Trading style rather than return: the role of pre-trading emotional predispositions in a stock market simulation

1 Introduction

Financial decision-making has long been framed as a process in which investors evaluate risk and return in a rational manner. A large body of work in behavioral finance, psychology, and neuroeconomics has shown that affective-emotional states are involved in judgment under uncertainty (Lerner and Keltner, 2001; Loewenstein et al., 2001). These states shape attention, confidence, risk perception, and action tendencies, thereby influencing how individuals approach risky choices. Prospect theory, for example, demonstrated that decision-making under risk systematically departs from the assumptions of standard rational choice (Kahneman and Tversky, 1979). Later research further showed that affective states alter the perceived attractiveness and riskiness of alternatives, making emotions central to economic behavior rather than peripheral to it (Kuhnen and Knutson, 2011; Lerner and Keltner, 2001; Lucey and Dowling, 2005). In the same perspective, the affect heuristic literature suggests that people often rely on immediate positive or negative feelings when evaluating risky options which implies that emotional tone may shape investment behavior even before any trade is made (Duxbury et al., 2020; Finucane et al., 2000).

In financial markets, affective and emotional predispositions may not necessarily be associated with performance in a direct manner. Prior research has shown that active trading is often associated with poorer outcomes among individual investors, which points to the importance of behavioral mechanisms in explaining financial performance (Barber and Odean, 2000). At the same time, studies on traders indicate that emotional reactivity, anticipatory feelings, and context-sensitive emotional signals may matter for profitability, although the direction and magnitude of these effects remain uncertain (Bossaerts et al., 2024; Hamelin and Bonelli, 2022; Hinvest et al., 2021; Lo et al., 2005). This ambiguity suggests that affective and emotional predispositions may be associated with how market participants expose themselves to the market without necessarily being directly associated with performance.

This distinction is central to our study. The analysis examines whether positive and negative pre-trading affective and emotional predispositions are more closely related to trading style than to performance. To address this issue, we rely on data collected during a four-hour stock market simulation involving 133 s-year undergraduate students in Management. Participants traded only CAC 40 stocks, faced transaction costs of 0.2%, received no instruction regarding either the number or the size of trades, and operated under a rank-based academic incentive scheme. This design allows trading behavior to emerge relatively freely while keeping the investment environment controlled and comparable across participants.

The article addresses the following general research question: Are pre-trading emotional predispositions more closely associated with portfolio return or with trading style in a short-horizon stock market simulation? More specifically, the study investigates three related questions. First, are positive and negative pre-trading emotional predispositions associated with portfolio return? Second, are they associated with trading-style indicators such as number of transactions, average transaction volume, portfolio standard deviation, or average cash holdings relative to initial capital? Third, if no mediation emerges, do these pre-trading emotional predispositions nevertheless differentiate distinct styles of market exposure among novices? These questions are theoretically important because the literature suggests that emotions matter in finance but remains less precise about whether they influence performance or the behavioral mechanisms through which performance may be produced (Duxbury et al., 2020; Kuhnen and Knutson, 2011; Lerner and Keltner, 2001). In the article, we distinguish between general pre-trading emotional tone, measured through PANAS items, and simulation-specific emotional expectations related to the trading task. This distinction is used to organize the empirical measures, not to claim that affects and emotions represent fully separate psychological categories.

Methodologically, the study distinguishes between general pre-trading emotional tone and simulation-specific emotional expectations by constructing six antecedent indices: positive pre-trading emotional tone, negative pre-trading emotional tone, positive simulation-specific emotional expectations, negative simulation-specific emotional expectations, and broader positive and negative emotional predisposition composites. The empirical strategy examines direct associations with return and exploratory mediation-style models in which return is the dependent variable and the behavioral indicators are trading intensity, average transaction volume, portfolio standard deviation, and average cash holdings. The analyses show no significant direct association between the six composite indices and return and no significant estimated indirect path through the tested behavioral indicators. At the same time, the number of transactions is strongly and negatively associated with return, whereas the other behavioral indicators do not significantly predict return but do reveal differences in trading style. Positive indices are associated with larger positions, greater portfolio variability, and lower cash holdings, while negative indices display the opposite pattern.

This article makes three contributions. First, it contributes to the behavioral finance literature by showing that pre-trading emotional predispositions are more closely associated with the way novice investors engage with the market than with portfolio return. Second, it contributes methodologically by combining measures of general pre-trading emotional tone, simulation-specific emotional expectations, and broader emotional predisposition composites with several trading-style indicators. Third, it shows that positive and negative pre-trading emotional predispositions are associated with distinct forms of market exposure, even though they are not robustly associated with portfolio return.

2 Literature review and hypothesis development

Within this behavioral perspective, an important issue is not simply whether emotional states matter in finance, but how they can be measured before trading and related to subsequent trading behavior. In our study, this question is addressed by distinguishing general pre-trading emotional tone from simulation-specific emotional expectations. General emotional tone and task-specific emotional expectations may be associated with financial behavior through different empirical patterns. General pre-trading emotional tone refers to diffuse positive or negative feeling states, often measured through the PANAS (Watson et al., 1988), whereas simulation-specific emotional expectations are more directly tied to the upcoming trading situation and may orient judgment and action in a more context-sensitive manner. Emotions may shape decision-making before any action is taken by influencing perceived risk and readiness to engage with the market (Kuhnen and Knutson, 2011; Loewenstein et al., 2001).

Literature suggests that positive and negative pre-trading emotional predispositions may be associated with different forms of financial behavior. Positive states tend to be associated with greater willingness to engage with risk, whereas negative states are often linked to caution and more pronounced sensitivity to potential losses (Kuhnen and Knutson, 2011; Lucey and Dowling, 2005). Appraisal theories show that different negative emotions may produce distinct behavioral consequences. Fear, for example, is generally associated with risk-averse choices, whereas anger is more often related to greater willingness to accept risk (Lerner and Keltner, 2001). Gambetti and Giusberti (2012) similarly found that anxiety and anger were associated with different investment tendencies, confirming that negative emotional states should not be treated as behaviorally uniform.

In this perspective, pre-trading emotional predispositions may be associated with portfolio return and trading behavior through two channels. On the one hand, they may be directly associated with performance, as confidence, optimism, fear, or aversion may shape judgment, perceptions of risk, and the quality of trading decisions. On the other hand, they may be indirectly associated with performance through trading behavior, by shaping the behavioral configuration through which participants engage with the market. Financial performance may depend on the quality of judgment, and how investors allocate capital, tolerate variability and manage cash in the portfolio.

The pre-trading dimension is relevant in the present context. Emotional tone and simulation-specific emotional expectations are measured before market entry and may be associated with how individuals approach the trading situation. In a financial setting, participants may anticipate gains, losses, disappointment, or excitement, and these expectations may be related to subsequent trading behavior. Prior evidence suggests that anticipatory emotional states matter in investment settings, although their relationship with performance remains difficult to interpret. These findings justify examining whether pre-trading emotional predispositions are associated with portfolio return and trading-style indicators. Hinvest et al. (2021), for example, show that anticipatory emotion can influence investment choices among non-professional investors, while Hamelin and Bonelli (2022) suggest that anticipatory feelings may be linked to performance. In our study, these pre-market psychological orientations are operationalized more cautiously as pre-trading emotional predispositions. Because pre-trading emotional predispositions may shape confidence, perceived risk, and readiness to engage with the market before the first trading decision, it remains theoretically relevant to test whether they are associated with short-horizon portfolio return, even though the direction and strength of this association remain uncertain.

At the same time, the literature on individual investors indicates that behavioral pathways are likely to play an important role. Barber and Odean (2000) show that greater trading activity is not associated with better outcomes, as more active investors tend to underperform. Related research indicates that sensation seeking and overconfidence help explain more intense forms of market participation (Dorn and Sengmueller, 2009; Grinblatt and Keloharju, 2009). These findings justify examining whether pre-trading emotional predispositions are associated with portfolio return and trading-style indicators, and whether these variables are statistically connected in exploratory mediation-style analyses.

In this perspective, our study examines whether pre-trading emotional predispositions are more closely associated with portfolio return or with the trading style through which novice participants engage with the market. The theoretical expectation is that positive and negative pre-trading emotional predispositions may be associated with distinct forms of market engagement. At the same time, the literature suggests that not all behavioral dimensions play the same role in short-term financial performance. It is therefore necessary to distinguish between associations with portfolio return, associations with trading-style indicators, and exploratory mediation-style paths linking pre-trading emotional predispositions, trading behavior, and return. On this basis, the following hypotheses are proposed.

2.1 Hypothesis 1: pre-trading emotional predispositions and exposed market engagement

H1a. Positive pre-trading emotional tone is associated with a more exposed trading style, reflected in larger average transaction volume, greater portfolio variability, and lower average cash holdings.

H1b. Positive simulation-specific emotional expectations are associated with a more exposed trading style, reflected in larger average transaction volume, greater portfolio variability, and lower average cash holdings.

2.2 Hypothesis 2: negative pre-trading emotional predispositions and defensive market engagement

H2a. Negative pre-trading emotional tone is associated with a more defensive trading style, reflected in smaller average transaction volume, lower portfolio variability, and higher average cash holdings.

H2b. Negative simulation-specific emotional expectations are associated with a more defensive trading style, reflected in smaller average transaction volume, lower portfolio variability, and higher average cash holdings.

2.3 Hypothesis 3: pre-trading emotional predispositions and portfolio return

H3. Pre-trading emotional predispositions are associated with portfolio return.

2.4 Hypothesis 4: trading intensity and performance

H4. The number of transactions is negatively associated with portfolio return. Among the four behavioral indicators, only trading intensity gives rise to a directional expectation regarding portfolio return. Prior research suggests that excessive trading activity may be detrimental to performance, whereas the effects of average transaction volume, portfolio standard deviation and cash holdings are less straightforward.

2.5 Hypothesis 5: exploratory mediation-style analysis

H5. Exploratory mediation-style analyses examine whether pre-trading emotional predispositions are statistically connected with portfolio return through trading behavior, namely number of transactions, average transaction volume, portfolio standard deviation, and average cash holdings relative to initial capital.

3 Method and research design

3.1 Research design and methodological positioning

This study relies on an experimental stock market simulation. This methodological choice is based on the logic of experimental finance, a research tradition that has become increasingly important for studying financial decision-making under controlled conditions (Huber and Kirchler, 2023; Kirchler and Weitzel, 2023). Laboratory and simulation-based asset markets allow researchers to observe decisions under uncertainty while controlling key features of the environment, such as the set of available assets, transaction rules, and information conditions. A long tradition in experimental asset market research has shown that such designs can generate theoretically meaningful behaviors, including overtrading, bubbles, mispricing, and strategic responses to incentives (Smith et al., 1988), and are especially valuable when the objective is to observe behavioral regularities that are difficult to study in naturally occurring financial settings (Merl et al., 2022).

This methodological framework is relevant for our study because the objective is to examine whether pre-trading emotional predispositions are associated with portfolio return and trading-style indicators under controlled conditions. Experimental asset market designs are especially useful when the aim is to relate psychological antecedents to performance and behavioral mediators under controlled conditions (Huber and Kirchler, 2023; Kirchler and Weitzel, 2023; Merl et al., 2022). In our design, this control is achieved by placing all participants in the same trading environment, with the same investment universe, transaction rules and incentive structure.

3.2 Use of student participants

The experiment was conducted with 133 s-year undergraduate students in Management at the University of Mons. The sample consisted of 77 male participants (57.9%) and 56 female participants (42.1%). Participants’ age ranged from 18 to 42 years, with a mean age of 20.90 years (SD = 3.48). Additional descriptive information on gender, age, and prior stock-market experience is reported in Supplementary Appendix B4. The use of student participants is common practice in experimental economics and experimental finance. Laboratory economics has historically relied on students because they provide relatively homogeneous participants, facilitate control over the decision environment, and make it easier to identify behavioral mechanisms. Kessler and Vesterlund (2015) argue that the relevant issue for laboratory external validity is often not whether exact quantitative magnitudes can be generalized, but whether the direction of behavioral relationships is significant across contexts.

However, the use of students should be interpreted with appropriate caution. Belot et al. (2015) show that students and non-students may differ in some respects, particularly in social preferences and strategic behavior. This does not invalidate student-based experiments, but their main strength lies in identifying behavioral regularities and mechanisms under controlled conditions rather than providing immediate population-level estimates. For the purposes of our study, the student sample is methodologically justified because the aim is to examine associations between pre-trading emotional predispositions, trading style, and performance. The results are more generalizable to early-stage market participants than to professional investors, whose behavior is more likely to be shaped by experience and trading routines. Among the 133 participants, 32 students reported prior stock-market experience, corresponding to 24.1% of the sample (Supplementary Appendix B4). This measure was self-reported and does not allow us to distinguish between occasional interest in financial markets, previous simulated trading, or actual investment experience. The sample should therefore be interpreted as mainly composed of novice participants. The sample size was determined by the number of students enrolled in the course and participating in the trading simulation. No a priori power analysis was conducted before data collection. With 133 participants, the study has approximately 80% power to detect bivariate correlations of about r = .24 at the 5% significance level. The design is therefore reasonably powered to detect moderate associations, but it may be underpowered to detect small associations, especially in exploratory mediation-style analyses where estimated indirect paths are typically smaller and require larger samples.

3.3 Experimental design

The simulation took place on December 11, 2025, from 1:30 p.m. to 5:30 p.m. in a lecture hall under instructor supervision. Participants were physically present in the same room, communication was allowed, and they used the ABC Bourse platform to manage an individual virtual portfolio. Before the simulation began, written instructions were distributed and summarized orally. A prior presentation of the CAC 40 index was also provided. Participants completed pre-simulation questionnaires measuring affective and emotional variables on five-point Likert scales before making any trading decision.

Each participant received virtual capital of EUR 100,000 and could trade only CAC 40 equities. Orders were placed individually through the platform and executed at prevailing market prices. Short selling was not allowed. Participants could buy eligible stocks, sell previously purchased holdings, or keep cash. The platform provided real-time market prices, real-time market information, portfolio composition, available cash, and portfolio value. Internet access and mobile phones were allowed, so participants could consult financial information on other stock-market websites during the session.

No instruction was given regarding the number of transactions or the size of positions. Participants were free to trade actively, trade occasionally, hold positions, or remain partly in cash. Transaction costs of 0.2% were applied. The platform recorded all transaction histories and portfolio states, which were used to compute the number of transactions, average transaction size, average cash holdings, final portfolio value, and portfolio variability.

The simulation included an academic incentive scheme based on final portfolio value: the best-performing participant received 2 bonus points, the second 1.5 points, the third one point, and all other participants 0.5 points. Rankings and performance feedback were available during the session. During the four-hour session, the CAC 40 increased by approximately 0.30%, from 8,061.40 to 8,085.76 points. The price movements and financial information were real and only the investment capital and portfolios were virtual.

The ranking and academic bonus system may have introduced a tournament-like component into the simulation. This feature was intended to increase engagement, but it may also have amplified competitive behavior and social comparison. At the same time, the use of academic credit does not invalidate the experimental setting. More generally, the use and design of incentives are central issues in experimental economics, including when non-cash incentives are used to motivate participants (Voslinsky and Azar, 2021). Ding et al. (2018) show that extra-credit incentives can be an effective reward medium in experimental asset markets, with bubble measures in extra-credit sessions not significantly different from cash sessions. Moreover, relative performance evaluation and ranking concerns are not entirely artificial in financial contexts, as tournament-like incentives and relative performance comparisons have been documented in the mutual fund industry (Brown et al., 1996; Chevalier and Ellison, 1997).

To improve transparency and replicability, the full description of the participant instructions, ABC Bourse platform, trading rules, information access, and recorded data is provided in Supplementary Appendix A.

3.4 Ethical approval and study registration

The study was not pre-registered. The hypotheses and analytical strategy were developed before the final statistical analyses but were not formally deposited in a public registry.

Data were collected in the context of a pedagogical stock market simulation conducted at the University of Mons. Under the institutional procedures applicable to this teaching activity, no formal external ethics approval was required. All students were informed of the objectives of the simulation before it began, as well as of the rules of the trading session and the academic incentive scheme. The analysis used data generated during the simulation and pre-simulation questionnaires. Data were treated confidentially and analyzed at the aggregate level. All trading decisions were made using virtual capital.

Additional methodological details are available from the Corresponding Author upon reasonable request.

4 Measures and variables

4.1 Independent variables: pre-trading emotional tone and emotional expectations

The empirical strategy relies on six pre-trading composite indices constructed from a pre-trading questionnaire. These indices capture two operational dimensions of participants’ emotional predispositions before the simulation: general pre-trading emotional tone, measured through PANAS items, and simulation-specific emotional expectations, measured through items related to the upcoming trading task. Four first-level indices were computed as the mean of their constituent items, all measured using a five-point Likert scale: positive pre-trading emotional tone (PA), negative pre-trading emotional tone (NA), positive simulation-specific emotional expectations (PE), and negative simulation-specific emotional expectations (NE). Two broader indices were also created by aggregating the corresponding positive and negative dimensions: global positive emotional predisposition (GPC) and global negative emotional predisposition (GNC). Positive and negative indices were retained separately because they are not statistical mirror images in the sample. The correlation matrix shows weak associations between the positive and negative poles: PA and NA are only weakly correlated (r = .077), while PE and NE are virtually uncorrelated (r = .007). This suggests that positive and negative pre-trading emotional orientations capture distinct dimensions. Retaining both poles is theoretically and empirically justified, as participants may simultaneously report low or high levels of positive and negative emotional predispositions.

The positive pre-trading emotional tone index is based on the ten positive PANAS items administered before the simulation: interested, excited, strong/energetic, enthusiastic, proud, alert, inspired, determined, attentive, and active. The negative pre-trading emotional tone index is based on the ten negative PANAS items: distressed, upset, guilty, scared, hostile, irritable, ashamed, nervous, jittery, and afraid. These two indices measure general emotional tone prior to the simulation.

The positive simulation-specific emotional expectations index includes six items referring to optimism about what will happen, feeling happy at the idea of participating, looking forward to the simulation, expecting good surprises, feeling confident, and being motivated by comparison with other students. The negative simulation-specific emotional expectations index contains six simulation-specific items referring to fearing that things may go badly, feeling fear, feeling already somewhat irritated, feeling sad, experiencing disgust toward this kind of exercise, and fearing bad surprises during the simulation. The six positive and six negative items were selected to represent two symmetrical simulation-specific emotional expectation poles. The positive pole reflects anticipation of opportunity and engagement, whereas the negative pole measures anticipation of threat and aversion. This choice is consistent with emotion theory in two respects. First, positive emotions are known to support exploratory and approach-oriented engagement (Fredrickson, 2001). Second, emotions can be differentiated by appraisal patterns and action readiness tendencies, with negative emotions being tied to defensive orientations (Frijda et al., 1989). The indices were constructed as two opposed task-specific emotional expectation configurations linked to different modes of behavioral engagement. In addition, the two global composites were designed to provide a broader representation of participants’ positive and negative pre-trading emotional predispositions. The global positive emotional predisposition composite combines positive pre-trading emotional tone and positive simulation-specific emotional expectations, whereas the global negative emotional predisposition composite combines negative pre-trading emotional tone and negative simulation-specific emotional expectations.

All six indices display strong internal consistency. Cronbach's alpha equals.899 for positive pre-trading emotional tone,.931 for negative pre-trading emotional tone,.861 for positive simulation-specific emotional expectations, .851 for negative simulation-specific emotional expectations, .926 for the global positive emotional predisposition composite, and .933 for the global negative emotional predisposition composite. These values indicate that the composite measures are coherent to be treated as reliable variables.

4.2 Dependent variable and behavioral variables

The dependent variable is portfolio return. This variable is calculated as the percentage change between the final value of the portfolio and its initial value. It serves as the performance indicator in direct association models and mediation-style analyses. In addition, the analysis considers four behavioral variables reflecting distinct dimensions of trading style. These variables are examined as indicators of market engagement and as behavioral indicators in mediation-style analyses linking pre-trading emotional predispositions and return. The first behavioral variable is the number of transactions, which reflects how actively participants engage with the market. The second is average transaction volume, which reflects the size of positions taken. The third is portfolio standard deviation. It is computed at the end of the simulation from the sequence of portfolio values recorded during the trading session. After each order, the platform recorded the updated portfolio allocation, cash position, and total portfolio value. Portfolio standard deviation is calculated from the evolution of the participant's total portfolio value across these portfolio-update points. The variable is expressed in euros. The fourth variable is average cash holdings relative to initial capital. After each order, the platform recorded the amount of cash remaining in the participant's portfolio. The variable is computed from the sequence of cash positions observed across these successive portfolio-update points and divided by the initial capital of EUR 100,000. It therefore reflects the average proportion of capital kept in cash over the course of the simulation, based on event-based portfolio records. This variable captures the extent to which participants kept resources uninvested during the trading session and indicates a preference for remaining in cash. This selection is consistent with experimental finance literature (Merl et al., 2022; Smith et al., 1988). Evidence suggests that activity may be detrimental to performance, whereas position size, portfolio variability, and cash retention may better reflect market exposure than determinants of return.

4.3 Empirical strategy

The empirical strategy proceeds in two stages. First, direct associations between the six pre-trading emotional predisposition indices and portfolio return are examined using the following model:Where Y denotes portfolio return, X denotes one pre-trading emotional predisposition index, and c measures the direct association between the index and the return.

Y=β0+cX+ε

Second, exploratory mediation-style analyses are conducted in which each index is tested separately through each of the four behavioral indicators. These models retain the formal structure of mediation analysis, with an antecedent variable, a behavioral indicator, and portfolio return, but they are used only as exploratory mediation-style models and are interpreted descriptively rather than causally (see below). The behavioral-indicator model is specified as:where M denotes the behavioral indicator, X denotes the pre-trading emotional predisposition index, and a captures the association between the emotional predisposition index and the behavioral indicator.

M=α0+aX+ε

The final mediation model is estimated as:where c′ represents the direct effect of the antecedent variable on return after inclusion of the behavioral indicator, and b captures the association of the behavioral indicator with return. The estimated indirect path is computed as a×b and assessed using bootstrap confidence intervals based on 5,000 resamples.

Y=β0+c′X+bM+ε

The models retain the structure of mediation analysis but are interpreted as exploratory mediation-style analyses rather than causal mediation tests. The sequential ignorability assumption required for causal mediation cannot be guaranteed because the behavioral indicators were not randomly assigned but resulted from participants’ own trading decisions. The estimated indirect paths should be read as exploratory associations rather than causal mediation effects.

The main models are estimated without controls to preserve comparability across specifications, while robustness checks including participant-level controls are reported in Section 5.8 and Supplementary Appendix B1. This approach follows the logic of experimental finance by observing financial decisions in a controlled trading environment while holding key features of the market setting constant across participants (Huber and Kirchler, 2023; Kirchler and Weitzel, 2023). All statistical analyses were conducted using Jamovi.

5 Results

This section presents descriptive statistics and bivariate correlations for the variables, providing an overview of the data structure and the relationships among pre-trading emotional predispositions, behavioral indicators and return. It then reports the direct associations between the six pre-trading emotional predisposition indices and return and the exploratory mediation-style analyses conducted through the four behavioral indicators of trading style. The results are organized to distinguish associations with performance from associations with the behavioral configuration of market participation. The section concludes with a summary table of hypothesis testing.

5.1 Descriptive statistics

Table 1 reports descriptive statistics for the variables used in the analyses. The results show important variation across the behavioral indicators and a broader dispersion for the trading variables than for the anticipatory composite indices. They also suggest that, before the simulation, participants were characterized by a more positive than negative affective-emotional climate. Mean levels are higher for positive than for negative indices, both for general pre-trading emotional tone (PA = 3.220 vs. NA = 1.620) and for simulation-specific emotional expectations (PE = 2.934 vs. NE = 1.692), with the same pattern appearing in the global indices (GPC = 3.113 vs. GNC = 1.647). This configuration may indicate that participants approached the trading session with a relatively favorable anticipatory orientation, possibly because participation was associated with academic rewards and participants had nothing to lose.

Table 1. Descriptive statistics.

5.2 Correlation matrix

Table 2 reports main bivariate correlations between the six emotional predisposition indices and the main trading outcomes. The table shows that portfolio return is only weakly related to the emotional indices. By contrast, several emotional indices are significantly associated with trading-style indicators. Positive emotional predisposition indices are generally associated with higher portfolio standard deviation and, in some cases, larger average transaction volume and lower cash holdings. Negative emotional predisposition indices are generally associated with lower portfolio standard deviation, smaller average transaction volume, and higher cash holdings.

Table 2. Main correlations between emotional predisposition indices and main trading outcomes.

Pearson correlations are reported. * p < .05, ** p < .01, *** p < .001.

5.3 Direct associations between pre-trading emotional predispositions and return

The direct association models show that none of the six pre-trading emotional predisposition indices is significantly associated with portfolio return. This result is consistent with the correlation matrix, in which portfolio return is only weakly correlated with the emotional indices. The strongest association concerns negative simulation-specific emotional expectations, but it remains non-significant. These findings suggest that pre-trading emotional predispositions are not significantly associated with final trading performance in the direct association models. Table 3 summarizes the empirical support for each hypothesis tested in the study.

Table 3. Summary of hypotheses and empirical support.

“Partially supported” indicates that the empirical evidence is consistent with the expected theoretical direction, but only some of the predicted relationships are statistically significant. In other words, the hypothesis is not fully validated across all its behavioral dimensions, yet the observed pattern remains coherent to provide partial empirical support.

5.4 Exploratory mediation-style analyses

The detailed mediation-style results are reported in Supplementary Appendix C. In the main text, the results are summarized to emphasize the distinction between associations with trading style and associations with portfolio return. When the number of transactions is used as the behavioral indicator, no significant estimated indirect path is observed between any emotional predisposition index and portfolio return. Trading intensity is negatively associated with return, but the emotional indices are not significantly associated with the number of transactions.

For average transaction volume, several emotional indices are associated with position size. Negative pre-trading emotional tone and the global negative emotional predisposition composite are associated with smaller average trades, whereas positive simulation-specific emotional expectations and the global positive emotional predisposition composite are associated with larger average trades. However, no significant estimated indirect path to portfolio return is observed.

Portfolio standard deviation displays a similar pattern. Positive pre-trading emotional tone, positive simulation-specific emotional expectations, and the global positive emotional predisposition composite are associated with higher portfolio variability, whereas the corresponding negative indices are associated with lower portfolio variability. Again, no significant estimated indirect path to portfolio return is observed.

Finally, average cash holdings show that negative pre-trading emotional tone, negative simulation-specific emotional expectations, and the global negative emotional predisposition composite are associated with higher cash retention, whereas positive simulation-specific emotional expectations are associated with lower cash retention. No significant estimated indirect path is observed.

5.5 Robustness checks

As a robustness check, we re-estimated the association models by including participant-level controls. The first specification included gender, age, and prior stock-market experience. A second specification additionally included the five OCEAN personality traits. The detailed results are reported in Supplementary Appendix B1. Since control variables are used only in robustness checks, Table 1 reports descriptive statistics for the main analytical variables, while the control-based specifications are reported in Supplementary Appendix B3. Supplementary Appendix B3 reports exploratory associations between participant-level characteristics and the main trading-style indicators. These analyses indicate that some participant characteristics are associated with trading style. Male participants are associated with larger average transaction volume, higher portfolio standard deviation, and lower average cash holdings. Prior stock-market experience is associated with larger average transaction volume, while age and openness are associated with higher portfolio standard deviation.

The robustness checks support the main interpretation of the article. First, pre-trading affective and emotional predispositions do not significantly predict portfolio return. This finding remains unchanged after the inclusion of controls and after the addition of OCEAN personality traits. Second, emotional predispositions do not explain trading frequency: none of the six affective or emotional indices significantly predicts the number of transactions. Third, negative pre-trading emotional predispositions remain consistently associated with a more defensive trading style. Fourth, trading intensity remains significant and negatively associated with portfolio return across specifications, making it the most stable performance-related result.

These robustness checks reinforce the interpretation of the findings. Pre-trading emotional predispositions appear to be more closely related to trading style than to final performance. Negative pre-trading emotional predispositions are associated with lower market exposure, whereas the emotional predisposition indices show no statistically significant association with return over the short simulation horizon.

We also applied a Benjamini–Hochberg false discovery rate correction to the 30 primary associations between the six indices and the five outcomes. The detailed results are reported in Supplementary Appendix B2. The correction does not change the main interpretation of the article. No pre-trading emotional predisposition index is significantly associated with return or trading frequency after adjustment. By contrast, several associations with trading-style indicators remain significant, particularly for negative pre-trading emotional predispositions, which are associated with smaller transaction volume, lower portfolio variability, and higher cash holdings.

5.6 Overall interpretation

The results indicate a stable pattern. None of the six pre-trading emotional predisposition indices is significantly associated with return, and none generates a significant estimated indirect path in the mediation-style analyses. The findings suggest that pre-trading emotional predispositions are not significantly associated with financial performance but are associated with the behavioral form of market engagement.

Indeed, the indices are consistently related to several dimensions of trading style. Positive pre-trading emotional predispositions are associated with larger average transaction volumes, greater portfolio standard deviation, and lower average cash holdings, which point to a more exposure-seeking position. Negative pre-trading emotional predispositions exhibit the opposite profile, being associated with smaller transaction volumes, lower portfolio variability, and higher cash retention, thereby reflecting a more defensive orientation. These results suggest that pre-trading emotional predispositions help differentiate levels of market exposure and capital commitment.

Accordingly, from our results, pre-trading emotional predispositions are not significantly associated with return, nor do the mediation-style analyses indicate a robust indirect association through the behavioral indicators tested. However, they contribute to distinguishing offensive from defensive trading configurations. In this perspective, our findings suggest that greater attention should be paid to how pre-trading emotional predispositions are associated with the style, intensity, and risk profile of trading behavior, rather than focusing only on whether they improve or reduce performance.

6 Discussion

The results provide evidence that pre-trading emotional predispositions are more closely associated with trading style than with portfolio return. Positive and negative predispositions do not translate into robust differences in performance, but they are associated with distinct ways of engaging with the market. The analyses suggest that positive predispositions are linked to greater market exposure, whereas negative predispositions are associated with a more defensive trading style, reflected in smaller average positions, lower portfolio variability, and higher cash retention. The absence of significant associations with portfolio return should nevertheless be interpreted cautiously. The simulation lasted only four hours, during which the CAC 40 increased by approximately 0.30%, and portfolio returns were highly compressed around zero. In such a setting, final return may reflect short-term price noise, timing effects, and limited market movement. The absence of significant return associations should not be read as definitive evidence of no relationship between pre-trading affective-emotional predispositions and performance. It indicates that no statistically significant return association can be detected in this short-horizon design.

This interpretation reinforces the central contribution of the article: the findings are more informative about market exposure and trading style than about short-term portfolio performance. This conclusion is consistent with behavioral finance showing that investor behavior often reflects psychological dispositions more clearly than financial efficiency. Shefrin and Statman (1985) showed that investor decisions are often shaped by behavioral patterns that deviate from rational choice, while Odean (1998) found that investors may overtrade without achieving superior performance. Statman et al. (2006) linked investor overconfidence to higher trading volume, suggesting that psychological variables may influence market engagement through which investors act. In our study, pre-trading emotional predispositions are reflected less in final return than in the way participants commit capital, tolerate portfolio variability, and manage cash.

Positive pre-trading emotional predisposition indices are associated with larger average transaction volumes, greater portfolio standard deviation, and lower average cash holdings, whereas negative pre-trading emotional predisposition indices show the opposite pattern. These associations point toward two contrasting styles of market exposure. On the one hand, positive pre-trading emotional predispositions seem to be associated with a more engaged posture, characterized by stronger capital commitment and a greater willingness to accept variability. On the other hand, negative pre-trading emotional predispositions are associated with a more defensive posture, marked by smaller average positions, lower exposure, and a stronger tendency to preserve cash. Such findings are in line with research showing that investor confidence, optimism, and self-perceived competence often translate into more assertive forms of market action, while doubt, anxiety and defensive expectations promote a more cautious form of participation (Bouteska et al., 2023; Hoffmann and Post, 2016; Nofsinger, 2005).

From a psychological perspective, this distinction can be understood in terms of action readiness. Positive pre-trading emotional climates appear to be associated with engagement, whereas negative pre-trading emotional climates appear to be associated with caution. This does not imply that positive emotions necessarily improve decision quality, nor that negative emotions necessarily reduce it. It only suggests that these states orient investors toward different ways of confronting uncertainty. In this regard, the findings resonate with work showing that mood and emotion affect financial decisions through changes in beliefs and readiness to act. Investor mood, for example, has been shown to influence market behavior and judgment even when its effect on returns is neither stable nor linear (Cohen-Charash et al., 2013; Shu, 2010). Our results extend this logic by showing that initial emotional orientation is associated with the configuration of novice investors’ trading behavior, even when they do not produce a significant channel to performance.

Because the participants are novice investors, they are less likely to rely on established routines and are likely to express their psychological orientation through their behavior. Da Costa et al. (2013) show that novice participants exhibit biases such as the disposition effect, while Li et al. (2023) indicate that early investment experience affects stock investment behavior through changes in risk perception. In our study, the participants’ inexperience may help explain why pre-trading emotional predispositions are reflected in behavioral style. Instead of being neutralized by prior trading habits, these orientations appear in more aggressive or more defensive market exposure. This makes our findings informative for the psychology of novice investors, because they suggest that, at an early stage of market engagement, pre-trading emotional predispositions are expressed more in trading style than in performance.

This point also helps explain why the exploratory mediation-style analyses did not reveal significant indirect paths. The absence of significant estimated indirect paths does not result from a lack of associations between pre-trading emotional predispositions and behavioral indicators. Several of those relationships are statistically significant and theoretically coherent but the behavioral indicators most closely associated with pre-trading emotional predispositions do not significantly predict return. This result is in line with syntheses emphasizing that emotions in finance often operate through behavioral processes rather than through effects on returns (Goodell et al., 2023). It also complements work on inexperienced investors showing that psychological traits influence trading behavior without necessarily improving financial returns (Bouteska et al., 2023; Brooks and Williams, 2022).

7 Conclusion

The study investigated whether positive and negative pre-trading emotional predispositions measured before a stock market simulation were associated with portfolio return, either through direct effects or through estimated indirect paths in exploratory mediation-style analyses. In doing so, the study followed the logic of behavioral finance, which has argued that financial decision-making must be understood considering psychological processes, judgment under uncertainty and investor heterogeneity (De Bondt and Thaler, 1985; Thaler, 1980).

Our results show that broad pre-trading emotional predispositions are not significantly associated with portfolio return, nor do the exploratory mediation-style analyses reveal significant estimated indirect paths through the behavioral indicators tested. In that sense, the answer to the first two research questions is negative: neither the six composite nor the four mediation pathways provide evidence of a return-transmission mechanism. However, this does not imply that pre-trading emotional predispositions are irrelevant. The analyses show that they are associated with several dimensions of trading behavior. Positive pre-trading emotional predispositions are linked to larger transaction volumes, greater portfolio variability, and lower cash retention, whereas negative pre-trading emotional predispositions are associated with smaller positions, lower variability, and higher cash holdings. The answer to the third research question is therefore: pre-trading emotional predispositions are more closely associated with market engagement than with the performance participants eventually achieved.

This conclusion suggests that the contribution of pre-trading emotional predispositions in financial contexts may lie less in determining success or failure than in organizing market posture. From this perspective, the study supports a view of investor psychology in which emotional predispositions are associated with how individuals confront uncertainty, allocate capital, tolerate exposure, and hold cash. Such an interpretation is consistent with behavioral finance that has shifted attention from normative models to psychological explanations of investor behavior (Campbell, 2006). It is also consistent with work showing that participation in financial markets depends on subjective readiness, competence, and psychological orientation (Van Rooij et al., 2011).

The study offers several added values. First, it contributes by distinguishing between general pre-trading emotional tone, simulation-specific emotional expectations, and broader emotional predisposition composites. This distinction helps clarify how different affective and emotional dimensions are associated with behavior. Second, it contributes by showing that affective and emotional states can be behaviorally meaningful even when they do not predict performance. Third, the study contributes by using a stock market simulation with novice participants under controlled conditions. In that respect, it provides insight into the early formation of investor behavior, before trading habits become stabilized through market experience. This is especially relevant given the growing interest in novice market participants in literature (Campbell, 2006; Slovic et al., 2004).

8 Limitations

This study should be interpreted considering several limitations. First, data were collected in a simulated stock market environment. This choice offered a high level of control over the trading setting, the investment universe, transaction costs and incentives. However, the observed behaviors are produced within a structured environment that simplifies some features of real financial markets. The results should be understood as evidence on behavioral mechanisms under controlled conditions rather than as a direct representation of all trading situations.

Second, the participants were second-year undergraduate students, which makes the study informative for the analysis of novice investors and early-stage market behavior. At the same time, this implies that the conclusions are relevant to inexperienced decision-makers rather than to professional traders.

Third, the simulation lasted four hours, which means that portfolio return was measured over a relatively short horizon. In such a setting, broad pre-trading emotional predispositions may be more easily expressed through trading style than through a stable performance effect. This may partly explain why the behavioral patterns appear more clearly than the direct or indirect effects on return. Accordingly, the results are informative with respect to market exposure, while conclusions regarding performance should be interpreted within the temporal boundaries of the simulation. In addition, the statistical power of the study should be considered when interpreting the absence of significant associations with portfolio return. With 133 participants, the study is sufficient to detect moderate bivariate associations, but it may be insufficient to detect small associations, especially in exploratory mediation-style analyses where estimated indirect paths are typically smaller and require larger samples. The absence of significant return associations should therefore not be interpreted as definitive evidence that such associations are absent.

Fourth, the study relies on composite indices of pre-trading emotional predispositions. Composite indices could attenuate some of the specificity of individual emotional states.

Fifth, academic bonus points and live rankings may have introduced a tournament-like component into the simulation, increasing engagement but also amplifying competitive behavior and social comparison. Although extra-credit incentives have been shown to be an effective reward medium in experimental asset markets, the combination of academic bonus points and live rankings limits the generalization of the findings to contexts in which ranking, comparison, and performance feedback are absent.

9 Directions for future research

The study also leaves several questions open. First, future work could examine specific emotional predispositions rather than relying only on broad composites. More specific affective and emotional states may be more informative for explaining behavioral mechanisms or links with performance.

Second, it would be interesting to compare novices and more experienced investors. A comparative design could determine whether these psychological influences remain equally visible once investors have developed more stable routines, greater discipline and stronger market experience.

Third, future research could adopt a longer trading horizon. In our four-hour simulation, broad pre-trading emotional predispositions appear mainly through differences in trading style rather than through differences in return. Over a longer horizon, these behavioral differences may accumulate and produce more significant performance effects, especially through repeated exposure and portfolio adjustments.

Fourth, future studies could introduce alternative experimental conditions. For example, varying transaction costs or incentive structures would facilitate determining whether the behavioral associations of pre-trading emotional predispositions depend on the features of the trading environment.

Fifth, an extension would be to combine pre-trading emotional predispositions with other variables, such as confidence, risk perception or ambiguity tolerance. This would facilitate assessing whether affective and emotional climates operate independently or whether they are part of a wider psychological configuration shaping trading behavior.

Finally, future work could examine the dynamic relationship between pre-trading, in-trading, and post-trading emotions. Our study focuses on pre-trading emotional predispositions measured before the simulation, but future research could investigate how these initial predispositions evolve during market participation and how they interact with emotional reactions to gains, losses, or unexpected market developments.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frbhe.2026.1836906/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: John L. Haracz, Indiana University, United States

Reviewed by: Benno Guenther, London School of Economics and Political Science, United Kingdom

Xuan Phan Ngoc Yen, University of Finance and Marketing, Vietnam