Abstract
The current research sheds new light on how individuals can best use the consideration of future outcomes as a self-control strategy to enhance their likelihood of goal attainment. Across three studies, the authors find that the effectiveness of positively versus negatively valenced outcome elaboration is dependent upon the construal level at which the potential outcomes are considered. This research demonstrates that positive outcome elaboration is more effective when it is abstract, whereas negative outcome elaboration is more effective when it is concrete. Moreover, the authors explore the process underlying these effects and demonstrate that the increased effectiveness of matching the outcomes’ valence and construal level is due to outcome elaboration fluency, as increased ease of generating outcomes that are positive and abstract or negative and concrete promotes more effective self-control.
Keywords
As children, we are extolled to consider the consequences of our actions before we take them as suggested by proverbial statements, suggesting we should “think before you act/speak,” “look before you leap,” and “measure twice before cutting once.” This conventional wisdom readily applies to decisions relevant to reaching one’s goals, given that much goal-directed behavior depends upon overriding immediate impulses in pursuit of longer term goals. Strategies like outcome elaboration, which specifically bring attention to future periods of time, can increase self-control effectiveness, as they allow individuals to consider consequences consistent with their higher order goals (Baumeister & Heatherton, 1996; Haws, Bearden, & Nenkov, 2012; Nenkov, Inman, & Hulland, 2008). But what is the best way to focus one’s future thoughts? In the current research, we focus on variations of outcome elaboration that work to enhance self-control and long-term goal attainment. We propose and demonstrate that the effectiveness of both positive and negative outcome elaboration can be maximized by varying the construal level of the outcome elaboration.
Prior research has provided conceptual (Baumeister & Heatherton, 1996) and empirical (Nenkov et al., 2008) evidence that considering potential outcomes supportive of the achievement of individuals’ future goals generally enhances present self-control by helping them focus on the future and transcend present temptations. In general, people can adopt either a positive or negative outcome focus in their goal pursuit (Carver & Scheier, 1998; Higgins, 1999) and can similarly encode anticipated end states with a positive versus negative valence (Nenkov et al., 2008). In other words, people can think about the potential positive outcomes of exerting self-control in the present in order to achieve a future goal (e.g., eating healthy food will help me lose weight and be happy) or the potential negative outcomes of indulging themselves in the present and failing to achieve a future goal (e.g., eating rich food will make me overweight and unhealthy). Since both positive and negative outcome elaboration can provide individuals with information as to whether a current decision has the potential to move them toward goal achievement success or away from goal failure, respectively, both of these approaches should lead to better self-control and more appropriate behavior modification.
In addition to valence, we consider the construal level of one’s future outcome elaboration. Construal level theory (Trope & Liberman, 2003) defines construal level as the degree of abstraction at which goals and actions are represented (see also Carver & Scheier, 1998; Vallacher & Wegner, 1987). For example, a consequence of goal achievement success (e.g., losing weight) can be framed at abstract construal level by emphasizing the overall, broad, and context-independent aspects of goal achievement (e.g., be healthy and happy). In contrast, a consequence of achieving the same goal can be framed at concrete construal level by emphasizing the specific, context-dependent details of goal achievement (e.g., get into a healthy daily routine and fit into my favorite jeans).
Construal level has been shown to be relevant to self-control–related outcomes (Fujita, Trope, Liberman, & Levin-Sagi, 2006; Laran, 2010). Generally, abstract construal has been associated with enhanced self-control (Fujita & Han, 2009), as self-control involves focusing on long-term or high-level construals as opposed to short-term or low-level construals (Trope & Fishbach, 2000). On the other hand, various research perspectives point to the potential for concrete construals to have positive effects on the successful exertion of self-control (Gollwitzer, 2011). To reconcile these two different streams of research, understanding conditions under which abstract or concrete outcome elaboration will be more effective for enhancing self-control is critical.
Recent research has started to establish a link between valence and construal level in action consideration (Eyal, Liberman, Trope, & Walther, 2004; Herzog, Hansen, & Wänke, 2007). Eyal, et al. (2004) suggested that pros (i.e., arguments in favor of taking an action) become more salient as temporal distance from the action increases, whereas cons (i.e., arguments against taking the action) become more salient when temporal distance decreases. Eyal et al. (2004) proposed that pros constitute a higher level of construal than cons, because cons are subordinate to pros in decision making, so that the cons are considered only if the level of pros is sufficient. Herzog, Hansen, and Wänke (2007) further showed that temporal distance also affects the ease of retrieval associated with generating arguments, such that it is easier for people to generate pros but more difficult to generate cons if an action pertains to the distant rather than the near future. Based on these findings, it seems that pro arguments might constitute a higher level of construal, and con arguments, a lower level of construal, although this research has only examined temporal shifts in the salience of pros and cons in the context of non-self-control relevant judgments that do not require intertemporal trade-offs.
Building on this research, we suggest that even though both positive and negative outcome elaboration pertain to the future, positive outcome elaboration, which refers to the pros of goal success, is more likely to be construed at an abstract level, whereas negative outcome elaboration, which refers to the cons of goal failure, is more likely to be construed at a concrete level. Further, we propose that in the context of outcome elaboration that is consistent with long-term goals, the match or mismatch between the valence and construal level of outcomes is likely to influence the effectiveness of outcome elaboration as a self-control strategy. Specifically, positive outcome elaboration should match with an abstract construal level and should promote more effective self-control when it is abstract and general. On the other hand, negative outcome elaboration should match with a concrete construal level and should promote more effective self-control when it is concrete and specific.
We further propose that these matches between the valence and construal level of outcome elaboration will enhance the ease with which potential outcomes are generated, giving rise to a processing fluency experience. Processing fluency pertains to the ease with which relevant thoughts can be generated (Schwarz, 2004). People rely on the subjective ease with which information comes to mind in forming judgments (Schwarz et al., 1991), for example, when recalling past events, ease of recall makes events seem more frequent. Similarly, it is possible that increased ease of generating potential outcomes might make these outcomes seem more likely to occur. If people perceive that the outcomes they expect to result from a given course of behavior are more likely to occur and move them toward goal achievement success or away from goal failure, they would be more inclined to undertake that behavior (Bandura, 1997; Carver & Scheier, 1998). As such, we expect that increased fluency of outcome elaboration that is consistent with long-term goal achievement should increase the perceived likelihood of outcome occurrence and should feed into one’s self-control decision (Novemsky, Dhar, Schwarz, & Simonson, 2007), increasing self-control effectiveness.
Study 1
The goal of this study was to test our proposition that positive outcome elaboration is naturally more abstract, whereas negative outcome elaboration is naturally more concrete. For this purpose, participants were randomly assigned to one of two conditions, where they participated in a thought exercise asking them to elaborate on either the potential positive outcomes of exercising self-control and choosing a more responsible option or the potential negative outcomes of failing to exercise self-control and choosing a more indulgent option. We then assessed the degree of abstraction of the consequences participants listed, which served as our main dependent variable.
Method
Participants
Eighty-five undergraduate students (45.2% female; M age = 21.24) participated in an online study in a controlled laboratory setting for course credit.
Procedure and Materials
Participants were first were asked to choose one of two different domains (financial management or time management issues) based on their personal relevance. Then, participants were asked to indicate the importance of their chosen issue, measured on a 7-point scale, 1 = not at all important, 7 = very important. Next, they were randomly assigned to one of two experimental conditions that manipulated the type of outcome elaboration they were asked to engage in: elaborate on the positive outcomes of exercising self-control or the negative outcomes of failing to exercise self-control.
Outcome Elaboration Instructions
We utilized a deliberative mind-set procedure adapted from Gollwitzer and Kinney (1989) to fit the different outcome elaboration conditions. Participants were told that they will participate in a thought exercise that is intended to focus their attention on the consequences of the things they do. They were given an example of how this thought exercise should be completed and then had to apply the exercise to their own self-control issue. Specifically, depending on their domain choice, participants were given one of the following two hypothetical scenarios and were asked to imagine themselves in these situations: Imagine that you just found out that your good friend is having a birthday party tonight. You really want to attend the party because all of your friends will be there and it will be fun, but there is a problem. You have an exam worth 30% of your overall grade in a difficult class the day after the party and if you attend the party you will not have enough time to study.
OR Imagine that you have just received your paycheck and you have some money left over after you pay your bills. You are trying to decide whether to put the money aside or spend it on things that might be fun or pleasurable right now.
Next, depending on condition, participants were asked to list five potential positive outcomes of choosing a more responsible option in the domain they selected (i.e., choosing to save, rather than spend their money; choosing to stay home and study, rather than go out and party) or five potential negative outcomes of choosing a more indulgent option (i.e., choosing to spend, rather than save; to party, rather than study).
Manipulation Check
After participants completed the outcome elaboration task, they indicated their agreement with the following two statements: “My focus was on the positive consequences that might occur” and “My focus was on the negative consequences that might occur” (9-point Likert-type scale; 1 = strongly disagree, 9 = strongly agree). Finally, gender and age measures were collected.
Results and Discussion
Choice of domain was evenly split among the two options, financial management: 45%; time management: 55%; χ2(1, N = 85) = 1.16, p < .30, indicating that these domains of self-control were equally relevant to our participants. We compared participants’ ratings of the importance of the self-control issue they chose across the two self-control domains and no significant differences emerged (all ps > .1). Therefore, we collapsed across the two self-control domains. We also ensured that our outcome elaboration valence manipulation was indeed successful. Our manipulation check revealed that participants in the positive outcome elaboration condition indicated a significantly stronger focus on the positive outcomes that might occur (M = 5.90), as compared to participants in the negative outcome condition, [M = 4.50], t(83) = 3.90, p < .001. On the other hand, participants in the negative outcome condition indicated a significantly stronger focus on the negative outcomes (M = 6.97), as compared to participants in the positive outcome condition, [M = 5.56], t(83) = 4.35, p < .001.
We next analyzed the consequences participants listed as part of the outcome elaboration task. Two independent judges blind to condition first coded each listed consequence as (1) abstract (i.e., consequence is more abstract and general and refers to reasons why one should or should not engage in the given behavior or to an end state one might achieve/not achieve; e.g., will build a successful career; will be unable to retire) or concrete (i.e., consequence is more concrete and specific, refers to contextual details or to aspects of the consequence related to implementational steps; e.g., will be tired on day of exam; cannot pay bills next month); (2) short-term (i.e., refers to the immediate future) or long-term (i.e., refers to the long-term future). The two judges’ ratings were highly correlated (r abstract/concrete = .90; r long-term/short-term = .94, ps < .01). Discrepancies in codes were resolved through discussion.
Since each consequence was coded as either abstract or concrete and either long-term or short-term and since these categories were completely dependent, we calculated the percentage of abstract and long-term consequences people listed by dividing the number of abstract/long-term consequences by the total number of consequences listed. As expected, participants in the positive outcome elaboration condition listed a significantly higher percentage abstract, t(83) = 10.04, p < .001, and long-term, t(83) = 6.61, p < .001, consequences, as compared to participants in the negative outcome elaboration condition; see Table 1. This evidence supports our contention regarding the natural tendency for positive elaboration to be more abstract and negative elaboration to be more concrete.
Study 1: Outcome Elaboration Coding.
Note. Standard deviations in parentheses. aPercent abstract consequences = Number of abstract/Number of total consequences. bPercent long-term consequences = Number of long-term/ Number of total consequences.
**Means are significantly different at p < .01.
To further examine the differences in outcome elaboration based on valence, we next utilized a coding scheme based on the linguistic categorization model (Semin & Fiedler, 1988) to analyze participants’ listed consequences. The two judges coded each predicate as belonging to one of four linguistic categories: descriptive action verb (DAV; e.g., study, save, and buy); interpretive action verb (IAV; e.g., help, learn, and need); state verb (SV; e.g., admire, enjoy, and feel); or adjective (e.g., good, quick, and hard). Past research has demonstrated that these four linguistic categories are organized along a dimension of concreteness to abstractness, with DAVs being the most concrete and adjectives being the most abstract. The two judges’ ratings were highly correlated (r DAV = .75; r IAV = .85; r SV = .87; r adjective = .93, ps < .01). Discrepancies in codes were resolved through discussion. After coding, we calculated a weighted abstractness index score from the number of DAVs, IAVs, SVs, and adjectives each participant used. To reflect the different levels of abstraction, we used a simple weighting schema (1–4) to weigh DAVs, IAVs, SVs, and adjectives, respectively (Semin & Fiedler, 1988). To control for length of the descriptions, we divided each weighted score by the number of coded predicates in the description.
Participants who elaborated on the positive outcomes of exercising self-control in the present used more abstract language than those who elaborated on the negative consequences of failing to exercise self-control, t(83) = 3.46, p < .001; see Table 1. The two conditions did not differ significantly in the number of predicates used. Analysis of each linguistic category revealed that although the conditions did not differ in the use of IAVs, or SVs, there was a significant difference in usage of DAVs and adjectives. Participants in the positive outcome elaboration condition used the most abstract type of predicate, adjectives more than participants in the negative outcome elaboration condition, t(83) = 2.48, p < .01, and the most concrete type of predicate, DAVs less than participants in the negative outcome elaboration condition, t(83) = 3.48, p < .001. These results provide strong evidence that positive outcome elaboration tends to be construed at a more abstract level, whereas negative outcome elaboration tends to be construed at a more concrete level.
Study 2
In this study, we independently manipulate the outcome elaboration valence and construal level, and we test our prediction that positive outcome elaboration will promote more effective self-control when abstract, whereas negative outcome elaboration will be more effective when concrete. In this study, we focus on the financial management self-control domain.
Method
Participants and Design
One hundred and thirty-one adults (56% female; M age = 33.0) from an online panel were paid to participate in an online study. Participants were randomly assigned to one of the four experimental conditions in a 2 (outcome elaboration valence: positive vs. negative) × 2 (outcome elaboration construal: abstract vs. concrete) between-subjects design.
Outcome Elaboration Instructions
We utilized a similar deliberative mind-set procedure employed in Study 1 but focused only on the financial management domain. After being presented with an example of how the exercise should be completed, participants engaged in a condition-specific elaboration task for the financial management domain. Participants in the positive (negative) and abstract condition were asked to list four general, broad positive (negative) consequences that they expected to occur in the future and consider the broad positive (negative) ways in which their current decision might impact their life overall, while those in the positive (negative) and concrete condition were asked to list four specific positive (negative) consequences that they expect to occur in the future and consider in detail the specific positive (negative) ways in which a current decision might impact their life in the future.
Self-Control Assessment
Following the outcome elaboration task, participants were provided with the same financial management hypothetical scenario used in Study 1 and were asked to indicate their preference for either the more responsible financial behavior or the more indulgent financial behavior on a 9-point scale ranging from will definitely spend now to will definitely save for later. As such, higher responses indicate more self-control effectiveness.
Manipulation Check
After participants completed the outcome elaboration task, they rated the consequences they listed as very specific—very general, positive—negative (measured on a 7-point semantic differential scale). Finally, gender and age measures were collected.
Results and Discussion
We confirmed that both the valence, 1 = positive, 7 = negative; M positive = 1.49, M negative = 6.30; t(129) = 769.29, p < .001, and construal, 1 = very specific, 7 = very general; M abstract = 3.48, M concrete = 2.65; t(129) = −2.93, p < .01, manipulations were effective. We next ran an analysis of variance (ANOVA) on our dependent variable—likelihood to choose the more responsible financial behavior (i.e., saving vs. spending)—with valence condition, construal level condition, and their interaction as the independent variables. There were no main effects for outcome elaboration valence, F1, 129 = .70, n.s., or outcome elaboration construal level, F(1, 129) = .32, n.s.. Importantly, there was the predicted significant interaction between the two conditions, F(1, 129) = 10.09, p < .01, as shown in Figure 1.
Study 2: Outcome elaboration valence and construal level interaction.
Additional analysis of the group means suggested that, in line with our predictions, participants were more likely to save their money when the positive elaboration was abstract (M = 7.55), as compared to when it was concrete, [M = 5.97], t(129) = −2.58, p < .01. However, the opposite was true for the negative outcome elaboration—participants were more likely to save their money when the negative outcome elaboration was concrete (M = 6.85), as compared to when it was abstract, [M = 5.86], t(129) = 1.89, p < .06. These results confirm our prediction that a match between outcome elaboration valence and level of construal promotes more effective self-control.
Study 3
While Study 2 shows the conditions under which valence and construal level of outcome elaboration have the most beneficial effects on self-control, we have yet to empirically establish the reason for these effects. The objectives of Study 3 are to both replicate our previous results and examine the underlying process.
We expect that a match between the outcome valence and construal level would give rise to a fluency experience that would hence promote more effective self-control. As suggested above, processing fluency generates feelings of ease and this ease influences decision making (Bandura, 1997; Carver & Scheier, 1998; Schwarz, 2004) making related outcomes, including those involving enhanced self-control, seem more likely to occur. As such, we expect that increased fluency of outcome elaboration that is consistent with long-term goal achievement should increase the perceived likelihood of outcome occurrence thereby increasing self-control effectiveness (Novemsky et al., 2007). Specifically, negative outcome elaboration will be perceived as easier when concrete, whereas positive outcome elaboration will be easier when abstract. Therefore, these pairings will result in a greater perceived ease of processing and enhanced goal-directed behavior.
Method
Participants and Design
A total of 151 undergraduate students (47% female; M age = 21.46) completed this experiment for course credit in a computer lab with privacy partitions. Participants were randomly assigned to one of the four experimental conditions in a 2 (outcome elaboration valence: positive vs. negative) × 2 (outcome elaboration construal: abstract vs. concrete) between-subjects design.
Outcome Elaboration Instructions
We utilized the same deliberative mind-set procedure as in Study 2.
Ease of Outcome Elaboration Assessment
After participants completed the outcome elaboration task, they were given three measures assessing the ease of outcome elaboration. These included two measures assessed on a 9-point scale ranging from 1 = strongly disagree to 9 = strongly agree: “Coming up with the consequences was easy”; “I had trouble coming up with potential consequences” (reversed), as well as one measure assessed on a 1 to 9 scale ranging from 1 = not at all difficult to 9 = very difficult: “How difficult was it to come up with the consequences that you listed in the thought exercise?” (reversed).
Self-Control Assessment
Following the outcome elaboration task, participants were provided with the same financial management scenario as in Study 2 and were asked to indicate their preference for either the more responsible or the more indulgent financial behavior on a 9-point scale ranging from will definitely spend now to will definitely save for later.
Manipulation Check
After participants completed the outcome elaboration task, we again checked whether our outcome elaboration manipulations successfully focused participants on the abstract, concrete, positive, or negative outcomes, using the same measures as in Study 2. Finally, gender and age measures were collected.
Results and Discussion
We confirmed that both the valence, 1 = positive, 7 = negative; M positive = 1.85, M negative = 5.79; t(149) = 20.48, p < .001, and construal, 1 = very specific, 7 = very general; M abstract = 4.59, M concrete = 3.19; t(149) = −5.50, p < .001, manipulations were effective. As in Study 1, we collapsed across the two self-control domains. We then ran an ANOVA on our dependent variable, relative preference for putting the money aside in order to save it for later, with valence, construal level, and their interaction as the independent variables. There were no main effects for outcome elaboration valence, F(1, 149) = .10, n.s., or outcome elaboration construal level, F(1, 149) = .01, n.s. Importantly, there was a significant interaction between the two conditions, F(1, 149) = 9.36, p < .01, as shown in Figure 2.
Study 3: Outcome elaboration valence and construal affect likelihood to save.
Additional analysis of the group means suggested that, in line with our predictions, participants were more likely to save their money when the positive elaboration was abstract (M = 7.14), as compared to when it was concrete (M = 6.18), t(149) = −2.30, p < .05. However, the opposite was true for the negative outcome elaboration—participants were significantly more likely to save their money when the negative outcome elaboration was concrete (M = 7.11), as compared to when it was abstract (M = 6.25), t(149) = 1.99, p < .05. These results confirm our contention that a match between outcome elaboration valence and level of construal promotes more effective self-control.
Next, we seek to examine whether or not the ease of processing experienced from the match between concrete construal and negative outcomes and abstract construal and positive outcomes explains the effects on the likelihood to exercise self-control in the current situation. We created a single ease of outcome generation score from the three measures (α = .88). We then ran an ANOVA on this score with valence condition, construal condition, and their interaction as the independent variables. There were no main effects for outcome elaboration valence, F(1, 149) = .61, n.s., or outcome elaboration construal level F(1, 149) = .01, n.s. Importantly, there was a significant interaction between the two conditions, F(1, 149) = 10.25, p < .01, as shown in Figure 3.
Study 3: Valence and construal affect outcome elaboration ease.
Additional analysis of the group means suggested that, in line with our predictions, participants found the outcome elaboration task significantly easier when the positive elaboration was abstract (M = 6.27), as compared to when it was concrete (M = 5.31), t(149) = −2.35, p < .05. However, the opposite was true for the negative outcome elaboration—participants found the task significantly easier when the negative outcome elaboration was concrete (M = 6.02), as compared to when it was abstract (M = 5.14), t(149) = 2.13, p < .05.
We then examined mediation using the bootstrapped estimation of conditional indirect effects (Preacher & Hayes, 2004). The estimated 95% confidence interval around the indirect effect of the interaction between the valence and construal level experimental conditions on saving intentions does not contain zero (.0015 to .2218), supporting mediation. That is, the differences in self-control outcomes are explained by the increased ease of processing experienced when the valence and construal level of the future outcomes considered are combined in the most optimal manner.
General Discussion
Many failures to sufficiently exercise self-control have been attributed to an overfocus on the present and a corresponding disregard for the future consequences of one’s behavior. Therefore, people are often encouraged to carefully consider the potential outcomes of their current behaviors. The current research reveals that there is more to the story than simply looking ahead, and the effectiveness of outcome elaboration as a self-control strategy depends on the valence and construal level of the future outcomes considered and the fluency experienced from the elaboration task. As such, we are able to provide new theoretical insights into the process of considering the future as well as clear recommendations for individuals regarding ways to maximize the effectiveness of their attempts to think about the potential consequences of their actions.
In our first study, we use participants’ own elaboration on potential future outcomes to establish the naturally occurring tendency for positive outcome elaboration to be more abstract and long-term and for negative elaboration to be more concrete and short-term. We also utilize the linguistic categorization model (Semin & Fiedler, 1988) to add further support to this finding. Next, in Studies 2 and 3, we examine how implementing outcome elaboration techniques using these naturally occurring matches among valence and construal level can lead to more successful self-control outcomes. In Study 3, we also provide evidence that it is the greater perceived ease of processing arising when the valence and construal level of outcome elaboration are combined in the most optimal manner that makes the matched outcome elaboration more effective.
The current research makes four important theoretical contributions. First, it contributes to the literature on outcome elaboration (Nenkov et al., 2008) and predecision deliberation (Gollwitzer, 2011) by demonstrating what forms of outcome elaboration most effectively promote self-control and identifying an important mechanism driving outcome elaboration’s effectiveness. Specifically, we find that both positively and negatively valenced elaboration on potential outcomes can be effective for enhancing self-control, but that their effectiveness depends on the level of construal at which potential outcomes are considered. Therefore, encouraging positive elaboration is more effective when more abstract outcomes are emphasized (e.g., “you will feel more relaxed and confident in your retirement days”), whereas negative elaboration is more effective when concrete outcomes are encouraged (“you won’t be able to pay the minimum payment on your credit card bill this month and will be assessed extra fees”). More appropriate use of future outcome elaboration based upon these valence and construal level congruencies can reduce self-control failures.
Second, our research contributes to recent work that has started to explore the interplay of valence and construal level in action consideration (Eyal et al., 2004; Herzog et al., 2007) by demonstrating both the natural associations between positive and abstract and negative and concrete thinking in considering the future. Third, the present research advances theory on construal level, as it relates to the successful operation of self-control by showing that not only abstract (e.g., Fujita et al., 2006) but also concrete construal level can have beneficial effects on self-control effectiveness. Finally, we contribute to the fluency literature (Schwarz, 2004) by demonstrating that a type of processing fluency pertaining to the ease with which relevant potential outcomes are generated has important implications for the effective exertion of self-control, extending the typically past-focused effects of fluency into the domain of future intentions and actions.
Overall, our research sheds light on how the interplay of valence and construal level of outcome elaboration affects present self-control decisions. It seems that future-oriented approaches to enhancing self-control are not a “one-size-fits-all” type of proposition, so rather than advising people to “look before they leap,” we might be better off, suggesting that they look at the specific hazards that they might be leaping into or consider the overall benefits they can reap by not leaping.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received the following financial support for the research, authorship, and/or publication of this article: The second author gratefully acknowledges financial support for this research from the Mays Business School, Texas A&M University.