stress ever experienced). Three additional items were included to allow participants to practice evaluating their SUD in response to situations that involve varying levels of stress (see Ponce, Lorber, Paul, Esterlis, Barzvi, Allen, & Pescatello, 2008), that is, (a) their average stress on a day at home, (b) their average stress on a day at work/university, and (c) their average stress scrambling to meet a deadline. This 100-point scaling has been widely used in previous research to measure distress and anxiety (e.g. Ironson, Freund, Strauss, & Williams, 2002; Ponce et al., 2008; Schmidt & Zvolensky, 2007).
Serenity was measured using one item from the PANAS-X (Watson & Clark, 1994). This is the extended version of the PANAS (Watson et al., 1988), which is a reliable and valid instrument to measure positive and negative affect (Egloff, Schmukle, Burns, Kohlmann, & Hock, 2003): participants were asked to indicate on a 5-point Likert scale (1 = not at all, 5 = extremely) how relaxed they currently felt.
Recovery experiences were measured with an adapted version of the Recovery Experiences Questionnaire (Geurts, De Bloom, Kompier, & Sonnentag, 2009; Sonnentag & Fritz, 2007). This is a reliable and valid questionnaire to measure the recovery experiences detachment, relaxation, mastery, and control. As the items in the original version of the instrument refer in general to the time people spend after work, items were reformulated so that they explicitly referred to the activity that was pursued in the experiment, and some items were excluded because they could not be meaningfully reformulated. Detachment (α= .86) was measured by means of four items (e.g. “During the meditation exercise, I forgot about the presentation assignment”) and mastery (α= .75; e.g. “By engaging in the meditation exercise, I learned something new”) and relaxation (α= .89; e.g. “During the meditation exercise, I used the time to relax”) were each measured by means of three items. All items were rated on a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree).
Intrinsic motivation was measured using the seven-item interest/enjoyment subscale of the Intrinsic Motivation Inventory (e.g. Ryan, 1982; Ryan & Deci, 2000). This is a multidimensional questionnaire that is intended to assess participants' subjective experience related to a target activity in laboratory experiments. It has been used in several experiments related to intrinsic motivation and self-regulation (e.g. Ryan, 1982; Deci, Eghrari, Patrick, & Leone, 1994). The original instrument assesses participants' interest/enjoyment, perceived competence, effort, value/usefulness, felt pressure and tension, and perceived choice while performing a given activity, thus yielding six subscale scores, of which we only used the interest/enjoyment scale. Items were reformulated to fit the specific stress-reducing activities of this experiment, e.g. “I enjoyed doing the meditation exercise/listening to the radio interview very much”. Items were rated on a 7-point Likert scale (1 = not at all true, 7 = very true; α= .93).
Participants came to the laboratory, read and signed an informed consent form, and were seated in individual cubicles that were equipped with a computer that displayed all materials. During the course of the experiment, each participant was alone in their cubicle. After having provided some demographic information, participants completed the first set of SUD (including the practice questions) and serenity ratings, and continued with a speech preparation task to induce stress.
Stress induction. Stress was induced by means of a speech preparation task, which is commonly employed for that purpose (cf. Egloff, Wilhelm, Neubauer, Mauss, & Gross, 2002; Fredrickson et al., 2000; Kaiser, Hinton, Krohne, Stewart, & Burton, 1995; Phillips & Giancola, 2008; Sher, Bartholow, Erickson, Peuser, & Wood, 2007). Participants were told that they had 3 minutes to prepare a speech about what they like and don't like about their body and physical appearance (Philips & Giancola, 2008). They were also told that their speech would be evaluated by experts. During the 3-minute preparation time, they saw a timer counting down from 180 to 0. At the end of the preparation time, participants were told that they had a 50 per cent chance that the computer would choose them to give their presentation and were requested to complete the SUD (without the practice questions) and serenity questions while the computer decided.
All participants were then informed that they were not chosen to give their speech, but instead would do another activity that lasted for 15 minutes. In the radio condition, they learned that they would listen to a radio interview on meditation, were instructed to put on their headphones, and spent the next 15 minutes listening to this radio fragment. In the meditation exercise condition, they learned that they would participate in a meditation exercise, were instructed to put on their headphones, and spent the following 15 minutes doing the guided imagery meditation exercise. The meditation exercise was recorded beforehand by a professional meditation trainer. Part of this exercise included instructions to release muscle tension, to let go of any thoughts and to imagine oneself on a tropical beach. In the radio interview condition participants listened to a recording of a radio interview about the effects of meditation performed by the same trainer. After the meditation exercise/radio interview, participants provided the second set of SUD and serenity ratings, and completed the recovery experience questionnaire and intrinsic motivation scale. Following these questionnaires, participants were debriefed and dismissed.
Table 1 presents means, standard deviations, and correlations of the variables under study.
Table 1. Means, Standard Deviations, and Correlations of the Variables under Study
| 1. Gender (1 = male, 2 = female)||1.70||0.46||1.00|| || || || || || || || || || || |
| 2. Age||21.99||5.39||−.08||1.00|| || || || || || || || || || |
| 3. SUD t11||23.58||20.01||−.14||.15||1.00|| || || || || || || || || |
| 4. SUD t2||46.98||23.66||.08||−.26*||.20*||1.00|| || || || || || || || |
| 5. SUD t3||15.57||15.73||−.06||−.19||.39**||.38**||1.00|| || || || || || || |
| 6. Serenity t1||3.83||0.85||.07||−.21*||−.46**||−.06||−.17||1.00|| || || || || || |
| 7. Serenity t2||2.85||0.98||.12||.01||−.19||−.61**||−.25*||.30**||1.00|| || || || || |
| 8. Serenity t3||4.04||0.81||.17||.14||−.03||−.16||−.62**||.29**||.32**||1.00|| || || || |
| 9. Intrinsic motivation||4.42||1.44||.14||.01||.05||−.05||−.38**||.14||.10||.49**||1.00|| || || |
|10. Detachment||4.22||0.83||.28**||.14||.13||.06||−.21*||−.04||.07||.23*||.41**||1.00|| || |
|11. Mastery||2.54||0.97||.19||.15||.06||.04||−.20*||.00||−.08||.28**||.61**||.20*||1.00|| |
First, we examined possible gender and age differences between the participants who meditated and those who listened to the radio interview by means of a X2-test and an independent sample t-test respectively. These revealed no differences between the two groups (gender: X2(1) = 0.32, ns; age: t(97) =−0.58, ns). Second, we examined the validity of the SUD measure by examining whether participants rated their stress levels while struggling for a deadline as higher than their stress levels at home or at work/university. A paired-sample t-test indicated that this indeed turned out to be the case (Mhome= 28.33 vs. Mdeadline= 70.94: t(98) =−21.50, p < .01; Mwork/university= 40.45 vs. Mdeadline= 70.94: t(98) =−18.77, p < .01).
Third, because of a possible conceptual overlap between the items used to measure serenity and the recovery experience “relaxation”, we conducted a confirmatory factor analysis to examine whether these were indeed empirically distinct constructs. Two models were tested against each other. In the first model the items measuring relaxation and the serenity measures after the speech preparation task and after the meditation/radio session loaded on one factor. This model fit the data reasonably well (χ2= 12.26, df= 5, RMSEA = .12, CFI =.97, NNFI = .94). In the second model, two factors were specified, one for relaxation and one for serenity. As this model provided a significantly better fit than the first one (χ2= 4.72, df= 4, RMSEA = .03, CFI = 1.00, NNFI = .99; Δχ2= 7.54, df= 1, p < .01), we concluded that the recovery experience relaxation and serenity can be considered two different constructs.
Fourth and finally, to examine whether our stress manipulation indeed evoked stress, we investigated differences between the SUD and serenity measures before and after the speech preparation task. Paired-sample t-tests revealed that participants reported higher levels of stress (as indicated by the SUD) and lower levels of serenity after the speech preparation task (MSUD= 46.98 and Mserenity= 2.85) than before that task (MSUD= 21.99 and Mserenity= 3.83; SUD: t(98) =−8.40, p < .01; serenity: t(98) = 8.96, p < .01).
Testing of Hypotheses
Hypothesis 1: Recovery Effects of Meditation. To investigate differences in stress recovery between those who meditated and those who listened to the radio interview, we submitted our dependent variables, subjectively experienced stress and serenity, to separate repeated measures ANOVAs with time (two levels, i.e. after the speech preparation task and after meditation/radio interview) as within-subjects factor and type of stress-reducing activity (two levels, i.e. meditation vs. radio interview) as between-subjects factor. For serenity, the results revealed a significant main effect of type of activity (F(1, 97) = 4.76, p < .05, ηp2= .05). Participants in the meditation condition generally reported higher levels of serenity (M= 3.58) than those in the radio condition (M= 3.26). We also found a significant main effect of time (F(1, 97) = 122.19, p < .01, ηp2= .56), which showed that levels of serenity were higher (M= 4.04) after the meditation/radio session than after the speech preparation task (M= 2.85). Finally, we found our predicted interaction among type of stress-reducing activity and time (F(1, 97) = 7.84, p < .01, ηp2= .08). To examine this interaction in more detail, we computed a difference score for serenity by subtracting the measure after the speech preparation task from the measure after the meditation/radio session. An independent sample t-test revealed that participants who did the meditation exercise showed a stronger increase in serenity (M= 1.44) than those who listened to the radio interview (M= 0.86; t(97) = 2.80, p < .01). It should be noted, however, that simple effects analyses for each stress-reducing activity showed serenity levels to increase in both conditions (radio: F(1, 97) = 29.58, p < .01; η2= .23; meditation: F(1, 97) = 113.09, p < .01; η2= .54).
Regarding subjectively experienced stress, the results revealed no significant main effect of type of activity (F(1, 97) = 0.01, ns) but did uncover a significant main effect of time (F(1, 97) = 180.23, p < .01, ηp2= .65), showing levels of subjectively experienced stress to be lower (M= 15.57) after the meditation/radio session than after the speech preparation task (M= 46.98). We also found our predicted interaction among type of stress-reducing activity and time (F(1, 97) = 4.44, p < .05, ηp2= .04). Again, we computed a difference score for the two measures and an independent sample t-test revealed that participants in the meditation group showed a stronger decrease in SUD (M=−35.50) than those in the radio group (M=−25.86; t(97) = 2.11, p < .05). However, subjectively experienced stress levels decreased significantly in both conditions (radio: F(1, 97) = 55.62, p < .01; η2= .36; meditation: F(1, 97) = 142.17, p < .01; η2= .59).
Thus, supporting Hypothesis 1, these results show stronger recovery from stress following meditation as compared to listening to the radio interview.
Hypothesis 2: The Mediating Effect of Recovery Experiences. To test for mediation of the effect of type of stress-reducing activity on our recovery indicators by recovery experiences, we computed a series of analyses on the lines of the assumptions set forth by Baron and Kenny (1986). Pertaining to the present study, these assumptions are: (1) type of stress-reducing activity should be related to recovery experiences, (2) recovery experiences should be related to recovery, (3) type of stress-reducing activity should be related to recovery, and (4) the relationship between type of stress-reducing activity and recovery should become smaller (partial mediation) after controlling for the recovery experience.
Regarding assumption 1, independent sample t-tests revealed no differences between the two stress-reducing activities with respect to detachment (t(97) = 0.17, ns), but showed that participants in the meditation condition reported higher levels of relaxation (t(97) = 2.61, p < .05; Mmeditation= 4.33, Mradio= 3.92) and mastery (t(97) = 2.16, p < .05; Mmeditation= 2.72, Mradio= 2.30). Given the prerequisites for mediation, these results imply that detachment cannot act as a mediator in the relationship between condition and recovery. For this reason, the other assumptions for mediation are only tested for relaxation and mastery.
To examine the relationship between recovery experiences and the recovery indicators (assumption 2), for each combination of recovery experience (predictor: relaxation or mastery) and recovery indicator (criterion variable: subjectively experienced stress or serenity) a regression analysis was conducted. Because we were interested in changes in recovery indicators due to the meditation exercise or listening to the radio interview, the subjectively experienced serenity or stress measures as completed after the stress task were included as control variables. Table 2 shows that relaxation and mastery are significantly negatively related to subjectively experienced stress and positively related to serenity, thus supporting assumption 2.
Table 2. Regression Analyses Examining the Associations between Condition, Recovery Experiences, and Recovery Indicators
|Serenity/SUD after speech preparation1||.09||.30||<.01||.15||.39||<.01|
|Serenity/SUD after speech preparation1||.24||.31||<.01||.19||.40||<.01|
|Condition|| ||.37||<.01|| ||−.20||<.05|
|Serenity/SUD after speech preparation1||.44||.24||<.01||.32||.36||<.01|
|Condition|| ||.25||<.01|| ||−.10|| ns |
|Relaxation|| ||.48||<.01|| ||−.37||<.01|
|Serenity/SUD after speech preparation1||.29||.33||<.01||.22||.40||<.01|
|Condition|| ||.32||<.01|| ||−.16|| ns |
|Mastery|| ||.24||<.01|| ||−.18||<.01|
Assumptions 3 and 4 were examined by computing a series of hierarchical regression models for each recovery indicator. Again, the subjectively experienced serenity or stress measures as completed after the stress task were included as control variables in the first step. In the second step, we added type of stress-reducing activity (0 = radio, 1 = meditation) as predictor to examine if this was related to recovery (assumption 3). Finally, in the third step we included relaxation or mastery to see whether this would influence the strength of the relationship between type of activity and the recovery indicators (assumption 4). Table 2 shows type of stress-reducing activity to be significantly positively related to serenity and negatively to subjectively experienced stress (assumption 3 supported). With respect to both serenity and subjectively experienced stress, Table 2 also shows that the effects of type of activity become smaller, but not zero, after controlling for relaxation or mastery. This implies that recovery experiences may play a partial mediating role. To examine whether this partial mediation effect was indeed significant, we employed the bootstrapping methodology as recommended by Preacher and Hayes (2008). Results of these analyses using 1,000 bootstrapping samples showed both the indirect effects of condition on serenity via relaxation (bootstrapping estimate = .1225; 95 per cent CI between .0288 and .2737) and mastery (bootstrapping estimate = .0514; 95 per cent CI between .0060 and .1396) to be significant. A similar pattern of results was observed for subjectively experienced stress (relaxation: bootstrapping estimate =−.1035; 95 per cent CI between −.2912 and −.0288; mastery: bootstrapping estimate =−.0393; 95 per cent CI between −.1165 and −.0031).
Altogether, detachment turned out not to act as a mediator in the relationship between condition and recovery (inconsistent with Hypothesis 2a), but the recovery experiences mastery (see Hypothesis 2b) and relaxation (see Hypothesis 2c) did.
Hypothesis 3: The Moderating Effect of Intrinsic Motivation. To examine whether intrinsic motivation affected the recovery potential of the meditation exercise and the radio interview, we first created two subgroups for intrinsic motivation, based on the lowest (≤3.71, i.e. “low-group”) and highest (≥5.29, i.e. “high-group”) tertiles of the intrinsic motivation scale. For each recovery indicator, we then conducted a repeated measures ANOVA with time (two levels, i.e. after the speech preparation task and after the stress-reducing activity) as within-subjects factor and intrinsic motivation subgroup (two levels, i.e. high and low) as between-subjects factor.
For serenity, in the radio condition there was no significant main effect of group (F(1, 25) = 3.15, ns), but there was a significant main effect of time (F(1, 25) = 13.93, p < .01, ηp2= .36), which showed serenity to increase from M= 2.85 before the radio interview to M= 3.74 afterwards. Finally, there was no significant group by time interaction (F(1, 25) = 0.15, ns).
In the meditation condition, there was no significant main effect of group (F(1, 31) = 1.68, ns), but a significant main effect of time (F(1, 31) = 86.31, p < .01, ηp2= .74; serenity increased from M= 2.82 before the meditation exercise to M= 4.36 afterwards) and a significant time by group interaction were observed (F(1, 31) = 8.31, p < .01, ηp2= .21). An independent sample t-test showed that participants in the high-group reported a stronger increase in serenity (M= 1.90) compared to participants in the low-group (M= 1.00; t(31) =−2.88, p < .01).
For subjectively experienced stress, both in the radio condition and in the meditation condition there were no significant main effects of group (radio: F(1, 25) = 1.33, ns; meditation: F(1, 31) = 0.47, ns), significant main effects of time (radio: F(1, 25) = 24.92, p < .01, ηp2= .50; meditation: F(1, 31) = 115.71, p < .01, ηp2= .79), but no significant group by time interactions (radio: F(1, 25) = 1.84, ns; meditation: F(1, 31) = 2.50, ns).
As intrinsic motivation positively affected the increase in serenity for the meditation condition, but not for the radio condition, these results provide partial support for Hypothesis 3.
The present study aimed to advance our insight into the stress-reducing qualities of meditation, by comparing the effects on stress recovery of a 15-minute session of guided imagery meditation to a control condition in which participants listened to a 15-minute radio interview on meditation. Additionally, we aimed to shed light on the mediating role of recovery experiences and the moderating role of intrinsic motivation in relation to meditation's effectiveness. Results showed that both practicing meditation and listening to the radio interview resulted in a significant decrease in subjectively experienced stress and a significant increase in serenity. However, and supporting Hypothesis 1, these effects were even more pronounced in the meditation condition.
More importantly, by demonstrating that the differences in recovery between the meditation and radio conditions could be partly explained by higher levels of relaxation and mastery experienced while meditating (see Hypotheses 2b and 2c), we were also able to provide more insight into the psychological mechanisms underlying the effectiveness of meditation. Although previous work (e.g. Sonnentag & Bayer, 2005) has shown the importance of detachment in stress recovery, we did not find support for a (partial) mediating effect of detachment. This may be due to the fact that—before starting the meditation or radio session—all participants were informed that they would not actually have to give their presentation, which suggests that they could have detached from the presentation task before they started meditating or listening to the radio interview. Because of these already high levels of detachment at the start of the meditation or radio session, engagement in these activities could not contribute differently to participants' levels of detachment (which is one of the requirements for mediation; Baron & Kenny, 1986). Our data indeed show that both in the meditation and in the radio condition levels of detachment were rather high (Mmeditation= 4.24, Mradio= 4.21 on a 5-point scale).
Finally, this study also showed that the effectiveness of meditation is affected by the extent to which participants are intrinsically motivated to meditate (see Hypothesis 3). Meditation was successful in increasing serenity for participants who are not intrinsically motivated, but effects were stronger for those who were highly intrinsically motivated, and, thus, considered meditating to be a pleasant activity. Although a similar effect for the expected decrease in subjectively experienced stress did not reach significance, it should be noted that results were in the hypothesised direction (decrease in high-group: M= 42.41, decrease in low-group: M= 31.54).
We believe that four points of concern regarding the current study should be discussed. First of all, the present study relied exclusively on self-report data to measure participants' (recovery from) stress, recovery experiences, and intrinsic motivation. Although stress, recovery experiences, and intrinsic motivation are ultimately defined as a subjective experience (Quick, Quick, Nelson, & Hurrell, 1997; Sonnentag & Fritz, 2007) and self-reports provide valid and useful information (Kompier, 2005), it is often argued that exclusive reliance on self-reports would result in an overestimation of the associations among variables due to common method variance. Spector (2006), however, thoroughly studied the potential problem of common method variance and mentions three factors that mitigate the problems associated with using self-report measures: (i) using self-reports does not guarantee finding significant results; (ii) potential biasing variables (e.g. negative affectivity, social desirability) do not generally inflate correlations among study variables; and (iii) mono-method correlations are not necessarily higher than multi-method correlations (also see Crampton & Wagner, 1994). Therefore, Spector (2006) concludes that “the popular position suggesting that common method variance automatically affects variables measured with the same method is a distortion and oversimplification of the true state of affairs” (p. 221). Based on these considerations, we cautiously conclude that common method variance did not seriously bias our results. However, we do note that supplementing these measures with physiological stress indicators, such as cortisol levels or blood pressure, may provide a more complete picture of the beneficial effects of meditation and the role that recovery experiences and intrinsic motivation play. Therefore we recommend future studies to include such measures as well.
A second point refers to the fact that serenity and subjectively experienced stress were each measured with only one item, which may raise questions regarding the validity of these measures. We nonetheless think that there are at least two arguments that support the use of these single-item measures. First, it should be noted that the validity of other single-item stress-related variables, such as fatigue and emotional exhaustion, have been established in previous studies (Van Hooff, Geurts, Taris, & Kompier, 2007; Rohland, Kruse, & Rohrer, 2004). Furthermore, the validity of a single-item SUD measure has already been shown in many previous studies (e.g. Ironson et al., 2002; Ponce et al., 2008; Schmidt & Zvolensky, 2007). Indeed, in the current study, we found that participants rated their stress levels at home/work as significantly lower than when struggling for a deadline. Second, we considered it important to employ the shortest possible stress measures in the present study. That is, on the one hand we wanted to prevent the recovery process starting while participants were still completing questionnaires, which would contaminate our manipulation of recovery activities (radio vs. meditation). On the other hand, completing relatively long questionnaires would be an additional burden for our participants, leading to an increase in stress levels independent of the speech preparation task, whereas single-item measures may evoke less participant boredom, fatigue, and frustration (Robins, Hendin, & Trzesniewski, 2001).
A third point of concern refers to the causal interpretation of the effect of intrinsic motivation. As this construct was measured after the meditation exercise, it is not possible to make causal inferences regarding the relationship between intrinsic motivation and recovery. However, as we did not know in advance how many participants would have experience with meditation (our meditation exercise was intended for novices and meditation practioners alike), it was not sensible to ask participants about their intrinsic motivation before they started the meditation exercise. Thus, reports of intrinsic motivation for an activity participants have not yet pursued would suffer from a serious lack of validity. Our data indeed showed that only 46 per cent of our participants had previous experience with one or more forms of meditation. This issue could be prevented in future research by selecting participants who have at least a little experience with meditation.
A final issue relates to the duration of the meditation exercise in the present study. Even though we found recovery to take place during a 15-minute session, it is not known whether this is indeed the optimal time interval (see Williams, 2010). It may well be that the main part of the recovery process already took place during the first part of the session, and that the remainder of the exercise was in fact redundant with respect to increasing recovery. The opposite may also be true, i.e. a longer meditation session may have resulted in significantly lower levels of subjectively experienced stress and higher serenity. Previous research (Ponce et al., 2008) nonetheless showed 15 minutes of progressive muscle relaxation (PMR) to be as effective as 25 minutes, whereas a 7-minute session was less effective than this latter one. Although this finding cannot be extrapolated directly to meditation, as PMR is considered to be a physical relaxation technique, whereas meditation is considered to be more cognitive in nature (Murphy, 2003), it underlines the potential of employing a 15-minute exercise. This notwithstanding, future research may further examine the optimal duration of a session of guided imagery meditation.
Contributions and Practical Implications
Despite its limitations, we believe that this study has advanced our knowledge regarding recovery in general and the recovery potential of meditation in particular. First of all, as both participants who meditated and participants who listened to the radio interview reported a decrease in subjectively experienced stress and an increase in serenity, this study indeed showed “low-effort” activities to be beneficial for recovery. However, our study also added to previous research regarding these activities (e.g. Sonnentag, 2001) by showing that not all “low-effort” activities are created equal. That is, participants in the meditation condition recovered more than participants who listened to the radio interview. As this difference could partly be explained by differences in the recovery experiences “relaxation” and “mastery”, our study confirms the importance of subjective experiences in understanding the recovery potential of activities one can engage in after a stressful experience. Furthermore, due to our examination of recovery experiences, we were among the first to provide insight into the psychological underlying mechanisms explaining the effectiveness of meditation. This is especially important given the finding that physiological mechanisms are not always sufficient to explain meditation's effectiveness (Cahn & Polich, 2006; Eifert & Heffner, 2003; Erisman & Roemer, 2010). Therefore, an interesting avenue for future research would be to integrate both perspectives by examining the complementary roles of physiological and psychological mechanisms in the relation between meditation and recovery from stressful events.
Furthermore, by showing that meditating is especially effective if the meditator is intrinsically motivated, we were able to provide insight into the possible moderating factors that may affect meditation's effectiveness. Because of the strong connection between intrinsic motivation and pleasure, this result supports the idea that pleasure may help reduce the stress response (Esch & Stefano, 2004; Fredrickson et al., 2000). However, as intrinsic motivation did not affect the recovery effects of the radio interview, it may be that intrinsic involvement is especially important for those activities that are strongly associated with recovery experiences. Indeed, compared to the radio interview, the meditation exercise is more focused on releasing tension and on generating pleasant thoughts. This could be further examined in future research.
From a practical point of view, our results show that one should refrain from applying a “one type fits all” strategy for the application of recovery activities. Specifically, meditation can most successfully be applied among participants who enjoy doing it. Participants who are less intrinsically motivated to meditate should preferably be matched to another activity for which they do have the motivation. Also, as our results reveal that levels of relaxation and mastery experienced while engaging in a potentially stress-reducing activity are positively related to recovery, it could be recommended that people engage in activities that allow them to experience mastery and relaxation.
Our findings have other practical implications as well. Most studies examining the effects of meditation so far have employed 6- to 12-week interventions that additionally required participants to practice in the time between sessions (Rausch et al., 2006). This may often be necessary, because most meditation techniques require considerable practice (e.g. mindfulness-based stress reduction takes an 8-week course; Kabat-Zinn, 1982) before they can be executed correctly, but is not always desirable from a practical point of view. For example, organisations looking for a method to reduce work-related stress would probably prefer a solution that their employees can use without having to invest a lot of time and effort in learning how to practice it. By showing that a single 15-minute session of guided imagery meditation is effective in recovery from stress, this study provided a practical tool for organisations wishing to introduce a stress-reduction program. An accessible 15-minute session would also be valuable for people to use as a recovery activity at home.
All in all, based on these theoretical and practical considerations, we believe our study provided valuable insights regarding the use of meditation as an activity that may contribute to recovery from stress.