The phenomenology of maintenance goals: lower threat and greater satisfaction with the current state

This research tested the mental experience of maintenance goals as distinct from goals to approach better outcomes (approach goals) and goals to avoid worse outcomes (avoidance goals). In Studies 1 and 2, participants reported personal goals and categorized them as one of the three goal types. We theorized that maintenance centers on existing positive outcomes, while avoidance centers on future (expected) negative outcomes. We therefore predicted a lesser experience of threat and a greater experience of satisfaction in maintenance compared to avoidance. Additionally, we predicted greater satisfaction in maintenance compared to approach goals, wherein motivation comes from the value of future (desired) positive outcomes rather than current ones. Confirming our predictions, participants rated personal maintenance goals as lower on threat (Study 1) and higher on satisfaction (Studies 1 and 2) compared to avoidance goals. Additionally, revealing the centrality of maintenance, maintenance goals played a major role in reported central life goals (Study 1) and the largest role in goals pursued in the previous week (Study 2). Finally, in Study 3, participants judged another person engaged in maintenance as more satisfied and less threatened than a person engaged in avoidance or approach. Taken together, this work informs about the phenomenological nature of maintenance and advances a theory-driven ternary taxonomy of basic goal types.

achieved; holding on to the love, intimacy and commitment that make the relationship last. However, while this new stage may be joyful and reassuring, it may also lead to the fear of losing what one has. Accordingly, the shift in people's goals as the romantic relationship matures can be a transition from approach to either maintenance or avoidance. In the current paper, we define approach goals as the pursuit of things that have not yet been attained. We additionally argue that maintenance and avoidance are two distinct goal types that people may form after attaining their desired state. Avoidance goals represent the attempt to ward off threats, whereas maintenance goals support stable positive states with minimal threat levels. As we will show, this distinction may have implications for the mental experience of goal pursuit.
The beginning of a love romance is typically characterized by desired end points that one wishes to reach, whether it is intimacy, an emotional bond, commitment, or marriage. As time goes by and the relationship settles, such goals are often fulfilled. Then, different types of goals arise. The main theme may now become to maintain what was already From a binary to a ternary goal distinction Throughout the history of psychological thought, the predominant view on motivated behavior was one of progress towards unattained objects (Hull, 1943;McDougall, 1908;Murray, 1938;Spence, 1951). Parallel to that unitary view on motivation, a binary view that received increasing attention in the last decades stresses the distinction between approach and avoidance (Elliot & Church, 1997;Elliot & Covington, 2001) as well as between the regulatory foci of prevention and promotion (Higgins, 1997). Approach goals direct behavior toward positive outcomes, whereas avoidance goals direct behavior away from negative outcomes. Following a similar pattern, promotion and prevention foci are often described as higher-order approach and avoidance goals, reflecting "more general desires for approaching positives or avoiding negatives" (Molden et al., 2008;p. 171).
Importantly to the current paper, both the category of "promotion goals" and that of "approach goals" may be construed broadly enough to include the pursuit of already attained positive states (maintenance goals). The research that stems from these frameworks, however, has largely equated "promotion" and "approach" with the attempt to reach unattained states (see, for instance: Elliot et al., 2001;Elliot et al., 2006;Förster et al., 2001).
Moreover, and perhaps more crucially, including maintenance goals within the goal category of "approach" blurs potential underlying differences in the motivational mechanisms underlying the pursuit of current states as compared to the pursuit of future states. As a result-despite the vast importance of human efforts to hold on to existing resources (nurturing children, maintaining relationships, sustaining the natural environment, etc.)-theory on self-regulation has mostly overlooked the motivation to maintain (Ecker & Gilead, 2018).
In contrast to the predominant binary framework, this paper follows a ternary framework, distinguishing between three basic goal types: maintenance, approach, and avoidance. Although less common, this framework is not novel (Ebner et al., 2006;Freund, 2006;Gong & Freund, 2020) and has received empirical support (Lappi & Wilkowski, 2020). We define approach goals as the attempt to change the current state. As conceptualized by Carver & Scheier (2000), the basic mechanism of such goals implicates a feedback loop aimed at closing a gap between the current state and a certain desired state. The mechanism of avoidance goals, in contrast, is assumed to reflect an attempt to increase the distance from an undesired state. Finally, the unique nature of maintenance goals is such that they implicate neither an attempt to move towards something positive (change the current state) nor move away from something negative (reduce threat). When pursuing maintenance, the individual invests in keeping both the current state and the current (hopefully, optimally low) level of threat. For instance, when one's relationship is secure, they may still invest goal-directed efforts in nurturing it, independent of a wish to makes things better or to counteract threats. Our ternary approach thus views maintenance, approach, and avoidance as separate basic goal categories.

Evidence of the distinct nature of maintenance goals
The current paper aims to understand the experience of maintaining current states, which we have termed maintenance goals, as distinct from the experience of pursuing better states (approach goals), and the experience of avoiding worse states (avoidance goals). Previous research has taken steps in this direction by examining the difference between maintenance and approach goals (Brodscholl et al., 2007;Stamatogiannakis et al., 2018;Yang et al., 2015). Unfortunately, however, this research has mostly neglected to compare maintenance to avoidance goals. For instance, Brodscholl et al. (2007) compared a goal to maintain previously acquired tokens throughout a task to a goal to attain similar tokens. However, because this study did not compare maintenance to avoidance goals, nor verify that participants construed their task as a maintenance task, it is hard to determine whether participants were primarily motivated by a goal to maintain the acquired token or a goal to avoid losing them.
Although maintenance is a major topic in theory and research on behavior change, it is portrayed there as distinct from approach but not from avoidance. Moreover, this literature often equates maintenance with avoidance in assuming that the maintenance stage relies on a prevention-focused mindset (Rothman et al., 2004). Research on motivation across the life span, in contrast, has compared all three categories-maintenance, avoidance, and approach (Ebner et al., 2006;Gong & Freund, 2020). However, this research examined general goal orientations rather than the characteristics of specific goals. Perhaps due to this choice, this research did not detect differences between maintenance and avoidance and therefore cannot inform us of the (potentially) distinct experience of maintenance. Differentiating maintenance from avoidance is particularly important for a theory of maintenance goals because both goals imply a disinterest in changing the current state. This commonality might mean that these two goals are distinct only on a communicational level. It is possible, for instance, that people categorize goals in certain domains (e.g., relationships) as maintenance rather than avoidance, while goals in both categories are actually characterized by a similar phenomenology. If maintenance goals are a distinct goal type, however, they should lead to consistent experiential differences from both approach and avoidance. Revealing and delineating these differences would be a valuable contribution to theory on the general mechanisms of goal-pursuit, as well as a potentially important insight for real-life problems (Kwasnicka et al., 2016;Nigg et al., 2008;Rothman, 2000).

Hypotheses: the phenomenology of maintenance goals
Maintenance goals represent the proactive attempt to maintain the current state of affairs (Ecker & Gilead, 2018). They guide individuals to satisfy the routine needs of stable positive states with optimally low levels of threat. This formulation of maintenance goals brings us to two testable hypotheses, centering on the distinction between maintenance and avoidance. First, (H1:) we hypothesize that the sense of threat will be lower in maintenance as compared to avoidance goals. This predicted contingency may result from two causal effects. When threat is low, there is no need to reduce it and hence no need to pursue avoidance. Consequently, situations with low threat would invite people to set maintenance rather than avoidance goals. Additionally, maintenance goals focus on a desired current state rather than on potential threats.
Our second and third hypotheses concern differences in sense of satisfaction with the current state. As in our rationale for H1, because maintenance goals center on valued current states, they are likely to evoke greater satisfaction than a focus on negative potential outcomes, as well as in comparison to a yet-to-be-attained valued state. Additionally, satisfactory current states are more likely to lead people to set maintenance goals as compared to goals aimed at changing or improving on the current state. Accordingly, (H2:) we hypothesize that maintenance goals will be characterized by a higher sense of satisfaction as compared to avoidance goals, and that (H3:) maintenance goals will be characterized by a higher sense of satisfaction as compared to approach goals.
In total, our theorizing leads us to specify two predicted differences between the experience of maintenance and that of avoidance goals, and a third difference that distinguishes the experience of maintenance goals from that of approach goals. We tested our hypotheses using reports of real goals and social perceptions of others' goals. In Studies 1 and 2, we measured participants' self-reported satisfaction and sense of threat in relation to their personal life goals. In Study 3, we tested social perceptions of others that engage in maintenance, avoidance, and approach.

The present research
We tested the hypotheses in three studies. In Studies 1 and 2, we ask participants about their central life goals (Study 1) and the goals they pursued in the previous week (Study 2). In both studies, we measured the extent to which participants experienced their personal maintenance, approach, and avoidance goals as threatening, as well as their satisfaction with the current state of each goal. In Study 3, we moved on to test our hypotheses on people's social perceptions of others' goals. Specifically, we tested the extent to which others who maintain, approach, or avoid were perceived as satisfied and under threat.
All procedures were conducted in accord with the American Psychological Association Ethical Principles of Psychologists and Code of Conduct (APA, 2017). At the senior author's institution (JGU Mainz, Department of Psychology, Ethics Committee), studies that did not involve deception, vulnerable populations, identifiable data, intensive data, or interventions were exempt from ethical approval and not evaluated at time of data collection. Across all studies, we report how we determined our sample size, all data exclusions (if any), all manipulations, and all measures. All data and materials pertaining to this project are fully available on the Open Science Framework (OSF; https://osf.io/ yj2pd/). We pre-registered our predictions in Studies 2 and 3. We did not per-register Study 1 because we considered it largely exploratory.

Study 1
To test our hypotheses about the phenomenology of maintenance compared to avoidance and approach, Study 1 examined how people experience their central life goals. We predicted that maintenance goals will be experienced as lower on threat and higher on satisfaction than avoidance goals, as well as higher on satisfaction compared to approach goals.

Participants
We collected data from 401 mTurk participants (all US residents; 43% women; Mage = 36.22, SD = 10.51). The sample size in this study, and in all studies in this paper, was determined before data collection began. We included all observations in the analysis because criterions for exclusion were not pre-registered. Subsequent simulations that we conducted with the SIMR package in r (Green & MacLeod, (avoidance vs. maintenance), leaving maintenance goals as the reference group. Dependent variables were left in their original metric. P-values for fixed effects were obtained with the LmerTest package (Kuznetsova et al., 2016), based on Satterthwaite's degrees of freedom method.

Results and discussion
Out of 401 participants, 354 reported at least one maintenance goal, 151 reported at least one avoidance goal, and 392 reported one or more approach goals. Figure 1 shows the percentage of maintenance, avoidance, and approach goals in each life domain. Participants reported a considerable amount of maintenance goals in all life domains, ranging from 42% in the romantic domain to a low of 18% in the financial domain. Avoidance goals appeared considerably less in all domains, ranging from 7 to 11%. Table 1 presents mean ratings on the dependent measures by goal type, with different superscripts indicating significant differences in the mixed-effects analyses. Tables S1 and S2 in the supplementary material present mean ratings on additional measures (Table S1), and the full results of the mixed-effects analyses (Table S2). Figure S1 in the supplementary material presents differences between goal types for each of the central measures by goal domain.

Study 2
Study 2 aimed to further test our hypotheses on a different type of personal goals. Instead of asking participants about central issues in different life domains, we asked them to report what they did last week. This enabled us to collect information about more mundane goals rather than the abstract goals that people may report as central in their life. 2016) concluded that the eventual power to find the predicted effects was very high (> 99%).

Procedure and measures
Participants were first instructed to name something that is important to them in each of seven domains: the household, workplace, parenting, romantic, financial, health, and the spiritual domain. Then, participants indicated whether their aspiration about each issue that they named is best described as "preventing things from getting worse" (a avoidance goal), "keeping/maintaining things as they are" (a maintenance goal) or "improving things" (a approach goal). Finally, participants were asked to formulate their aspirations as goals (e.g., "my goal is to lose weight"). These formulated goals were then presented to them in a series of questions.
The first question was "how important to you is each of the goals you listed?". Then, participants answered two sets of questions in random order. In one set, participants indicated for each goal the extent to which it feels boring, repetitive, exciting, motivating, and difficult. In another set, participants reported the extent to which they feel satisfied, threatened, and unstable about the current state of each goal. All responses were given on a sliding scale ranging from 0 to 100. Questions within each set appeared in random order. Lastly, participants were asked to indicate their current situation in relation to each goal, using a 9-point response scale ranging from "much less than ought level" to "much more than ideal level" ("ought" and "ideal" levels were defined in the instructions). The full verbatim account of this study is openly available on the OSF.

Analytic procedure
We first conducted descriptive analyses on the raw data. Then, for statistical inference, we conducted linear mixed effects analyses using the lme4 package (Bates et al., 2012) in R (R Core Team, 2012). This analytical method allowed us to maximize the informational value of our data by observing ratings of all of the goals that participants reported rather than calculating average scores by participant. We entered goal type as the fixed effect and intercepts for participants and life domains as random effects. We treated both measures implicated in our predictions-satisfaction and sense of threat-as dependent variables 1 . Goal type was added to the analysis using two dummy variables: Approach goals were dummy coded as 1 on the first dummy variable (approach vs. maintenance), and avoidance goals were dummy coded as 1 on the second dummy variable removing participants that did not report their weekly activities properly in open text questions. Because data quality was very high, as reflected in fully coherent responses to the open text questions, we kept all 280 participants in the analysis (73% women; Mage = 33.65, SD = 11.99). Simulation conducted with the SIMR package in r (Green & MacLeod, 2016) concluded that the eventual power to find the predicted effects was acceptable for H1 and H3 (ranging from 72 to 100%) but, unfortunately, very low for H2 (10.00%).

Procedure
Participants were instructed to report ten activities that they engaged in in the last week in a table with ten rows. For each activity, they also indicated in the table the weekdays they engaged in it and the category to which it belonged: Household, Parenting, Friends/social, Dating/relationship,

Participants
We collected data from 280 UK residents on Prolific. We pre-registered (https://aspredicted.org/blind.php?x=j226ki)  intercept for participants because a random intercept for goal domain did not allow for all models to converge. Other than that, the analyses were similar to Study 1.

Results
Out of 280 participants, 276 participants reported goalrelated activities, 246 reported at least one maintenance goal, 250 reported at least one approach goal, and only 90 participants reported one or more avoidance goals. The percentage of goal-related activities out of all activities was 52%. Out of the goal-related activities, 46% were maintenance goals, 44% were approach goals, and 10% were avoidance goals. Thus, in this sample, most of people's daily goals were maintenance goals. Figure 2 shows the percentage of maintenance, avoidance, and approach goals out of the total number of goal-related activities in each life domain. For ease of presentation, we grouped together the relational domain-parental, relationship and social -under the category "relational". Table 2 presents the mean ratings on the relevant dependent measures by goal type, with different superscripts indicating significant differences in the mixed-effects analyses. Tables S3 and S4 in the supplementary material present mean ratings on additional measures (Table S3), and the full results of the mixed-effects analyses (Table S4). Figure   Workplace, Resting, Health/body care, Hobbies, Studying, or Other. After listing all activities, participants were asked to answer one question about each activity: "Is this activity related to a goal?". The response options were "Related to a goal to keep/maintain something that I value" (maintenance), "Related to a goal to avoid something or prevent an unwanted event" (avoidance), "Related to a goal to attain something I desire or make something better" (approach), or "Not related to a goal".
Finally, participants answered seven questions on a 7-point scale and in random order about all their goalrelated activities: "To what extent do you feel that your situation in relation to each goal is threatened?" (Sense of threat), "How satisfied are you with your situation in relation to each of these goals?" (Satisfaction), "To what extent does your situation in relation to each goal make you feel happy?", "To what extent is each goal related to your relations with close others -family, friends, or romantic partners?", "How motivated do you feel to pursue each of these goals in the future?", "To what extent do you feel that pursuing each goal challenges you?", and "How confident do you feel in your ability to pursue each goal?".

Analytic procedure
We used similar methods to Study 1 in describing and analyzing the data. As random effects, we entered only an Additionally, perceptions of self-efficacy were higher in maintenance compared to both approach and avoidance goals. We therefore tested whether self-efficacy can account for the higher levels of threat in approach compared to maintenance goals, as well as for the greater level of satisfaction in maintenance compared to approach and avoidance goals. Similar to challenge ratings, including self-efficacy in a regression model predicting threat from goal type led to a non-significant effect of maintenance compared to approach on threat, b = 0.08, SE = 0.08, t = 1.10, p = .273. Furthermore, including self-efficacy in a regression model predicting satisfaction from goal type led to a non-significant effect of maintenance compared to avoidance on satisfaction, b = -0.3, SE = 0.11, t = -0.25, p = .805. In contrast, the effect of maintenance compared to approach on satisfaction remained significant, b = 0.67, SE = 0.02, t = 8.11, p < .001.

Discussion
In Study 1, we asked participants about subjective experiences pertaining to central life goals in seven life domains, whereas in Study 2 participants reported experiences pertaining to the goals they pursued in the previous week. Overall, the results of Studies 1 and 2 lend support to our three hypotheses. In Study 1, participants' sense of threat was considerably higher in avoidance compared to maintenance. Although this was not the case in Study 2, the descriptive difference in threat in Study 2 suggests that insufficient observations of avoidance may have led to low power to find the effect. Both Studies 1 and 2 supported H2 and H3, showing a greater sense of satisfaction in maintenance compared to avoidance and approach. Taken together, these findings confirm the unique phenomenology of maintenance pursuit.
Both Studies 1 and 2 also revealed the prevalence of maintenance goals in people's reports of personal goals. In Study 1, participants were first asked to name something that is important to them and only then to phrase a goal that relates to it. This may have biased participants to focus on the things they already have and therefore on maintenance goals. In Study 2, however, participants were instructed to report activities that they engaged in throughout the previous week and then the goal that guided each activity. Importantly, this procedure seems unlikely to have biased participants to report more maintenance goals. Nevertheless, participants viewed their previous week's activities as motivated by maintenance goals more than by approach goals, avoidance goals, or no goal at all.
Interestingly, the results of Study 2 also show that greater self-efficacy in maintenance can potentially account for S2 presents differences between goal types on the central dependent measures by goal domain.

Main analysis
In this study, the main effect of maintenance vs. avoidance on threat was not significant, and thus did not lend further support for H1, b = 0.12, 95% CI [-0.16, 0.39], SE = 0.14, t = 0.84, p = .402. However, the descriptive differences in threat suggest that the effect was not found because there may not have been sufficient observations in the avoidance goals condition to detect an effect. A descriptive difference of the same size in the threat means of approach and maintenance did reach significance, b = 0.22, 95% CI [0.06, 0.38], SE = 0.08, t = 2.71, p = .007, and reflected a higher sense of threat in approach compared to maintenance goals.

Additional analysis
As Table 2 shows, perceptions of challenge were higher in approach goals compared to both maintenance and avoidance. We tested whether the greater experience of challenge in approach accounts for differences in threat and satisfaction in progress compared to maintenance goals. Indeed, including challenge ratings in a regression model predicting threat from goal type led to a non-significant effect of maintenance compared to approach on threat, b = 0.09, SE = 0.08, t = 1.11, p = .265. When including challenge in a regression predicting satisfaction, in contrast, the effect of maintenance concludes that a person engaged in maintenance feels more satisfied than a person engaged in avoidance, they are likely to rely on the extent that they themselves feel satisfied when doing one or the other. To the extent that this is true, judgments of others who maintain, make approach, or prevent, inform us about participants' own experiences.
An important advantage of observing perceptions of others (rather than the self) is that they are less influenced by social desirability. In other words, when judging others rather than themselves participants are less likely to embellish their judgments according to what they may view as desirable. Additionally, manipulating the descriptions of others' goal striving behavior allowed us to achieve greater experimental control than in previous studies in which we could only measure associations with self-generated goals.
We asked participants to read about three types of people -one who likes to maintain, one who likes to make progress, and one who likes to prevent. After each text, we asked participants to evaluate how satisfied and threatened the person seemed in relation to his goals. As in Study 2, we pre-registered (https://aspredicted.org/blind.php?x=gf5ti4) the predictions that perceptions of satisfaction would be higher towards the person that engages in maintenance (we will term him "the maintainer") compared to the person that engages in avoidance ("the avoider"), whereas perceptions of threat would be lower towards the maintainer compared to the avoider. Additionally, we predicted that perceptions of satisfaction will be higher towards the maintainer compared to the person engaged in approach ("the approacher") 2 .

Participants
We predetermined and invited 300 participants from the United Kingdom on the Prolific platform. The actual sample included 297 participants (53% women; Mage = 37.83, SD = 13.73), reaching 80% power to find and effect size of d = 0.15 in a paired t-test.

Procedure
Each participant read about three men in random order -Michael who is passionate about gardening, David who is passionate about fitness, and Brian who is passionate about his job. Participants were randomly assigned to one of six possible variations of the pairing between scenario and goal type. The three goal type variations in each scenario varied only in key words that reflected the specific goal type. The full verbatim of these scenarios are included in the the higher levels of satisfaction in comparison to avoidance goals. One possible explanation for this finding is that greater satisfaction in maintenance is caused by higher self-efficacy. It is also possible, however, that the pursuit of maintenance affects both goal satisfaction and satisfaction with one's self in a similar manner. Because the pursuit of maintenance focuses on positive current outcomes rather than negative expected outcomes, it may remind people of the good they have already obtained rather than the bad that may occur. In doing so, maintenance goals may make people feel more capable and efficacious than when thinking of what might happen as in avoidance pursuit. Additionally, it is possible that the higher self-efficacy in maintenance is the result of greater satisfaction with the goal state. Indeed, it seems reasonable that one would feel more capable of achieving desired outcomes when satisfied with what they have thus far achieved.
It is noteworthy that both Studies 1 and 2 controlled for individual differences in the tendency to experience satisfaction and threat. Such tendencies may be related to participants' goal orientation (Ebner et al., 2006;Elliot & Thrash, 2002) or to their chronic regulatory focus (Shah et al., 1998). For instance, people with a prevention regulatory focus may experience greater threat than participants in a promotion regulatory focus. Assuming such participants also tended to report more avoidance goals, this could bias the results towards higher estimations of threat for avoidance goals when using a linear regression. However, by including random intercepts for participants in mixed-effects models predicting satisfaction and threat, we allowed the base level of these variables to change between participants. Our estimations of the effect of goal type on satisfaction and threat are therefore independent from individual base rates on these variables.

Study 3
In this last study, we examined social perceptions of goals. Taking a step beyond the phenomenology of one's personal goal-pursuit experiences, we examine whether people experience other people differently depending on the type of goals they pursue. One's social environment may seem essentially different according to whether it centers on maintaining, preventing, or making approach. One determinant of social judgments about goals may be naïve theories and beliefs about the way people are (Gopnik & Wellman, 1992). For instance, people may believe that people who maintain things are more satisfied than people who avoid things. Additionally, we reason that one's judgment about mental states of others relies, to a large extent, on their own mental experiences (Gordon, 1986). For instance, when one

General discussion
This research tested the phenomenology of maintenance goals as compared to approach and avoidance goals. We theorized that maintenance striving involves a focus on positive current states, whereas avoidance striving involves a focus on negative future states, and approach striving involves a focus on positive future states. Accordingly, we predicted that the experience of threat would be more salient in avoidance than in maintenance goals (H1), whereas the experience of satisfaction would be more salient in maintenance goals than in both avoidance (H2) and approach goals (H3). These hypotheses were confirmed in three studies. Participants rated personal maintenance goals as lower on threat (Study 1) and higher on satisfaction (Studies 1 and 2) than both personal approach and avoidance goals. Given that maintenance goals are not centerstage in motivation research, it seems particularly noteworthy that respondents in these two studies classified the largest share of their personal goals as maintenance goals. Finally, further confirming our three hypotheses, Study 3 showed that people's social perceptions of others that engage in maintenance, approach, and avoidance are in line with the personal experiences measured in Studies 1 and 2. Participants assigned more satisfaction to maintainers than to both avoiders and approachers, and less threat to maintainers compared to avoiders.
Past theorizing on motivation has largely neglected the distinction between maintaining current states and the pursuit of change, focusing instead on a binary distinction between approach and avoidance (Ecker & Gilead, 2018). Elliot and Fryer (2008), for instance, defined goals as cognitive representations of future objects that one is committed to approach or avoid. Although it was often acknowledged that people may also pursue current states, the distinct characteristics of this type of striving was rarely explored. Moreover, previous work that centered on maintenance striving has distinguished it only from the pursuit of change, thus potentially leading to the conclusion that maintenance is synonymous with avoidance. Contradicting this notion, our results show that participants' personal maintenance goals, as well as perceptions of others' maintenance goals, were characterized by considerably less threat and more satisfaction compared to avoidance. The role of threat in maintenance goals may be reduced because such goals are the attempt to keep threats at bay rather than to remove them. Maintenance striving enables a circular movement wherein threats are neither materialized nor removed. The threatening experience of feeling hungry, for instance, can never be permanently removed; instead, it is stabilized in a cycle wherein hunger arises and is then satisfied. Surely enough, this cycle includes subordinate approach and avoidance supplementary material. After reading about each person, participants were instructed to "try to form an impression of what kind of person he may be", and were then asked to answer a set of questions.
The first set of questions included the central pre-registered measures. Participants indicated on a 7-point scale "To what extent does David seem like the kind of person that…." -"Often feels happy when working (tending to his garden/ fitness)", "Often feels threatened about his work (garden's/ fitness) situation" (threat), "Is generally satisfied with his work (fitness/ garden)" (satisfaction), "Likes to connect to other people through is work (fitness hobby/ gardening hobby), "Likes to think people that he cares about at work would enjoy collaborating with him (people that he cares about would enjoy fitness activities with him/ would enjoy his garden)", "Is interested in sharing his work (fitness/ gardening) experiences with friends and family". In the second set of questions, participants rated each man on five statements adapted from the Caring subscale of the Parental Care and Tenderness (PCAT) questionnaire (Buckels et al., 2015), as well as on 14 emotions from the Positive and Negative Affect Scale (PANAS; Watson et al., 1988).

Results
We conducted paired t-tests to compare between goal type conditions. Table 3 presents average scores and standard deviations on threat and satisfaction by goal type. Results on measures that are not relevant to the current paper are presented in the supplementary material. Confirming H1, participants perceived avoiders as considerably more threatened than maintainers, t(296) = 13.66, p < .001, d = 0.79. Confirming H2, participants perceived maintainers as more satisfied than avoiders, t(296) = 6.75, p < .001, d = 0.39. Finally, confirming H3, maintainers were also evaluated as more satisfied than approachers, t(296) = 3.94, p < .001, d = 0.23. In sum, the results show that people perceive other people who engage in maintenance as more satisfied than people who engage in either approach or avoidance, and less threatened than people who engage in avoidance. (1.14) C Notes: Different superscripts in one row indicate a significant difference between the two values. The response scales on all measures ranged from 1 to 7.
Funding Open Access funding enabled and organized by Projekt DEAL.

Data availability
The data that support the findings of this work are openly accessible in Open Science Framework at https://osf.io/yj2pd/.

Declarations
Conflict of interest Yael Ecker declares that she has no conflict of interest. Michael Gilead declares that he has no conflict of interest. Roland Imhoff declares that he has no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/. goals, but the overarching goal that governs it-monitoring and allocating resources for proactive efforts (e.g., making sure there is food in the fridge)-is a maintenance one.
A shortcoming of the current paper lies in its reliance on participants' introspective ability to distinguish between goal types. Future research may establish ways to manipulate the pursuit of maintenance compared to avoidance and approach. This would allow an examination of the experience of maintenance that is independent of potential individual differences in naïve theories about goals and introspective ability. Moreover, future research may build on the current findings to go beyond phenomenology and test whether maintenance relies on different motivational underpinnings than approach and avoidance. While approach goals are energized by biological systems that react to reward, and avoidance goals are energized by systems that react to threat (Carver & White, 1994;Corr et al., 2013;Gray, 1990), it remains to be examined what are the motivational systems that underlie maintenance striving. One possibility is that maintenance goals govern broader motivational processes that inherently encompass both approach and avoidance regulation. For instance, when maintaining a relationship, one may both avoid fights and approach positive joined experiences. As a result, maintenance motivation may benefit from the underlying motivational systems of both approach and avoidance.
An additional shortcoming of this research is that it collected information only from residents of English-speaking countries, that is, from people in a highly individualistic cultural environment. People in collectivist societies may have a greater emphasis on maintenance than those in individualistic societies, potentially leading to differences in the subjective experience of maintenance striving. Future research should build on the results of the current paper in conducting a cross-cultural comparison of the characteristics of the three goal types.
In sum, the current paper identifies systematic differences in the mental experience of maintenance, approach, and avoidance goals. Although this might seem like a modest goal, it paves the way for future research to investigate whether the observed phenomenological differences between maintenance, avoidance, and approach goals are implicated within different underlying mechanisms. Ultimately, it is of utmost importance to gain a better understanding on the human capacity to maintain what already is, on an individual and global level. For that singularly valuable goal, an understanding of the unique characteristics of the motivation to maintain is an asset.