Maximize when valuable: The domain specificity of maximizingdecision-making style

Abstract

The maximizing decision-making style describes the style of one who pursuesmaximum utility in decision-making, in contrast to the satisficing style, whichdescribes the style of one who is satisfied with good enough options. Thecurrent research concentrates on the within-person variation in the maximizingdecision-making style and provides an explanation through three studies. Study 1(N = 530) developed a domain-specific maximizing scale and found thatindividuals had different maximizing tendencies across different domains.Studies 2 (N = 162) and 3 (N = 106) further explored this mechanism from theperspective of subjective task value through questionnaires and experiments. Itwas found that the within-person variation of maximization in different domainsis driven by the difference in the individuals’ subjective task value inthe corresponding domains. People tend to maximize more in the domains theyvalue more. Our research contributes to a comprehensive understanding ofmaximization and provides a new perspective for the study of the maximizingdecision-making style.

1 Introduction

The maximizing decision-making style, based on the rational choice model, describes a style where one tends to pursue maximum utility in decision-making, whereas the satisficing style, based on the bounded rationality model, describes a style where one is satisfied with sufficient options due to complicated contexts and limited cognitive resources (Schwartz et al., 2002). Individuals with different maximizing decision-making styles perform differently in activities such as goal setting and alternative search, which in turn affect the process and results of one’s decisions. Apart from inter-person variation, one could also have different decision-making styles in different contexts. For example, a student may weigh every word in course descriptions repeatedly when choosing an elective course. However, the same student may put on a random shirt — the one that happens to be on top of the pile — heading to the class he carefully selected.

Decision-making style can vary either between persons or within a person. Traits describe behavior stably over a long period and on a wide range of occasions, while a typical person’s behavior can be highly variable across situations (Fleeson, 2004). Some researchers suggest that individuals choose different decision strategies based on their assessment of a specific field, namely, “decide how to decide” (Maldonato & Dell’Orco, 2011). For example, the intuitive-analytic decision-making style appears to be domain-specific and has significant differences across various domains (Pachur & Spaar, 2015).

In the present research, we aimed to investigate whether the maximizing decision-making style is domain-specific and explore its underlying mechanism. From the perspective of daily experiences, many people show inconsistent maximizing styles across domains¹. For example, some “A grade students” may pursue the best in academic performance but may lower their standard in food and clothing. On the other hand, “foodies” are picky in food they consume but may be indifferent to other issues. According to the expectancy-value model of motivation (Eccles, 1984), one’s motivation and performance in a task partially depend on their subjective assessment of the task. Therefore, owing to the different subjective value assessments of different domains, one may have different motivational effects and behaviors, forming different maximizing-satisficing decision styles.

1.1 Maximizing-satisficing decision-making style

Schwartz et al. (2002) proposed a maximizing-satisficing decision-making style based on differences in goal-setting when making decisions. They developed the Maximization Scale (MS) to explore individual differences in maximization, with higher scores indicating a higher maximizing tendency. People with a higher maximizing tendency who show greater eagerness towards seeking the best option are maximizers, whereas those with a lower maximizing tendency who are satisfied with good enough options are satisficers.

Previous research has mainly focused on the differences between maximizers and satisficers, including points of focus, performances, and outcomes when making decisions. Maximizers focus on outcomes rather than processes or effort (Chan, 2021; Hsieh & Yalch, 2020), and desirability rather than feasibility (Hsieh & Yalch, 2020; Luan & Li, 2017). Thus, compared to satisficers, maximizers are willing to invest more effort in searching for alternatives while trying to obtain the best outcomes (Dar-Nimrod et al., 2009; Weaver et al., 2015; Rim, 2017), which corresponds to two key components for maximization: the maximization goal of choosing the best and the strategy of alternative search (Cheek & Schwartz, 2016). They regarded the former as more central and the latter as the way to achieve the best goal. These characteristics can lead to positive outcomes, as they have been found to have a positive relationship with in-role performance (Giacopelli et al., 2013) and more job offers, better jobs, and higher salaries after graduation (Iyengar et al., 2006). However, it can sometimes lead to negative emotional experiences such as regret and decision dissatisfaction (Chowdhury et al., 2009; Iyengar et al., 2006; Leach & Patall, 2013; Sparks et al., 2012; Schwartz et al., 2002; Shiner, 2015).

With increased attention to the role of context, researchers have started to explore maximizing decision-making styles in specific domains. The concept of maximizing is no longer limited to the consumption domain alone, but gradually extends to romantic relationships (Mikkelson & Pauley, 2013; Mikkelson & Ray, 2020), friendship choices (Newman et al., 2018), career choices (Iyengar et al., 2006; Voss et al., 2019), academic achievement (Dahling & Thompson, 2013), military decisions (Shortland et al., 2020), moral issues (Goldsmith et al., 2018; Soltwisch, Brannon & Iyer, 2020), and other domains. To date, most studies still use general maximizing scales, with only a few having developed scales in specific domains (e.g., relational maximizing scale, RMS, Mikkelson & Pauley, 2013; Mikkelson & Ray, 2020; career maximizing scale, CMS, Voss et al., 2019). However, it may be necessary to explore maximization in specific domains. Maximizing tendency in one domain could predict performance in this domain more precisely than a general scale. Mikkelson and Pauley (2013) found that the RMS predicted relational outcomes better than the general maximization scale (Schwartz et al., 2002). Moreover, some domains are crucial for individuals, such as the academic domain for students and career domain for employees; therefore, it may be helpful to assess maximization in these domains.

1.2 Within-person variations of maximization

As summarized above, most researchers focused on the between-person variation of maximizing tendency, whereas little research concentrated on within-person variations, that is, whether people maintain the same maximizing decision-making style in their different contexts. Do maximizers pursue their best in each domain? Will satisficers be consistently satisfied with sufficiently good options? Maximizing tendencies are widely expected to vary across occasions and domains (Schwartz, 2004; Cheek & Schwartz, 2016). Schwartz (2004), one of the proposers of maximization concept, said that maximizing-satisficing orientations tend to be “domain specific”. For example, he was not a maximizer in most domains (e.g., shopping, investments), but he was indeed a maximizer in his “selected area” — consuming delicious foods. The exploration of within-person variations in maximization would help us to understand whether and why individuals maximize differently in different domains and to avoid imprecise predictions resulting from observations in other domains. Theoretically, this enriches the meaning and understanding of maximization.

Recent evidence has shown that maximizing transcends decision domains. Kokkoris (2019a) measured people’s intention to make the best choice in three large domains (consumer goods and services, experiences, and life decisions) consisting of 29 situations and found that general maximizing positively correlated with maximizing in most situations. Moyano-Díaz and Mendoza-Llanos (2021) expanded the range of domains by considering the health and finance domains and replicated the domain-spanning results. Mikkelson and Ray (2020) found that general maximizing is correlated with maximizing relationships. These findings showed that individuals higher in general maximizing tend to maximize in most specific decision domains.

However, meaningful within-person variations exist across occasions. For example, people are more likely to use a maximizing strategy when making material purchase (versus experiential purchase) decisions, as material choices are more comparable than experiential choices (Carter & Gilovich, 2010; Kokkoris, 2019a). When making decisions in private contexts, maximizers no longer behave like maximizers (Luan & Li, 2019). When making decisions for others, satisficers are more likely to use maximizing strategies (Luan et al., 2018). Furthermore, Ma and Roese (2014) situationally activated maximizing or satisficing mindsets, under which people exhibited the characteristics of maximizers or satisficers.

In addition, concepts that positively correlate with maximization have also been found to be domain specific. For example, perfectionism shares the common feature of pursuing high standards with maximization (Kokkoris, 2019b; Schwartz, 2002), which has been found to differ between sports and academic domains (Mcardle, 2010). Specifically, high-achieving students showed low perfectionism in the sports domain but high scores in the academic domain, indicating the domain specificity of perfectionism. Moreover, risk aversion is positively correlated with maximization (Lai, 2010). Risk-averse individuals are more inclined to search for alternatives (Parker et al., 2007), which is one of the main characteristics of maximizers. Researchers have found that individuals tend to show different risk-seeking or risk-averse preferences in different decision-making domains (Schoemaker, 1990; Weber et al., 2002).

1.3 The view of subjective task value

In this study, we investigate within-person variations of maximization that is founded on the idea that people are more likely to maximize in the domains that they value.

According to the choice-goal framework, people consider both maximizing utility and minimizing efforts when making decisions (Bettman, et al., 1998). Maximizers are more likely to give up the latter for the former, whereas satisficers are more likely to give up the former for the latter (Hsieh & Yalch, 2020; Luan & Li, 2017). Therefore, willingness to invest effort is one of the key determinants of people’s decision-making styles, which reflects their level of motivation to achieve a goal.

The expectancy-value model of motivation proposed by Eccles (1984) links personal choice, persistence, and performance to subjective task value, indicating that motivation and performance in a task partially depend on the individual’s subjective assessment of that task. The model outlines four motive components of the subjective task value. Attainment value assesses the importance of tasks, and tasks with high achievement values are related to self-realization. Intrinsic value measures whether a task meets one’s intrinsic interests, such as the pleasure acquired from completing the task. Utility value evaluates the perceived usefulness of the task, which is a measure of the task’s relevance to the individual’s current or future goals. Finally, cost describes the effort required to complete the task.

Assessment of subjective task value occurs before decision making and directly affects one’s motivation, choice, effort investment, and performance (Eccles, 2005). Motivation increases when the subjective value of a task is high (Eccles & Wigfield, 2002). This model emphasizes the motivational consequences of task value by constructing a path from subjective task value to motivation. People’s evaluation of task value affects their motive level, which subsequently affects their willingness to invest effort. Roberts, Shaddy and Fishbach (2021) found that when making decisions for liked things, people perceived a larger difference between different options and would wait longer for better quality. If people attach great value to a task, they spend more time and make greater effort for the task (Cole et al., 2008; Liem et al., 2008), whereas low motive levels can compromise the maximum utility for minimizing effort.

Therefore, motivation in a specific domain can be affected by one’s subjective value in that domain, which helps explain the domain specificity of maximization. If one’s subjective task value evaluation for a certain domain is high, it increases the motivation to persist and pursue higher goals, thus leading to more effort investment (Eccles, 2005; Wigfield & Eccles, 2020; Wu et al., 2020; Yildirim, 2012), resulting in a maximizing style. As mentioned above, high standards in goal pursuit and effort investment in alternative search are two important characteristics for maximizing decision-making styles. Therefore, this study proposes that subjective task value may partially explain why maximizing tendencies vary across different domains. For example, some academic masters could have a higher evaluation of task value in the academic domain, thus having higher motivation and maximizing characteristics in the academic domain, but this may not be the case in other domains.

1.4 Current Research

The current study attempts to explain the domain specificity of maximizing from the perspective of subjective task value in a student sample. Therefore, we first developed a tool for measuring the maximizing tendency in different domains for the student sample and preliminarily verified the domain-specificity of the maximizing tendency (Study 1). Then, we measured the subjective task values of different domains and used them to interpret maximizing tendencies in corresponding domains by regression analyses (Study 2). Finally, we investigated the validity of the scale and domain specificity using behavioral performance indicators of maximization (Study 3). The domain specificity of maximization proposed in this research contributes to the literature on within-person variation of maximization, and understanding the domain specificity from the perspective of subjective task value also provides a new view for subsequent research on maximizing decision-making styles.

2 Study 1

The main purpose of Study 1 was to verify the construct, convergent, and discriminant validities of the domain-specific maximization scale (for item generation details, see Supplement S1). Confirmatory factor analysis was conducted to examine the construct validity of the scale. We also included two related scales – the self-direction subscale of the simplified multidimensional perfectionism scale (Hewitt et al., 2008) and the regret scale (Schwartz et al., 2002) – to examine their correlations with the domain-specific maximization scale. Furthermore, this study provided a glimpse of the person-domain interaction based on self-report and preliminarily explored the domain specificity of maximizing decision-making style by comparing scores between different types of participants among different domains.

2.1 Method

2.1.1 Participants

In total, 331 students completed a paper-based questionnaire for financial compensation. Sixteen participants were excluded because they failed to correctly answer the attention check questions, leaving a final sample of 315 (186 males, M _age = 22.25 years, SD _age = 4.59).

2.1.2 Materials

The domain-specific maximization scale (Cronbach’s α = .928) developed in Study S1 (in the Supplement) contains six domains: relationship (Cronbach’s α = .767), study (Cronbach’s α = .812), food (α = .867), health (α = .713), clothing (α = .893), and travel (α = .873) — with a total of 29 items.

The self-direction subscale of the simplified multidimensional perfectionism scale (five items; Hewitt et al., 2008; α = .722) was employed as an external criterion. Self-oriented perfectionism refers to the belief that striving for perfection and being perfect are important. The concept of both maximization and perfectionism involves the pursuit of high standards (Kokkoris, 2019b), and many studies have demonstrated that they are related (Schwartz et al., 2002; Chang et al., 2011; Kokkoris, 2019b). Schwartz et al. (2002) used the self-direction subscale as an external criterion when compiling the Maximization Scale, and found a significant correlation (r = .25, p < .001). Each item is rated from 1 (strongly disagree) to 7 (strongly agree), with higher scores indicating higher perfectionism tendencies.

The Regret Scale (five items; Schwartz et al., 2002; α = .759) was employed as another external criterion. Since those who spend energy, time, or money to seek alternative paths may experience dissatisfaction or regret later (Moyano-Díaz et al., 2014), maximization could cause regret afterwards. Diab et al. (2008) used the scale as an external criterion when compiling the Maximizing Tendency Scale and found a significant correlation (r = .27, p < .001). Each item was rated from 1 (strongly disagree) to 7 (strongly agree), with a higher score indicating a higher sense of regret after the decision.

In addition, for exploratory purposes, we asked participants to indicate whether they valued study, entertainment, and beauty. Participants had to answer a multiple-choice question asking which domains (study, beauty, or entertainment) they valued in their daily life. They could choose either none or several options. We categorized the different types of person-domain relationships based on their responses. People who chose the corresponding domain were categorized as “value,” otherwise “not value”. The study domain was a special domain for the student sample. We wanted to explore whether people who value learning show more maximization in the study domain, but not in other domains. Moreover, whether people who place more value on food/clothing maximize in the corresponding domain but not in the study domain.

2.2 Results

2.2.1 Construct Validity

We conducted a confirmatory factor analysis with maximum likelihood estimation using Mplus 8.3 to verify the construct validity of the scale. The model provided an adequate fit to the data (χ ² = 805.078, df = 362, χ ², df = 2.224, RMSEA = .062, CFI = .905, TLI = .893, SRMR = .051; for detailed structure see Supplement S3.1). This structure also suggests a stronger correlation within the domain than between the domains.

Table 1 shows the correlations between the factors (ps < .01). The coefficients of the factor-to-factor correlation mostly varied between 0.18 and 0.6, and the coefficients of the factor-to-total correlation varied between 0.55 and 0.86. The average heterotrait-monomethod correlation (the average correlation between different factors; Harris et al., 1993) was 0.44, which shows a relatively moderate correlation compared to the average factor-to-total correlation of 0.73, indicating that factors were consistent in direction but with discrimination.

Table 1: Heterotrait-monomethod correlation matrix.

All correlations were significant at p < .001, except .18, which is .01.

As predicted, self-directed perfectionism and regret correlated significantly with the total scale and subscales (Table 2). The correlation coefficients between the total maximization score and perfectionism and regret were 0.47 and 0.35, respectively, which is higher than the 0.25 (Schwartz et al., 2002) and 0.27 (Diab, Hamilton & Schmidt, 2008) coefficients that were found in previous studies. Each subscale had a moderate correlation with perfectionism and a low-to-moderate correlation with regret.

Table 2: Correlations between subscales of maximization scale and perfectionism/regret scale

Notes: * p < 0.05, ** p < 0.01, *** p < 0.001.

2.2.2 Comparison of the average scores of different types of subjects in each field

Participants were divided into groups based on their answers to the question of whether they valued study, entertainment, or beauty. In each domain, an independent samples t-test was conducted to compare the maximizing tendencies of participants who valued study/entertainment/beauty and those who did not.

The results are shown in Figure 1. Except for the health domain (t = 1.61, df = 313, p = .109), students who valued study scored higher than those who did not (Figure 1a). Except for the study domain (t = –.24, df = 313, p = .809), students who valued entertainment scored higher than those who did not (Figure 1b). Except for the study domain (t = 1.582, df = 313, p = .115), students who valued beauty scored higher than those who did not (Figure 1c).

Figure 1: Multi-domain maximizing scores of different types of participants.

(a) Maximizing scores of those who values study vs. those who do not

(b) Maximizing scores of those who value entertainment vs. those who do not

(c) Maximizing scores of those who value beauty vs. those who do not

2.3 Discussion

Confirmatory factor analysis showed that the domain-specific maximization scale had high structural validity. The heterotrait-monomethod correlation showed good discriminant validity between the domain factors. The correlations between the scale and the self-directed perfectionism and regret scales also correlated with related concepts. Overall, the domain-specific maximization scale had good reliability and validity, and can be a reliable tool for use in subsequent studies.

The exploratory analysis preliminarily showed that maximizing tendencies varied among different domains. For example, participants who valued entertainment and beauty may be maximizers in travel, food, clothing, health, andrelationship domains but not in the study domain. Participants who valued study may be maximizers in study, travel, food, clothing, andrelationship domains but not in the health domain. These results, to some extent, support the domain-specificity of maximization in that a maximizer in one domain is not always a maximizer in another domain. Although some variations existed, the value and maximizing tendency did not precisely correspond. The correlations (Supplement Table S2.4) showed that people’s self-reports of whether they valued study/entertainment/beauty were significantly related to each other to a moderate degree, suggesting that people who valued (or did not value) each domain could overlap. This means that a person who values beauty could also value the entertainment/study domain, causing the results to be indistinguishable.

Interestingly, we found that gender was correlated with maximizing in the clothing domain (r = –.299, p < .001). Women were maximizing more than men in the clothing domain, which was quite reasonable because women often attached higher value on clothing domain.

However, the dichotomous measurement of whether one valued study/entertainment/beauty was only an exploratory trial, was not precise enough, and did not perfectly match the domains in the scale. Therefore, in Study 2, we introduced subjective task values to specifically measure the personal evaluation of the domain. In addition, we observed that someone could value multiple domains, whereas others could value none of them. To avoid overlapping effects of different domains, we compared two specific domains in subsequent studies.

3 Study 2

This study introduced a subjective task value to measure the personal evaluation of the domain, and attempted to explain the domain-specificity of maximization. The number of items that remained in the “clothing” and “study” domains in the domain-specific maximization scale was the most (six items), which are the two domains students frequently make daily decisions on. Therefore, these two domains were chosen for further exploration in this study, and the subjective task value scale was adapted to the two domains for the corresponding measurement.

3.1 Method

3.1.1 Participants

A total of 191 students completed the questionnaire on financial compensation. Twenty-nine participants were excluded because they failed to correctly answer the attention check question, resulting in a final sample size of 162 (94 females, M _age = 21.21 years, SD _age = 4.55).

3.1.2 Materials

Domain-specific maximizing tendency. The study (α = .851) and clothing (α = .851) subscales of the domain-specific maximization scale measured maximizing tendency in corresponding domains, with six items in each subscale. Each item was rated from 1 (strongly disagree) to 7 (strongly agree), with a higher score indicating a greater tendency towards maximizing decision-making style. An attention check item (“This question is used to test your attention, please calculate 2+3 and choose the correct answer”) was inserted into the questionnaire.

Subjective task value. Eight items were employed to assess the subjective task value of the participants (Supplement S6). Four of the items were related to the study domain (α = .784) and the other four pertained to the context of clothing (α = .780). These items measured participants’ subjective value evaluation of each domain, which included four aspects regarding intrinsic value (e.g., I think nice clothes is really appealing to me.”), utility value (e.g. “I think it’s useful to get good grades.”), attainment value (e.g. “I think it’s important to get achievement in studying.”), and cost (e.g., “I think studying well requires a lot of efforts, such as time, money, and energy.”). Each item is rated from 1 (strongly disagree) to 7 (strongly agree). These items have been used in previous research (Eccles, 1984; Mcardle, 2010) and proven to have excellent psychometric properties.

Demographic information, including age, education, sex, and major, was collected at the end of the questionnaire.

3.2 Results

3.2.1 The value of subjective tasks and the maximizing tendency of corresponding domains

Correlations suggested that the maximizing tendency in the clothing domain was significantly related to subjective task value in the clothing domain (r = .740, p < .001) but not the study domain (r = .141, p = .073), and maximizing tendency in the study domain was significantly related to subjective task value in the study domain (r = .79, p < .001) but not the clothing domain (r = .119, p = .132). People are maximizers only in their subjective valuable domains.

T-tests were also conducted for illustrative purposes. Participants were divided into two groups according to their subjective task value scores in the study and clothing domains, with the median as the split point (Moyano-Díaz & Mendoza-Llanos, 2021). An independent samples t-test was conducted to compare the maximizing tendency of participants with high and low subjective task values in the clothing and study domains. Results showed that participants with high subjective task value in clothing (M = 4.76) had a significantly higher maximizing tendency in the clothing domain than the low group (M = 3.48; t(160) = –8.49, p<.001, Cohen’s d = 1.34), but no significant differences in maximizing tendency in the study domain (high clothing value group: M = 4.73; low clothing value group: M = 4.65; t(160) = –.44, p = .658). Similarly, participants with high subjective task value in the study (M = 5.36) showed significantly greater maximizing tendency in the study domain than the low group (M = 3.97; t(160) = –10.39, p < .001, Cohen’s d = 1.64), with no significant differences in maximizing tendency in the clothing domain (high study value group: M = 4.33; low study value group: M = 4.00; t(154.34) = –1.96, p = .061).

3.2.2 Explanation of the maximizing tendency of various fields in the cognitive evaluation process

Hierarchical regression analyses were employed to explore the relationship between the subjective task value and domain-specific maximization. Two separate regressions were conducted for both the study and clothing maximization. Two blocks of independent variables were included for each analysis. In Step 1, gender, age, and education were entered to control for their effects. In Step 2, the domain-specific subjective task value was entered. Table 3 presents the results of the regression analyses. A significant change in R ² at Step 2 indicated that subjective task value predicted maximization in the study and clothing domains above and beyond demographic factors. In line with these expectations, the regression coefficients showed that the subjective task value of study (STV-Study) was a significant predictor of maximizing tendencies in the study domain but not in the clothing domain. Similarly, the subjective task value of clothing (STV-clothing) was a significant predictor of maximizing tendencies in the clothing domain, but not in the study domain.

Table 3: Hierarchical regression results of subjective task value on maximizing tendencies of clothing and study domains.

Notes: * p < .05, **p < .01, ***p <.001.

3.2.3 Elimination of common method bias

Since the variables of this study were collected using self-report questionnaires at the same time, Harman’s one-factor test (Podsakoff et al., 2003; Tehseen et al., 2017; Whitson et al., 2019) was conducted to rule out the possibility of common method bias. There is a common method bias if a single factor or any dominant factor accounts for the majority (50% or more) of the covariance of all research variables. We conducted principal component analysis of the variables concerned. The findings suggest that the main factor accounted for 32.63% of the total covariance. Therefore, common method bias was not strong enough to complicate our findings. We also conducted a random intercept item factor analysis suggested by Maydeu-Olivares and Coffman (2006) (Supplement S4) to capture how the common self-report method factor impacts items in both the STV and domain-specific maximizing tendency scale.

3.3 Discussion

The present study found that students with high subjective value of study had a greater maximizing tendency in the study domain, but not in the clothing domain, and vice versa. After controlling for the demographic factors, the results of the regression analyses showed that the subjective value of one domain was a good predictor of the maximizing tendency in the corresponding domain but not in other domains. This discrimination provides additional support for the domain-specific nature of maximization. At the same time, it offers an explanation of people’s differences in the subjective evaluation of different domains.

4 Study 3

In previous studies, we used a self-report measurement of maximization. In Study 3, behavioral indicators were used to measure maximization. On one hand, it could help verify the external validity of the questionnaire. On the other hand, it can also extend the conclusion to the performance level of maximization. The clothing domain used in Study 2 may have been significantly affected by gender. Thus, a more general domain, travel, was selected for comparison with the study domain in Study 3. The Subjective Task Value Scale was adapted to the two domains for the corresponding measurement.

4.1 Method

4.1.1 Participants

A total of 113 students participated in the study for financial compensation. Seven participants were excluded because they failed to correctly answer the attention check question, resulting in a final sample size of 106 (57 females, M _age = 22.24 years, SD _age = 2.62).

4.1.2 Procedure

We developed a web search paradigm to measure the participants’ maximizing tendencies in the travel domain. Participants were required to choose a travel destination for vacations. The instructions were as follows:

The National Day Golden Week holiday is approaching; you plan to visit New Zealand, and now you have to make travel plans. Please check the overview of the scenic spots on the website, select a few scenic spots you want to visit, and write down the name of the scenic spots you select in the file.

Notes:

1. (1) You have only seven days; please choose the spots you want to go to the most, about once per day.

2. (2) You do not need to make an itinerary; that is, assume that the distance between the scenic spots is 0.

3. (3) To ensure that all participants see the same information, please do not use any sorting or filtering functions.

Afterwards, the “Mafengwo” travel website (http://www.mafengwo.cn with New Zealand set as the destination was opened for the participants. They were allowed to browse New Zealand-related travel information by themselves. The time spent browsing before submitting their answers was recorded as a behavioral indicator of subjects’ maximizing tendency in the travel domain.

After submitting the answers, the participants were required to complete a questionnaire. The questionnaire included study and travel domain maximization scales (study: α = .896, travel: α = .860) and subjective task value scales (study: α = .717, travel: Cronbach’s α = .682) of the study and travel domains (for detailed items, see Supplement S6).

At the end of the scale, we collected GPA as a behavioral indicator of maximizing tendency in the study domain. Demographic information, including age, education, sex, and major, was also collected.

4.2 Results

4.2.1 Behavioral indicators in study and travel domains

Because the time of travel decision was not normally distributed, subsequent analyses were performed after logarithmic transformation of time. The self-reported maximizing tendency of the study domain was significantly correlated with GPA (r =.33, p =.001). The self-reporting maximizing tendency of the travel domain was significantly correlated with the participant’s browsing time before making decisions (r =. 31, p = .001). Correlation analyses showed that GPA and web browsing time were reliable indicators of maximizing tendencies in the study and travel domains.

4.2.2 The subjective task value and the maximizing performances in the corresponding domain

Similar to Study 2, correlations suggested that maximizing performance in the travel domain was significantly related to subjective task value in the travel domain (r = .359, p < .001), but not in the study domain (r = .170, p = .087), and maximizing performance in the study domain was significantly related to subjective task value in the study domain (r = .237, p < .001), but not in the travel domain (r = .074, p = .469). To explore whether people only show maximizing characteristics in domains with great subjective value, participants were divided into two groups according to their subjective task value scores in the study and travel domains, respectively. The medians were used as the split points. Results showed that participants with high subjective task value in travel (M = 2.85) spent more time browsing travelling website (maximizing indicator in travel domain) than the low group (M = 2.73; t(100) = –2.98, p = .004, Cohen’s d = 0.60). However, GPA (maximizing indicator in study domain) did not differ significantly (high travel value group: M = 3.45; low travel value group: M = 3.43; t(97) = –.28, p = .784). Similarly, participants who had a high subjective task value in study (M =3.51) had a higher GPA than the other group (M = 3.34; t(61.11) = –2.02, p = .048, Cohen’s d = 0.52), whereas there was no significant difference in travel website browsing time (high study value group: M = 2.83; low study value group: M = 2.74; t(100) = –1.90, p = .06).

Table 4 shows a simple analysis of the main result for the maximizing tendency (MTS) and subjective task value (STV) in each domain. For each participant, we asked whether MTS in the study domain (MTS-Study) was greater than, less than, or equal to MTS in the travel domain (MTS-Travel). We also asked this for subjective task value for the two domains (STV). The table shows the classification. Within the table, the classification (coded 1, 0, –1) in MTS is correlated with the classification in STV (r = .329, p = .001). Participants who value one domain more than the other are more likely to say that they maximize in that domain, more than the other.

Table 4: Numbers of participants in each category, classified by comparison of subjective task value (STV) in the travel and study domains and comparison of domain-specific maximizing tendency (MTS) in the two domains.

4.2.3 Explanation of the process of cognitive evaluation to the maximizing tendency of each domain

Hierarchical regression analyses were conducted as described in Study 2. Two separate regressions were conducted for both study and travel-maximizing performance. Gender, age, and education were entered to control for their effects in Step 1. In Step 2, domain-specific subjective task values were entered. Table 5 presents the results of the regression analyses. A significant change in R ² at Step 2 indicated that subjective task value predicted maximization in the study and travel domains after adjusting for demographic factors as measured. Consistent with the results of Study 2, the regression coefficients showed that the subjective task value of study (STV-Study) was a significant predictor of maximizing performance in the study domain, but not in the travel domain. Similarly, the subjective task value of travel (STV-Travel) was a significant predictor of maximizing performance in the travel domain, but not in the study domain.

Table 5: Hierarchical regression results of subjective task value on maximizing performances of travel and study domains.

Notes: *p < .05, **p < .01, ***p<.001

4.3 Discussion

We replicated the findings of Study 2 in a different domain, demonstrating the role of subjective task value in domain-specific maximization and expanded the findings from tendency to performance. Furthermore, we demonstrated that the performance indicators (GPA and travel planning decision time) successfully measured participants’ degree of maximization in the study/travel domain.

5 General Discussion

Our research explores the intra-individual variation in maximizing-satisficing decision-making style through three studies. The results confirmed the existence of domain specificity, meaning that individuals may show different maximizing tendencies in different domains. The decision-making style in a domain depends on its subjective value evaluation. People are more inclined to be maximizers in fields with high subjective value evaluations and satisficers in fields with low subjective values. Meanwhile, we demonstrated in a preliminary way how it is possible to devise a domain-specific maximization scale for the student samples.

5.1 Maximization: Within-person variations

Maximizing decision-making style has attracted the attention of researchers in recent years (Richardson et al., 2014). While existing literature focuses more on individual differences between maximizers and satisficers, researchers rarely investigate within-person variations of maximization. Our findings revealed one type of intra-person variation and proposed that people maximize more in the domain they value. Previous studies exploring within-person variations have found systematic variations (all people in one direction) in maximization among people. For example, people are maximizing more in material (versus experiential) purchases (Carter & Gilovich, 2010) and public (versus private) contexts (Luan & Li, 2019) when deciding for others (versus for self, Luan et al., 2018). However, the variation found in our research is not systematic, because it depends on people’s subjective perception of the domain, which is also the reason why Kokkoris (2019a) and Moyano-Díaz & Mendoza-Llanos (2021) failed to detect this domain specificity. Our finding is consistent with the view that decision-making style is flexible (Maldonato & Dell’Orco, 2011). A person showing maximization in a certain domain does not mean that the person is equally maximizing in all the domains. Moreover, the results from Mikkelson and Pauley (2013) are in line with our findings, which found that when predicting the outcomes of decision-making in romantic relationships, the relational maximization scale was more reliable than the general maximization scale.

Discussions on between-person and within-person variations in maximization are valuable. The findings on between-person variations contribute to identifying maximizers and satisficers and predicting related decision consequences. Exploration of within-person variations helps clarify whether and why a person maximizes differently in different domains, which leads to a deep integration of the situation and personality and contributes to the literature and explanation of maximization. Our findings, together with those of other maximizing studies in specific domains, highlight the necessity of developing a maximizing measurement in subdivided domains. When the concept of maximization is extended to various domains, considering domain specificity can help to draw more accurate conclusions.

However, our findings did not conflict with the use of a general maximization scale and score. We also found meaningful correlations among maximizing scores in different domains, as suggested by Kokkoris (2019a) and Moyano-Díaz & Mendoza-Llanos (2021). In addition, we conducted a bifactor model using data from Study 1 (for details, see Supplementary Material S5), and the index ECV and ω H suggested that the general maximizing tendency also played a non-negligible role. Ma and Roese (2014) primed participants’ maximizing mindset in a non-consumption domain that led to maximizing performance in the consumption domain, suggesting that maximization has a certain degree of commonality across domains. All evidence shows that the commonality and specificity of maximization coexist among the domains.

5.2 Subjective value perception: the key to choosing appropriate styles

Abundant studies have investigated characteristics and consequences of maximization (Cheek & Schwartz, 2016; Khare et al., 2021; Zhu et al., 2019) of maximizers. This study explores the factors that influence people’s strategic choices of maximization in different domains. Our research demonstrated that the domain specificity of maximization was partially derived from the differences in subjective task values in different domains. Individuals show a greater maximizing tendency in domains to which they attach greater value. According to the regression results of Studies 2 and 3, subjective value evaluations of the study, clothing, and travel domains all significantly predicted maximizing tendency and performance of the corresponding domain, but not of other domains. The maximizing style may be a decision strategy that people automatically choose after evaluating the decision-making domain, which is naturally domain specific.

Based on the expectancy-value model of motivation (Eccles, 1984), an individual’s motivation to complete a task depends on his or her subjective evaluation of the task, namely the subjective task value. If one evaluates the task in the domain as useful, interesting, or very important, that person develops strong motivation and would be willing to make more efforts for higher standards and thus choose the maximization strategy; conversely, when the subjective value is low, the person is likely to show unwillingness to make continuous efforts to pursue greater utility and thus choose a satisficing strategy. This kind of adjustment is adaptive to a certain extent because it allows individuals to selectively behave as satisficers in less important domains to save their cognitive efforts and resources. This is also in line with previous findings. Roberts et al. (2021) found that when people made decisions about things they liked more, they perceived a larger difference in subjective value between small-sooner and larger-later options. Thus, they were more willing to wait or pay for better quality things they liked. From the side of satisficers, when extrinsically motivated (e.g., by liking), they changed their “good enough” minimum threshold and preferred distinct evaluative information (Smallman & Becker, 2017). Luan and Li (2019) found that maximizers are willing to put in more effort than satisficers in public situations but not in private situations. From the perspective of subjective task value, the existence of social comparison in a public situation strengthens the achievement and utility values of the task. As maximizers are more sensitive to social comparison (Schwartz et al., 2002; Weaver, 2015), they are motivated to make efforts.

5.3 Theoretical and practical contributions

The theoretical contributions of our research are mainly reflected in the following aspects. First, we responded to the question of whether there is domain specificity in maximization (Cheek & Schwartz, 2016) and showed that the maximizing decision-making style had within-person variation across domains. Our study showed the existence of the domain specificity of maximization. Thus, to draw more accurate conclusions when exploring problems in a specific domain, it is necessary to pay more attention to the domain of interest when developing scales. Second, we found that people have high motivation to adopt maximizing strategies in the domain they assign high value to. The high explanatory power of subjective value for maximizing tendency and performance could help researchers to understand and explore maximizing strategies from the perspective of cognitive evaluation, providing new directions for subsequent research.

The link between subjective value and maximization is also significant in practice. Our results indicated that maximization depends on individuals’ perception of domain values, suggesting that individuals can change their maximizing tendency and performance in a domain as their subjective value evaluations of the domain change. The maximizing tendency may lead to good consequences (better performance, achievement, etc.), but they may also experience poor emotional experiences (decision difficulty, higher regret, lower satisfaction, etc.). Therefore, when we seek better performance, we can try to enhance the subjective value of the corresponding domain. However, when we want to reduce the adverse emotional reactions caused by maximizing, we can try to do so by reducing the perceived value of the domain.

5.4 Limitations and future directions

There are several limitations. First, we used a student sample and only a few domains. The homogeneity of our sample was relatively high, as the subjects were students from various universities. Future research can further verify these results in a more representative sample and a larger range of domains. Even the study domain, we captured only one or several aspects (academic performance) of students’ academic life, and other aspects (such as course selection option) should be considered in future research. Second, it is not persuasive enough to draw causal conclusions from cross-sectional data. The inference of causality came mainly from the theoretical derivation of our study. Future studies could use multiple methods to further verify the causality of subjective task values and maximizing tendencies, such as manipulating subjective values. Third, our domain-specific maximization scales were preliminary; more validity/reliability tests and revisions are needed before they can be used by others. The high correlation with perfectionism suggests that future research should include more choice-related items in the academic domain maximization scale.

There are still several concerns regarding the domain specificity of maximization that need to be further explored. For example, in addition to subjective task value, other cognitive appraisal variables, such as self-efficacy, may also have an impact on maximizing tendency. Additionally, previous studies have shown that the domain specificity of risk-taking styles is affected by personality traits. Individuals with higher neuroticism scores are more likely to have completely different risk preferences in different domains (Weller & Tikir, 2011). It would be interesting to explore whether the domain specificity of maximizing decision-making styles is affected by similar traits.