https://orcid.org/0000-0003-1767-0265
Abstract
Precommitment shows promise in reducing decision-making impulsivity; however, it may be invalidated, leading to self-control failures. Therefore, this study aimed to explore the effectiveness of revocable precommitment in promoting commitment and reducing decision-making impulsivity through two studies. Experiment 1 used event-related potentials to explore whether revocable precommitment was beneficial to make individuals more inclined to commit. Experiment 2 explored the effectiveness of revocable precommitment in reducing decision-making impulsivity. The results showed that (i) compared with the precommitment condition, there is a higher proportion of precommitment selection under the revocable precommitment condition. Besides, in short delay time condition, the proportion of individuals choosing precommitment is higher than that of the other two conditions. Additionally, the average amplitudes of N1 and P300 in the revocable precommitment condition are significantly lower than those in the precommitment condition. (ii) There is a higher proportion of choosing larger-later (LL) options in the revocable precommitment condition compared with the other conditions. Moreover, the proportion of choosing LL options in short-term conditions is higher than that in medium-term conditions, which is higher than that in long-term conditions. Our findings indicated the efficiency of revocable precommitment in reducing decision-making impulsivity.
Introduction
Decisions between outcomes available at different points in the future often require inhibiting impulsivity to maximize rewards, and how to make far-sighted decisions has been focused in a growing body of research (Peters and Büchel 2010, Wang and Dvorak 2010). Currently, extensive evidence has proposed the efficiency of precommitment in reducing decision-making impulsivity (Kurth-Nelson and Redish 2010, 2012, Crockett et al. 2013, Studer et al. 2019, Ljusic et al. 2023). However, people often succumb to the immediate temptations after making precommitment in real life, which indicates the clear need for valuable strategies in reducing decision-making impulsivity.
Delay discounting is the core performance of decision-making impulsivity (Myerson et al. 2017, Zhou et al. 2022), which refers to the phenomenon that individuals always tend to give less weight to future gains than immediate gains (Kable and Glimcher 2007, Wang et al. 2023). The delay discounting rate could get bigger with the increase of delay time (Kable and Glimcher 2007, Kalenscher and Pennartz 2008, Myerson et al. 2017, Wang et al. 2019). Broadly, there is accumulating evidence that people always tend to use precommitment strategies to make short-term choices, so that they can resist the temptation of immediate options and reduce decision-making impulsivity (Webb and Sheeran 2003, Layton 2014, Studer et al. 2019).
Precommitment means that individuals make far-sighted decisions in advance to avoid choosing temptation options (usually irrevocable after making commitments) and reduce self-control failure in subsequent decisions (Crockett et al. 2013, Layton 2014, Engebø et al. 2022). However, recent research has shown that precommitment may get invalidated (Baumgartner et al. 2009, Breman 2011). Two recent studies on promised repayment and tax rebate savings confirmed that precommitment may be invalidated during commitment fulfillment (Calluso et al. 2018, Roll et al. 2019). Given that, is there a more effective strategy that can help to reduce decision-making impulsivity? According to the self-control resource model (Baumeister 2002, Baumeister et al. 2007), self-control resources are limited. Irrevocable precommitment increases the risk of depleting these resources and subsequently leads to commitment failure. In contrast, does revocable precommitment effectively reduce consumption of self-control resources and enhance commitment effectiveness? If so, what is the brain mechanism involved in this process?
In order to explore the vital role of recoverable precommitment in making effective commitments, the event-related potential (ERP) technique could be introduced to uncover the robust mechanism of the above-committing process. Considering that precommitment involves processes such as self-control resource consumption and goal trade-off (Baumeister et al. 2007, Baumgartner et al. 2009), this study focused on the related amplitudes of N1 and P300 ERP components. N1, a negative ERP component occurring 100 ms after stimulation, reflects recognition of decision-making conflict situations, and the amplitude of N1 induced by conflict situations is larger than those induced by mutually beneficial situations (Boudreau et al. 2009). Compared with revocable precommitment, further self-control resources are needed to weigh between long-term goals and short-term temptations when making precommitment, and thus the N1 amplitude increases larger in precommitment. P300 is a positive ERP component that usually occurs ∼300–500 ms following the presentation of a stimulus and is related to cognitive resources. There is a wealth of literature from observational studies that demonstrates the positive correlation between cognitive resource investment and the amplitude of P300 (Christie et al. 2013). In addition, P300 is related to individual impulse and control (Martin and Potts 2009). Compared with healthy individuals, individuals faced with difficulties in self-control (e.g. addicted people) showed a larger increase in P300 amplitude in impulse control tasks (Dong and Zhou 2010). As precommitment represents a strategy that requires the support of self-control resources, the amplitude of P300 may be greater when individuals make precommitments (Ariely and Wertenbroch 2002, Christodoulou et al. 2010, Crockett et al. 2013, Layton 2014, Brevers et al. 2016).
Although previous studies found that precommitment could be helpful in reducing decision-making impulsivity (Kurth-Nelson and Redish 2010, 2012, Crockett et al. 2013, Studer et al. 2019), it still remains limited in efficacy. Is revocable precommitment associated with greater effectiveness in reducing decision-making impulsivity compared to regular precommitment? Specifically, although individuals are likely to limit the time and amount, they gamble beforehand, the actual time and amount always exceed, showing an increase in decision-making impulsivity (Lalande and Ladouceur 2011). Meanwhile, based on the self-control resource model, self-control resources’ premature depletion may further contribute to the failure of self-control in subsequent tasks (Christodoulou et al. 2010) and finally lead to an increase in decision-making impulsivity (Baumeister et al. 2007). However, in the task of revocable precommitment, fewer self-control resources are consumed when committing, enabling individuals to effectively resist temptation and reduce decision-making impulsivity.
To this end, we aimed to examine the role of revocable precommitment in promoting precommitment making and reducing decision-making impulsivity through two studies. Specifically, we first explored the efficiency of revocable precommitment on commitment making through ERP in Experiment 1 and revealed the effectiveness of revocable precommitment in reducing cognitive resource consumption by observing the changes of N1 and P300 amplitudes. We hypothesized that individuals prefer to make commitment in the revocable precommitment task, rather than in the precommitment task. In addition, compared to a revocable precommitment task, a greater magnitude of the amplitude of N1 and P300 may be shown when individuals engage in precommitments within a precommitment task. To verify the effectiveness of revocable precommitment in reducing decision-making impulsivity, Experiment 2 further explored the effect of revocable precommitment on intertemporal decision-making. We hypothesized that the impact of revocable precommitment on reducing decision-making impulsivity surpasses that of precommitment alone.
Experiment 1
Methods
Participants and power analysis
A prior power analysis (1 − β = 0.80, α = 0.05, f = 0.25) indicated that the sample size is 19 participants. In order to ensure the quality of the experiment, a total of 24 participants from a university in China were involved in the experiment, of which three with poor data quality due to fatigue were not included in the statistical analysis. Finally, there were 21 effective participants (14 females, Mage = 23.19, s.d.age = 0.93). All the participants were in good health, normal vision, and no brain injury or surgery history.
The participants signed informed consent forms and were told they could withdraw from the research at any time. All procedures in this study were in accordance with the ethical standards of the Academic Board of Shandong Normal University and the 1964 Helsinki Declaration and its later amendments.
Tasks
In this experiment, the precommitment intertemporal decision-making task and the revocable precommitment intertemporal decision-making task were designed, combining the restricted selection paradigm in precommitment (Crockett et al. 2013) with the dynamic intertemporal decision-making task (Xu et al. 2020). Participants who chose the delay option [larger-later (LL) option] in both tasks were required to wait in real time. The delay time (i.e. the waiting time of LL options) included 7 s (short delay time), 15 s (medium delay time), and 29 s (long delay time). Participants had to wait in real time if they chose the LL option and the waiting time was shorter than the previous settings of 1 week or 1 month. Therefore, the amount of the smaller-sooner (SS) option in this study was fixed at 50 cents, and the ratios between the LL option and the SS option were 110%, 120%, 130%, 140%, 150%, 160%, 170%, 180%, 190%, 200%, 210%, and 220%. There were 12 trials in each delay time, 36 trials in each task in total. The specific tasks used in this study were as follows (Fig. 1).
Task flow. Notes: Process A and Process B (Fig. 6).
Precommitment intertemporal decision-making task
There are three stages in this task: choose stage, delay stage, and reward stage. Before choose stage, participants need to decide whether to restrict their choice of SS options (i.e., whether to make a precommitment) in choose stage after 4 s. If they decide to restrict the SS option, participants could press the “J” key, else they could press the “F” key. Then, participants could only choose the LL option in the choose stage, waiting for the corresponding delay time after entering the “delay stage” and finally getting the delay reward; if not, participants can choose among the SS and LL options in the “choose stage,” in which if they choose the SS option by pressing the “F” key, they will be rewarded immediately; if they select the LL option by pressing the “J” key, they should wait for the corresponding delay time after entering the “delay phase” and finally get the delayed reward.
Revocable precommitment intertemporal decision-making task
The task flow is the same as the precommitment intertemporal decision-making task, except that participant can press the “F” key at any time to get an immediate reward in delay stage when they choose the LL option.
Procedur
After entering the EEG laboratory, researchers introduced the experimental process to participants, demonstrated the function of each button, and then asked participants to practice fully to ensure that they were familiar with the experimental process. At the end of the exercise, the formal experiment was conducted immediately. The formal experiment was divided into two blocks (pre-commitment task and revocable precommitment task), with a 5-min break between two blocks. Each block contains 108 trials, and there were 216 trials in total. Random half of participants involved in the procedure that precommitment task prior to the revocable precommitment task, the rest half of participants involved in the procedure that revocable precommitment task prior to the precommitment task. Participants were told to keep their heads as stationary as possible and reduce blinking during the experiment. At the end of the experiment, they were asked to fill in the demographic information.
ERP data acquisition and process
Raw ERP data were amplified using a 64-channel Quick Amp amplifier with the sampling rate of 500 Hz (Brain Products, Munich, Germany). One electrode was placed on the AFz electrode site as the ground and the recordings were referenced to FCz. The vertical electro-oculographic (EOG) activities were monitored from the electrode placed at the lower right-eye orbit, and the horizontal EOG activities were recorded from the electrode placed at the outer canthus of the left eye. The impedance of all electrodes was kept <10 kΩ during the ERP recording.
The toolbox of EEGLAB and letswave 7 on Matlab 2013b were used to process the offline data. The eye movement correction procedure was developed by Gratton et al. (1983), the stimulus-lock ERP was initially rereferenced to averaged mastoids (M1 and M2), and bad channels were interpolated using a spherical method. Continuous ERP was epoched from −500 to 1000 ms post the stimulus onsets, and baseline correction was proceeded using a −200 ms to 0 ms prestimulus time interval. Next, the epochs were filtered with a zero-phase shift low-pass filter of 30 Hz and a high-pass filter of 0.05 Hz, and those with incorrect responses or with artifacts exceeding ±100 μV were rejected.
The selection of the time windows, as well as the electrodes, used to derive P300 and N1, was based on previous research (Wu et al. 2016, Fang et al. 2019, Ni et al. 2019) and the visual inspection of the grand averaged ERP waveforms. The mean amplitudes of P300 and N1 were calculated from the time windows of 320–400 and 120–200 ms, respectively, following the stimulus onset. The final average N1 amplitude was derived by averaging N1 amplitudes at the PO7 and PO8 electrodes (Figs 2 and 3), and the final average P300 amplitude was derived by averaging P300 amplitudes at the Pz electrode (Figs 4 and 5).
Grand averaged ERP waveforms for the interaction between task types and delay time. (a) The N1 component at the PO7 electrode; (b) the N1 component at the PO8 electrode. Notes: Precom-s = precommitment—short delay time condition, Precom-m = precommitment—medium delay time condition, Precom-l = preccommitment—long delay time condition; Revocable-s = revocable precommitment—short delay time condition, Revocable-m = revocable precommitment—medium delay time condition, Revocable-l = revocable precommitment—long delay time condition.
The topographic distribution of the average N1 amplitudes (120–200 ms) for the interaction between task types and delay time.
Grand averages of P300 component for the interaction between task types and delay time at the Pz electrode.
The topographic distribution of the average P300 amplitudes (320–400 ms) for the interaction between task types and delay time.
Data analysis
This study conducted a within-subjects design with the task types and delay time as within-subjects factors. Behavioral (the proportion of precommitment) and ERP outcomes (i.e. N1 and P300 amplitudes) were analyzed using a two-way repeated-measure analysis of variance (RM ANOVA): 2 (task types: precommitment intertemporal decision-making task vs. revocable precommitment intertemporal decision-making task) × 3 (delay time: short, medium vs. long). A Greenhouse–Geisser correction was used when the sphericity assumption was violated. The univariate ANOVA and paired t-tests with Bonferroni’s corrections were used in subsequent analyses. Partial eta-squared (ηp2) was used to indicate small (0.01–0.059), medium (0.06–0.139), and large (>0.14) effect sizes (Cohen 1973). Statistical analyses were carried out using SPSS 19.0 with an alpha set at P < .05.
Results
Behavioral results
The result of the two-way ANOVA revealed that the main effect of task types was significant, F(1, 38) = 5.449, P = .031, ηp2 = 0.223. The proportion of making commitment was higher in the revocable precommitment group (M = 0.44, s.d. = 0.38) than in the precommitment group (M = 0.38, s.d. = 0.34). The main effect of delay time was significant, F(2, 38) = 19.771, P < .001, ηp2 = 0.51. The proportion of precommitment in short delay discounting was the highest (M = 0.59, s.d. = 0.34), followed by medium delay discounting (M = 0.39, s.d. = 0.34) and long delay discounting (M = 0.24, s.d. = 0.31). The interaction between task types and delay time was significant, F(2, 38) = 5.224, P = .01, ηp2 = 0.22. Furthermore, simple effect analysis showed that the proportion of precommitment in the revocable precommitment group was higher than the precommitment group in medium delay time condition, while there were no significant differences in short delay time and long delay time conditions.
ERP results
N1 amplitudes
The result of two-way ANOVA revealed that the main effect of task types was significant, F(1, 20) = 7.49, P = .01, ηp2 = 0.27. Compared with the revocable precommitment group, the precommitment group showed a larger N1 amplitude. The main effect of delay time and the interaction effect between task types and delay time were not found, F(2, 40) = 0.39, P = .68; F(2, 40) = 0.26, P = .74 (Table 1).
The descriptive statistics of task types and delay time for N1 amplitude (M ± s.d.).
| Task types | Short delay time | Medium delay time | Long delay time |
|---|---|---|---|
| Revocable precommitment | −1.18 ± 2.29 | −0.95 ± 2.27 | −0.96 ± 2.20 |
| Precommitment | −1.50 ± 2.27 | −1.55 ± 2.53 | −1.44 ± 2.46 |
| Task types | Short delay time | Medium delay time | Long delay time |
|---|---|---|---|
| Revocable precommitment | −1.18 ± 2.29 | −0.95 ± 2.27 | −0.96 ± 2.20 |
| Precommitment | −1.50 ± 2.27 | −1.55 ± 2.53 | −1.44 ± 2.46 |
The descriptive statistics of task types and delay time for N1 amplitude (M ± s.d.).
| Task types | Short delay time | Medium delay time | Long delay time |
|---|---|---|---|
| Revocable precommitment | −1.18 ± 2.29 | −0.95 ± 2.27 | −0.96 ± 2.20 |
| Precommitment | −1.50 ± 2.27 | −1.55 ± 2.53 | −1.44 ± 2.46 |
| Task types | Short delay time | Medium delay time | Long delay time |
|---|---|---|---|
| Revocable precommitment | −1.18 ± 2.29 | −0.95 ± 2.27 | −0.96 ± 2.20 |
| Precommitment | −1.50 ± 2.27 | −1.55 ± 2.53 | −1.44 ± 2.46 |
P300 amplitudes.
The result of two-way ANOVA revealed that the significant main effect of task types was significant, F(1, 20) = 5.28, P = .03, ηp2 = 0.21, with a larger P300 amplitude for the precommitment group compared with the revocable precommitment group. There were no significant main effect of delay time and significant interaction effect between task types and delay time, F(2, 40) = 0.54, P = .59; F(2, 40) = 1.58, P = .22 (Table 2).
The descriptive statistic of task types and delay time for P300 amplitude (M ± s.d.).
| Task types | Short delay time | Medium delay time | Long delay time |
|---|---|---|---|
| Revocable precommitment | 3.56 ± 2.45 | 3.35 ± 2.67 | 3.53 ± 2.64 |
| Precommitment | 4.05 ± 2.48 | 4.62 ± 2.34 | 4.04 ± 1.82 |
| Task types | Short delay time | Medium delay time | Long delay time |
|---|---|---|---|
| Revocable precommitment | 3.56 ± 2.45 | 3.35 ± 2.67 | 3.53 ± 2.64 |
| Precommitment | 4.05 ± 2.48 | 4.62 ± 2.34 | 4.04 ± 1.82 |
The descriptive statistic of task types and delay time for P300 amplitude (M ± s.d.).
| Task types | Short delay time | Medium delay time | Long delay time |
|---|---|---|---|
| Revocable precommitment | 3.56 ± 2.45 | 3.35 ± 2.67 | 3.53 ± 2.64 |
| Precommitment | 4.05 ± 2.48 | 4.62 ± 2.34 | 4.04 ± 1.82 |
| Task types | Short delay time | Medium delay time | Long delay time |
|---|---|---|---|
| Revocable precommitment | 3.56 ± 2.45 | 3.35 ± 2.67 | 3.53 ± 2.64 |
| Precommitment | 4.05 ± 2.48 | 4.62 ± 2.34 | 4.04 ± 1.82 |
The above results revealed that individuals showed a higher proportion of precommitment in the revocable precommitment task rather than precommitment tasks, which were also shown in the ERP results. More specifically, no matter the N1 component or the P300 component, the amplitudes of them induced in the precommitment task are larger than that in the revocable precommitment task, which indicates that in the precommitment task, individuals consume more cognitive resources and experience greater cognitive conflict when making a commitment. Summing up the above results, it can be seen that the revocable precommitment task is more conducive to promoting individual commitment making, in which individuals consume less resources and experience less conflict. In this case, does revocable precommitment have the same effect on reducing individual decision-making impulsivity? To make further exploration, Experiment 2 was conducted to explore the effect of revocable precommitment on reducing the decision-making impulsivity.
Experiment 2
Methods
Participants and power analysis
A prior power analysis (1 − β = 0.80, α = 0.05, f = 0.25) indicated that the number of participants needed for the analysis of repeated measurement variance is 124. A total of 134 participants from a university in China were involved in the experiment (104 females, Mage = 18.73, s.d.age = 1.15). All the participants are in good health and normal vision.
The participants signed informed consent forms and were told they could withdraw from the research at any time. All procedures in this study were in accordance with the ethical standards of the Academic Board of Shandong Normal University and the 1964 Helsinki Declaration and its later amendments.
Tasks
There were tasks included in Experiment 2, which were the revocable intertemporal decision-making task, precommitment intertemporal decision-making task, and revocable precommitment intertemporal decision-making task. Each task contains 36 trials that randomly present. (Fig. 6)
Task flow.
Intertemporal decision-making task
Participants were told to press the space bar and wait for 4 s before making a decision (matching the choose stage of the precommitment task). After waiting, they should choose between the SS option by pressing the “F” key and the LL option by pressing the “J” key. Participant would be rewarded immediately if the SS option is selected, if the LL option is selected, participants will enter the delay stage after making the choice, and the delay reward will be obtained at the end of the delay stage.
Revocable intertemporal decision-making task
After selecting the LL option, participants enter the delay stage, and they can press the “F” key to change their choices at any time, so as to get an immediate reward. The rest of procedures are the same as intertemporal decision-making task.
Precommitment intertemporal decision-making task.
It is same as Experiment 1.
Revocable precommitment intertemporal decision-making task
It is same as Experiment 1.
Procedure
This experiment was carried out in a quiet laboratory. After entering the laboratory, researchers introduced the experimental procedure to the participants, demonstrated the function of each button, and then asked the participants to practice until they ensured to be familiar with the experiment process. After exercising, the formal experiment was conducted immediately. At the end of the experiment, experimenters asked participants to fill in the demographic information.
Data analysis
The study employed a mix design with the task types as between-subjects factor and delay time as within-subjects factor. Behavioral results (the proportion of precommitment and the proportion of LL options) were analyzed using a two-way RM ANOVA: 4 (task types: intertemporal decision-making task, revocable intertemporal decision-making task, precommitment intertemporal decision-making task vs. revocable precommitment intertemporal decision-making task) × 3 (delay time: short, medium vs. long). A Greenhouse–Geisser correction was used when the sphericity assumption was violated. The univariate ANOVA and paired t-tests with Bonferroni’s corrections were used in subsequent analyses. Partial eta-squared (ηp2) was used to indicate small (0.01–0.059), medium (0.06–0.139), and large (>0.14) effect sizes (Cohen 1973). Statistical analyses were carried out using SPSS 19.0 with an alpha set at P < .05.
Results
The proportion of precommitment
The result of two-way ANOVA revealed that the main effect of task types was significant, F(1, 63) = 5.089, P = .03, ηp2 = 0.08, with a higher proportion of making precommitment for the revocable precommitment group (M = 0.55, s.d. = 0.36) than the precommitment group (M = 0.41, s.d. = 0.33). Also, there was a significant main effect of delay time, F(2, 126) = 107.47, P < .001, ηp2 = 0.63. The proportion of precommitment in short delay time was the highest (M = 0.70, s.d. = 0.31), followed by medium delay time (M = 0.49, s.d. = 0.32) and then long delay time (M = 0.26, s.d. = 0.24). The interaction between task types and delay time was significant, F(2, 126) = 3.56, P = .04, ηp2 = 0.053. Furthermore, simple effect analysis showed that the proportion of precommitment in the revocable precommitment group was higher than that of the precommitment group in short and medium delay time conditions, while there was no significant difference in long delay time condition.
The proportion of LL options
The result of two-way ANOVA revealed that the main effect of task types was significant, F(3, 130) = 2.78, P = .04, ηp2 = 0.06. The Bonferroni post hoc test showed that the proportion of LL options for the revocable precommitment group (M = 0.60, s.d. = 0.21) was higher than the other three groups (P < .05). There were no significant differences among the precommitment intertemporal decision-making group (M = 0.50, s.d. = 0.31), intertemporal decision-making group (M = 0.48, s.d. = 0.29), and revocable intertemporal decision-making group (M = 0.48, s.d. = 0.33) (P > .05). There was a significant main effect of delay time, F(2, 260) = 305.378, P < .001, ηp2 = 0.701. The proportion of LL options in short delay time condition was the highest (M = 0.77, s.d. = 0.23), followed by medium delay time condition (M = 0.51, s.d. = 0.26) and long delay time condition (M = 0.27, s.d. = 0.24). The interaction between task types and delay time was not significant, F(6, 260) = 0.35, P = .909 (Fig. 7).
The descriptive statistic of task types and delay time for the proportion of LL options.
The results of Experiment 2 showed that individuals are more likely to choose the delay option in the revocable precommitment intertemporal decision-making task than that in the precommitment task, intertemporal decision-making task, and revocable intertemporal decision-making task, which indicates that the revocable precommitment task setting is more conducive to reducing individuals’ decision-making impulsivity.
General discussion
Although previous research has revealed the effect of precommitment on intertemporal decision-making, there is an invalidated phenomenon of the role of precommitment played in the process of intertemporal decision-making. Based on this, this study provided the first evidence to explore the role of revocable precommitment in promoting effective commitment making and reducing decision-making impulsivity.
The results revealed a significantly higher precommitment proportion in the revocable precommitment group compared to the precommitment group. Additionally, the N1 and P300 amplitudes were found to be significantly higher in the precommitment group than those in the revocable precommitment group. Specifically, N1 components are related to early visual discrimination processing, visual attention, and cognitive resource allocation (Vogel and Luck 2000, Wu and Zhou 2012, Wu and Wang 2014). However, the differences of revocable and irrevocable precommitment tasks in visual presentation were consistent, which indicated the explanation of the allocation of attention resources, rather than early visual discrimination processing and visual attention. Also, relevant research also supported the consistent changes between N1 amplitude and resources devotion to attentional goals (van der Lubbe and Woestenburg 1997). In this experiment, a crucial finding was that irrevocable precommitment induced a larger N1 amplitude than revocable precommitment. This may be attributed to the fact that individuals were unable to change their choices after making a precommitment. In precommitment task of the current study, participants can only choose the LL option and cannot change their decision. Participants might gather more information and consider more possibility before making precommitment in precommitment task (Crockett et al. 2013, Anjali et al. 2015) than in revocable precommitment task, which leading them to invest more attention resources in making trade-offs before making precommitments. For instance, when an individual makes a consequential decision that will alter the course of their life, they will allocate additional cognitive resources to information search and invest heightened attentional resources in evaluating the available alternatives. Another possible explanation for the higher precommitment proportion in the revocable precommitment group compared to the precommitment group is that participants may prefer the freedom and lack of constraint that a revocable precommitment provides. Compared with the restrictions on subsequent choices in the precommitment task, there are fewer restrictions in the revocable precommitment task, which could allow individuals feel freedom in subsequent choices, which may lead to a diminished allocation of attentional resources in the revocable precommitment task. In addition, consistent with relevant studies, it is especially vital to note that P300 is a component related to cognitive processing and control, and its amplitude depends on task-induced mental state and cognitive effort (Yancey et al. 2013, Fan et al. 2020). In this study, the irrevocable precommitment task induced a larger P300 amplitude than the revocable precommitment task, resulting in the fact that individuals in the precommitment group consumed more self-control resources when making precommitments. To sum up, compared with irrevocable precommitment task, revocable precommitment task made it easier for individuals to make trade-offs and consume less self-control resources, which in turn made individuals more inclined to make precommitments.
The observed significant differences in the proportion of delay options chosen across the revocable precommitment group, the precommitment group, and the revocable intertemporal decision-making group may be due to the fact that the revocable precommitment is less restrictive to their choices and consumes fewer resources in precommitment making. It allows individuals to retain more self-control resources in the subsequent decision-making stage to resist temptation of SS options and make more far-sighted decisions. Similarly, ERP results empirically supported that individuals in precommitment tasks consumed more cognitive and self-control resources, resulting in increased impulsiveness in intertemporal decision-making (Dou et al. 2014a). Also, the self-control resource model points out that limited self-control resources are used in different ways of self-control, such as behavioral control and psychological control. Once the process of self-control is triggered, the self-control resources would be temporarily depleted (also known as ego depletion), resulting in a decline in subsequent tasks involving self-control process (e.g. intertemporal decision-making) (Muraven and Baumeister 2000, Baumeister 2002, Baumeister et al. 2007). Strong evidence on ego depletion also found that high-wastage tasks could lead to the increase of delay discounting rate in intertemporal decision-making (Dou et al. 2014a, which confirmed the effectiveness of revocable precommitment in reducing decision-making impulsivity.
In line with previous studies, this study found that participants in the precommitment group showed more patience than those in the intertemporal decision-making group (Crockett et al. 2013, Studer et al. 2019). However, the results showed that the differences between these two groups were not significant, which was probably due to the experimental materials. Different from erotic pictures used in previous studies, the stimulating material used in this study is monetary stimulation, which was more universal and stable for individuals. In addition, in previous studies using erotic pictures, participants are required to rate on the attractiveness of stimulating pictures, and then the low attractiveness group would be considered as SS option material and the high group as LL option material. Participants only need to choose according to the attractiveness of materials in subsequent decision-making tasks (Studer et al. 2019), making it easier for them to make choices. In this study, money was used as a stimulus, the amount of the SS option was fixed, and the amount of the LL option was increasing, making participants weighed the amount and delay time of the two options at each trial, which increased processing load (Yu et al. 2020).
Finally, this study found that with the increase of delay time, the proportion of precommitment making decreased gradually, and the difference of the precommitment proportion making among different tasks became smaller. This may be due to the fact that with the increase of delay time, the effect on the proportion of precommitment making covers the role of task types. In addition, in line with previous studies, the proportion of choosing LL option decreases with the increase of delay time (Crockett et al. 2013, Studer et al. 2019), which also confirmed the fact that in the hyperbolic discounting model (Mazur 1984), as the waiting time increases, individuals’ subjective value of LL options decreases, which in turn increases individuals’ preference for SS options and leads to an increase in decision-making impulsivity.
In this study, no differences in the amplitude of N1 and P300 components were found at different delay time, which may be that the delay time is relatively short in this experiment compared with the waiting time in actual life, and the cognitive resources allocated and control resources consumed to make precommitment in different delay time might be not notably different. Therefore, the degree of conflict and self-control individuals experienced conducted in different delay time may not differ significantly.
Based on the above results, in the precommitment task, individuals exhibit greater caution in making precommitments due to their irrevocable nature. They allocate more self-control resources toward decision-making and experience heightened intensity of conflicts during the process. On the other hand, individuals who assigned in the revocable precommitment task consumed less self-control resources and experience fewer decision-making conflicts when making precommitment, so there were sufficient self-control resources to resist temptation in delay stage, which indicated that for intertemporal decision-making in the domain of money, revocable precommitment may be a more effective strategy to reduce individuals’ decision-making impulsivity.
Limitations and future directions
There are several limitations in our study that should be considered. First, given the particularity of precommitment, it remained difficult for participants to really wait when the influence of revocable setting on intertemporal decision-making is explored in the domain of money. Most studies available use erotic pictures as stimulating materials (Studer et al. 2019); however, when exploring the role of revocable precommitment in the intertemporal decision-making in the domain of money, there are some more strict requirements for the setting of time and amount in the experimental tasks. Future research should further explore the effect of revocable precommitment on intertemporal decision-making under real context, and strengthen the ecological validity of the influence of revocable precommitment on reducing individual decision-making impulsivity. Second, an argument is that the revocable option is set after a formal decision in this study, and each trial requires precommitment making and can be reversed. However, precommitment events in real life involve multiple decisions, such as dieting, learning, donations, and other tasks that take a long time to accomplish, in which revocation may occur in any one decision when achieving the precommitted goals. In light of this, future research should deeply explore the influence of different ways of precommitment and revocable setting on intertemporal decision-making. The third aspect is the social and cultural background of people from different countries. In the current study, we only explained to participants the method and function of restricting choice, but did not tell them that it was a “promise,” because the precommitment is not completely consistent with the promise, and the former put more emphasis on the limitation of external means without moral implication. In Chinese culture, “breaking a promise” could be shameful, which will lead to individuals’ negative moral experience. Therefore, individual personality traits, moral emotion, and social and cultural background may be important factors that will affect the role of revocable precommitment in reducing decision-making impulsivity. At last, consistent with relevant empirical studies (Baumeister 2002, Baumeister et al. 2007, Dou et al. 2014a, 2014b), the results of this study supported the resource model of self-control used in this study to produce hypotheses. However, there are some conflicting results, such as the brief of limited or unlimited willpower and the role of self-affirmation played in the process of ego-depletion (Schmeichel and Vohs 2009, Job et al. 2010). Considering that, we will further consider the role of factors such as personal willpower and self-affirmation in the depletion of self-control resources in future studies.
Conclusions
To conclude, this study found that compared with the irrevocable precommitment task, the revocable precommitment task could promote individuals to make precommitment. There were also differences in ERP amplitudes between the two tasks when making precommitment, and the average amplitudes of N1 and P300 components in the revocable precommitment task were lower than those of the precommitment task, which means that individuals expend more cognitive resources and encounter heightened cognitive conflict when making a commitment in the precommitment task. In summary, these findings indicate that the revocable precommitment task facilitates commitment formation by reducing self-control resource consumption and minimizing conflicts in making precommitment.
The chosen proportion of LL options in the revocable precommitment intertemporal decision-making task was higher than that in the precommitment intertemporal decision-making task, revocable intertemporal decision-making task, and intertemporal decision-making task. This suggests that the utilization of a revocable precommitment task setting is more effective in reducing decision-making impulsivity among individuals.
Conflict of interest
None declared.
Funding
This study was supported by the General Project of Education in 2021 for the 14th Five-Year Plan of the National Social Science Fund: The Development, Influencing Factors and Intervention System of Middle School Students’ Moral Shading: Based on Decision-Making Process Theory (grant no. BEA210108).
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
Author notes
Co-first author.
