Implicit measurement of the moral self-image using the Go/No-Go Association Task (GNAT): An empirical investigation of the convergent validity between explicit and implicit measures

Bläßer, Louisa Felicitas Article Implicit measurement of the moral self-image using the Go/No-Go Association Task (GNAT): An empirical investigation of the convergent validity between explicit and implicit measures Junior Management Science (JUMS) Provided in Cooperation with: Junior Management Science e. V. Suggested Citation: Bläßer, Louisa Felicitas (2025) : Implicit measurement of the moral self-image using the Go/No-Go Association Task (GNAT): An empirical investigation of the convergent validity between explicit and implicit measures, Junior Management Science (JUMS), ISSN 2942-1861, Junior Management Science e. V., Planegg, Vol. 10, Iss. 4, pp. 966-984, https://doi.org/10.5282/jums/v10i4pp966-984 This Version is available at: https://hdl.handle.net/10419/334184 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Junior Management Science 10(4) (2025) 966-984 Junior Management Science www.jums.academy ISSN: 2942-1861 Editor: DOMINIK VAN AAKEN Advisory Editorial Board: FREDERIK AHLEMANN JAN-PHILIPP AHRENS THOMAS BAHLINGER MARKUS BECKMANN SULEIKA BORT ROLF BRÜHL KATRIN BURMEISTER-LAMP CATHERINE CLEOPHAS NILS CRASSELT BENEDIKT DOWNAR KERSTIN FEHRE MATTHIAS FINK DAVID FLORYSIAK GUNTHER FRIEDL MARTIN FRIESL FRANZ FUERST WOLFGANG GÜTTEL NINA KATRIN HANSEN ANNE KATARINA HEIDER CHRISTIAN HOFMANN SVEN HÖRNER STEPHAN KAISER NADINE KAMMERLANDER ALFRED KIESER ALEKSANDRA KLEIN NATALIA KLIEWER DODO ZU KNYPHAUSEN-AUFSESS SABINE T. KÖSZEGI ARJAN KOZICA CHRISTIAN KOZIOL MARTIN KREEB WERNER KUNZ HANS-ULRICH KÜPPER MICHAEL MEYER JÜRGEN MÜHLBACHER GORDON MÜLLER-SEITZ J. PETER MURMANN ANDREAS OSTERMAIER BURKHARD PEDELL ARTHUR POSCH MARCEL PROKOPCZUK TANJA RABL SASCHA RAITHEL NICOLE RATZINGER-SAKEL ASTRID REICHEL KATJA ROST THOMAS RUSSACK FLORIAN SAHLING MARKO SARSTEDT ANDREAS G. SCHERER STEFAN SCHMID UTE SCHMIEL CHRISTIAN SCHMITZ MARTIN SCHNEIDER MARKUS SCHOLZ LARS SCHWEIZER DAVID SEIDL THORSTEN SELLHORN STEFAN SEURING VIOLETTA SPLITTER ANDREAS SUCHANEK TILL TALAULICAR ANN TANK ANJA TUSCHKE MATTHIAS UHL CHRISTINE VALLASTER PATRICK VELTE CHRISTIAN VÖGTLIN BARBARA E. WEISSENBERGER ISABELL M. WELPE HANNES WINNER THOMAS WRONA THOMAS ZWICK Volume 10, Issue 4, December 2025 JUNIOR MANAGEMENT SCIENCE Marie-Claire Joyeaux,Work Less, Live More? The Impact of an Introduction of the Four-Day Working Week on Happiness in the Context of the Icelandic Four-Day Working Week Experiment Niklas Janßen, Integrating Sustainability in Risk Management and Internal Control Systems: An Empirical Assessment of ESG Reporting of German DAX40 Firms Tobias Keserü,A New Dimension of Transparency: ESG Disclosure and Its Effect on Shareholder Behavior Antonia Engel, ESG Regulation Across the Globe: Does ESG Regulation Pay Off? Finn Matthes Gooßen, The Impact of Female Board Members on ESG Performance: An Empirical Analysis Louisa Felicitas Bläßer, Implicit Measurement of the Moral Self-Image Using the Go/No-Go Association Task (GNAT) - An Empirical Investigation of the Convergent Validity Between Explicit and Implicit Measures Laura Wiredu, Who Bears the Costs of the UK Soft Drink Tax? An Empirical Study of Medium-Term Effects Thiemo Janßen, From Pictures to Perceptions: Exploring the Strategic Use of Visuals in CSR Reports and the Impact of Regulatory Mandates Vanessa Jeske, Determinants of Corporate Bond Mutual Fund Flows Felix Yumuşak, The Influence of Leadership Style on the Acceptance of Generative AI in the Workplace -The Role of Organizational Commitment, Job Insecurity and Interaction Frequency 831 858 876 904 940 966 985 1009 1028 1053 Published by Junior Management Science e.V. This is an Open Access article distributed under the terms of the CC-BY-4.0 (Attribution 4.0 International). Open Access funding provided by ZBW. ISSN: 2942-1861 Implicit Measurement of the Moral Self-Image Using the Go/No-Go Association Task (GNAT) - An Empirical Investigation of the Convergent Validity Between Explicit and Implicit Measures Louisa Felicitas Bläßer Technical University of Munich Abstract While people are increasingly aware of climate change, many still resist lifestyle changes. Research now focuses on understanding conscious (explicit) and unconscious (implicit) attitudes to encourage sustainable behavior. This thesis used the Go/No-Go Association Task (GNAT) to measure participants’ implicit moral self-image and examine its correlation with an explicit moral self-image questionnaire, indicating convergent validity and effective application of the GNAT as an implicit measure of the moral self-image. After applying exclusion criteria, 68 participants were randomly assigned to two groups with different word lists. Results showed that repeated exposure to fewer words in group A led to little or no correlation, while group B, using more varied words, showed higher correlation and good convergent validity. This demonstrates that the GNAT effectively measures moral self-image when learning effects are avoided. The findings offer insights into implicit attitudes that influence decisions and yield practical implications for different stakeholders. This thesis contributes through its experimental design, adapted exclusion criteria, and sample correction of all perfect responses, validating the GNAT as an implicit measure and offering a foundation for future research. Keywords: convergent validity; explicit measures; go/no-go association task (GNAT); implicit measures; moral self-image 1. Introduction Have you ever ordered something online, used nonrecyclable packaging, chosen a non-organic product, not separated food waste appropriately, or traveled by plane? The answer is likely yes, as we face many sustainable decisions daily. Unsustainable behavior is perceived as immoral, as people are increasingly aware of their impact on climate change (Sachdeva et al., 2015). But why do people engage in unsustainable and, consequently, immoral behaviors? I would like to thank Konrad Kober, my Bachelor’s thesis supervisor, for his dedicated support, continuous encouragement, and constructive feedback throughout the duration of this thesis. I am especially grateful for his valuable assistance and expertise in the experimental design. Additionally, I would like to thank Prof. Dr. Alwine Mohnen from the Chair of Corporate Management at the Technical University of Munich for serving as my examiner. There has been an increasing focus on understanding the psychological drivers that motivate immoral behavior in the last decades, especially with the intensifying global climate crisis (Sachdeva et al., 2015). The ultimate goal is to use this knowledge to nudge people further into being more sustainable (Fischer et al., 2012; Sachdeva et al., 2015). Therefore, it is essential to understand why people behave immorally and how people’s morality can be measured. Traditional explicit measures often fall short of capturing the perception of people’s moral selves. Social desirability biases influence the answers given in such self-report questionnaires (Crowne & Marlowe, 1960). In recent decades, various implicit measures have been developed to measure unconscious attitudes. The most famous method is the Implicit Association Test (IAT), which was further developed into the Go/No-Go Association Task (GNAT). Previous research has already implicitly assessed the moral self-image (perception DOI: https://doi.org/10.5282/jums/v10i4pp966-984 © The Author(s) 2025. Published by Junior Management Science. This is an Open Access article distributed under the terms of the CC-BY-4.0 (Attribution 4.0 International). Open Access funding provided by ZBW. L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984 967 of one’s own morality) with the IAT, yielding promising results (Perugini & Leone, 2009). However, little research has been devoted to the GNAT, and it has only been used once to capture the moral self-image (Ferguson, 2018). This bachelor’s thesis uses the GNAT to measure the implicit moral self-image and to examine the convergent validity (the ability of two measures to capture a joint construct (Carlson & Herdman, 2012)) of this implicit measurement method by correlating it with an explicit moral selfimage questionnaire. It aims to answer the following research question: “To what extent can the Go/No-Go Association Task (GNAT) be effectively applied to measure moral self-image, and is there a correlation between the outcomes of this method and the explicit moral self-image?” The timing for this research is crucial, as it aligns with the growing interest in sustainable practices and the need to deepen the understanding of the internal, implicit motivations behind a behavior change (Mazar & Zhong, 2010; Sachdeva et al., 2015; Schlegelmilch & Simbrunner, 2019). Applying implicit psychological tests, such as the GNAT, to consumer behavior or marketing strategies opens up a new field of research. These measures improve the assessment of implicit attitudes toward products or brands because the implicit test procedures are based on less biased, unconscious answers and reactions and, therefore, are very valuable for subsequent analyses. An effective and validated tool to determine people’s moral self-image could provide essential insights for different stakeholders to influence consumers toward more sustainable and moral choices. This thesis is structured as follows: Chapter 1briefly introduces the topic and aim of this thesis. Chapter 2reviews current research literature, presenting theories of moral behavior, important definitions, and measurement methods, leading into Chapter 3, which covers the methodology and detailed research design of the performed GNAT experiment. Chapters 4and 5present the compelling results and elaborate on the experiment’s key findings, discussing its limitations, suggestions for future research, and implications for practitioners in management and policy. Chapter 6provides a comprehensive summary of the main findings of the experiment, reflecting on the research’s significance. 2. Theoretical Background A wide range of theories are trying to explain why people behave immorally. The rational economic model expects people to behave immorally whenever their potential gain exceeds their expected punishment since it is the best choice economically (Becker, 1968). Following this reasoning, people should behave immorally every time they could potentially gain more than they would lose. In contrast, it can be observed that people intrinsically limit their immorality and avoid too much lying if it threatens their perception of their own morality (moral self-image) (Mazar et al., 2008; Sachdeva et al., 2009). It appears that the idea of an entirely rational person (e.g., homo economicus (Melé & Cantón, 2014)) does not apply to most people and situations. Instead, an internal force seems to restrict people from exploiting the potential benefits of cheating to its full extent (Cornelissen et al., 2013; Mazar et al., 2008). It becomes evident that people face an internal conflict whenever they have an opportunity to cheat (Barkan et al., 2015; Mazar et al., 2008). This ethical dissonance is a state of tension that occurs when people are either tempted to benefit from their immoral behavior or to uphold a positive moral self-image, also known as moral self-concept1(Mazar et al., 2008). Festinger (1957) describes this state as cognitive dissonance and argues that its presence motivates people to subsequent action, which reduces this dissonance. People developed different strategies to engage in immoral behavior to resolve this internal conflict and distressing state, particularly without updating (and depressing) their moral self-image. Erikson (1964) explains this motivating force as the intrinsic need for people to act according to their (moral) identity. Researchers interpret moral identity, defined as “the use of moral values to define the self” (Johnston et al., 2013, p. 209), as a moderator and motivational driver to act morally (Aquino & Reed, 2002; Blasi, 1993; Erikson, 1964). The following section will introduce different theories that investigate why people behave immorally. 2.1. Moral Theories 2.1.1. Self-Concept Maintenance Mazar et al. (2008) propose a theory of self-concept maintenance. They argue that people try to balance maintaining an honest self-concept and gaining from lying. According to Mazar et al. (2008), people would cheat to a certain extent as long as they do not need to update their moral self-concept (of being honest). This compromise allows them to benefit from cheating without negatively impacting their moral self-image. The authors also suggest that people use different techniques to decide on this motivational dilemma and to determine the degree to which cheating aligns with their moral self. A powerful technique is, for instance, self-serving justification (Shalvi et al., 2015). It suggests that people would try to find reasons for questionable behavior to make it seem less immoral when their moral self-image is threatened. Shalvi et al. (2015) distinguish between pre-violation justification (before the immoral action) and post-violation justification (after the immoral action). Pre-violation justification excuses the immoral action and thus reduces the threat to the moral self-concept beforehand (Shalvi et al., 2015). There are several strategies for this pre-violation justification. Examples are ambiguous actions, altruistic cheating, and moral licensing (Shalvi et al., 2015). Whenever the norms and rules for a situation are ambiguous, the actor could invent facts and reasons to justify his2 1These terms can be used interchangeably (Jordan et al., 2015) 2For better readability, only the pronouns “he/him/his” are used throughout this thesis L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984968 actions. If a lie does not cause harm to other people but instead would help to benefit the actor and other people, it is more likely to observe cheating (altruistic cheating) (Erat & Gneezy, 2012). Through moral licensing, people justify their bad behavior with their initial good actions (Merritt et al., 2010; Shalvi et al., 2015). In contrast, post-violation justification is a tool to justify immoral behavior after the action has already been conducted. This could be, for instance, through (partially) confessing or distancing themselves from their action by looking at others’ immoral behavior (Shalvi et al., 2015). People generally try to maintain or even enhance their moral self-image (Jordan et al., 2015; Mazar et al., 2008; Shalvi et al., 2015). They do this by behaving morally or biasing their cognitive perception with examples like selfserving justifications (Monin & Jordan, 2009). Monin and Jordan (2009) argue that people who value morality greatly pay more attention to their moral self-image, and deviations from their moral self-concept impact their self-worth more significantly compared to people with a lower importance on being moral. This moral self-image can also be influenced by previous and current situations. A deviation from their aspired level motivates people to take subsequent actions to reduce this dissonance (Cornelissen et al., 2013; Jordan et al., 2015). Monin and Jordan (2009) refer to that as behaviorgenerating power. To predict peoples’ behavior, their individual moral self-image, which fluctuates and deviates over time (Jordan et al., 2015), must be considered. 2.1.2. Moral Balancing Model There are two contrasting approaches when predicting peoples’ moral actions after they have acted morally or immorally. Either the actor behaves consistently with his initial action, or the subsequent behavior is the opposite of his previous action (moral balancing). Freedman and Fraser (1966) introduced the Foot-In-TheDoor-Technique, which is nowadays widely used in negotiation strategies and a great example of consistent behavior. They elaborate that people who already agreed to do a small favor were more likely to agree to do a second, even larger favor. An example of consistent moral behavior would be if a man returns a lost wallet to the owner after offering a seat in public transport to an elderly woman. The negative case, which still reflects consistent behavior, would be that the man does not offer his seat to the elderly woman and keeps the wallet as well. Mullen and Monin (2016) argue that people show consistent behavior when they focus abstractly on values and their initial behavior. In contrast, people exhibit a balancing behavior when they think more concretely about their initial behavior and what they have accomplished with it. This alternating pattern is described as moral balancing. After a previous immoral action, the actor behaves morally in the subsequent action, or vice versa. This moral balancing model was developed in 1990 by the psychologist Mordecai Nisan. It states that people consider previous behavior when making moral decisions. According to Nisan (1990), people try to balance their current moral self around a fixed personal moral standard (equilibrium). This personal reference point is essential for people as they constantly compare their current state with this self-set standard, which they want to maintain over time (Miller & Effron, 2010). Nisan (1990) assumes that when their moral status drops below a personal tolerable level, people will refrain from doing an immoral action. However, this satisfactory level of morality is lower than the ideal level, and those minimum requirements are determined mainly by a person’s moral identity (Nisan, 1990). Following this reasoning, a person who recently did something immoral would instead choose an altruistic action in order to compensate for the previously generated deficit in his own moral balance (moral cleansing). A person who is currently in moral surplus would be more likely to perform a subsequent selfish action (moral licensing) (Nisan, 1990). In other words, balancing happens when a moral initial behavior leads to the opposite in a subsequent behavior (Jordan et al., 2011; Mullen & Monin, 2016; Zhong et al., 2010). When balancing a previous action, these two directions can be observed: moral cleansing and moral licensing. Moral Cleansing Moral cleansing (or moral compensation) happens when a previous immoral behavior causes a subsequent moral behavior (Mullen & Monin, 2016; Perkins et al., 2024). This can be explained by an analogy of a moral bank account, the moral credits model (Perkins et al., 2024). If a person’s metaphorical moral bank account is in deficit, he wants to rebalance it with a subsequent moral behavior (Nisan, 1990). This could be done by performing a morally good action or refraining from immoral actions, such as cheating (Cornelissen et al., 2013). Continuing with the previous example, a man who did not offer his seat on the bus to an elderly woman would be more likely to return a lost wallet to its owner to compensate for this deficit in his moral balance. Researchers explain this effect through people’s motivation and willingness to invest effort to repair their shortfalls (Jacobsen et al., 2018). Additionally, there is strong evidence that people need to physically cleanse themselves after behaving immorally. Zhong and Liljenquist (2006) show that people who recall an immoral act would be more likely to choose antiseptic wipes compared to other products. They explain that the participants need to wash away their sins and cleanse themselves after their moral purity has been threatened. Moral Licensing The moral licensing effect describes the contrasting and somewhat counterintuitive observation: Good previous behavior leads to less positive or even bad behavior. In other words, people justify their bad behavior with their previous L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984 969 good action (Jacobsen et al., 2018; Merritt et al., 2010). Moral licensing can be explained from two different perspectives: the moral credits model and the moral credentials model. The moral credits model explains the licensing effect as people accumulate credits in their hypothetical moral bank account when they do something good. They can use these credits and “withdraw” them to justify subsequent negative behavior while maintaining an overall positive balance (Effron & Monin, 2010; Merritt et al., 2010; Miller & Effron, 2010). Moral licensing starts with a surplus in the moral bank account and withdraws credits to allow people to perform a negative action (Perkins et al., 2024). The second explanation, the moral credentials model, explains the moral licensing effect with a different interpretation of the subsequent behavior. According to Monin and Miller (2001), people are less likely to interpret their subsequent behavior as immoral after they have performed an initial moral act. Instead of earning a right to perform this immoral act without punishment, the initial moral behavior has provided a lens through which the following behavior is interpreted differently (Mullen & Monin, 2016). This process is more likely when the subsequent behavior is ambiguous and can be interpreted positively (Mullen & Monin, 2016). For example, by recommending a woman for one job, people built positive credentials as being someone without prejudice and were more willing to express that a man was better suited for a second job (Monin & Miller, 2001). In this experiment by Monin and Miller (2001), the second behavior was ambiguous. It could be explained by illegitimate or legitimate motives (sexism or pragmatism). The credentials of not being sexist, e.g., established through actively recommending a woman for the first job, help to interpret the second action positively, e.g., favoring a man for the second job, due to the actor’s history, without affecting the actor’s moral self-image (Monin & Jordan, 2009; Monin & Miller, 2001). Both models explain that a previous moral action can lead to immoral or questionable behavior later on. The key difference is that in the moral credits model, the actor is fully aware of the second action’s immorality but decides to afford this decrease in his overall moral balance. In contrast, in the moral credentials model, the positive first action helps to disambiguate and interpret the second action differently (Monin & Jordan, 2009). To clarify this tension between the two models, Monin and Jordan (2009) suggest that the moral credits model is at work in unambiguous cases, where the meaning of the target behavior is clearly interpreted as immoral and unaffected by the previous action, while the moral credentials apply for ambiguous cases. However, both models predict the same behavior and support the importance of acknowledging a dynamic moral self-image. These models suggest that recent actions shape a person’s moral self-image and influence his future moral behavior (Monin & Jordan, 2009). Moral Self-Image in the Moral Balancing Model The moral credits model describes a mechanism for people to balance their moral or immoral behavior with an accumulated or depleted moral bank account, reflecting the increase or decrease of the moral self-image, respectively (Merritt et al., 2010; Zhong et al., 2010). This enables people to repair their moral self-image by compensating for their selfish actions afterward (moral cleansing) (Perkins et al., 2024; Schlegelmilch & Simbrunner, 2019) or using their bolstered moral self-image (from a previous action) to perform a subsequent immoral act (moral licensing) (Cornelissen et al., 2013; Effron & Monin, 2010; Monin & Jordan, 2009; Nisan, 1990). This emphasizes that the moral self-image plays a central role in moral decision-making. Its discrepancies from the actor’s personal standard (equilibrium) motivate balancing behavior (Nisan, 1990). While the balancing could be observed, and there is empirical evidence (Cornelissen et al., 2013; Lee & Hsieh, 2013; Ploner & Regner, 2013), measuring the moral self-image is also important. Cornelissen et al. (2013) first attempted to measure the moral self-image with a scale of differences between the desired and the perceived moral self. Jordan et al. (2015) developed this scale further to provide a tool that actively and explicitly measures the moral self-image. However, there is still little empirical evidence of the deviations and fluctuations in time of the moral self-image (Perkins et al., 2024). 2.2. Definitions As the previous section illustrated, different theories try to explain immoral behavior. Given the important role of the moral self, moral psychology increasingly shifted its focus to it to extend moral reasoning and predict behavior (Monin & Jordan, 2009). Before introducing some measurement methods, two terms must be defined accordingly in the context of the moral self: moral identity and moral self-image. 2.2.1. Moral Identity Aquino and Reed (2002) define moral identity “as a selfconception organized around a set of moral traits.” (p. 1424). They suggest that moral identity is relatively stable over time and identify two dimensions: Internalization and Symbolization. Internalization describes how important it is for a person to have (nine) moral traits: “caring, compassionate, fair, friendly, generous, helpful, hardworking, honest, and kind.” (Aquino & Reed, 2002, p. 1426). Symbolization describes the degree to which a person wants to be seen as moral or demonstrate these traits through their actions to others (Aquino & Reed, 2002). The researchers propose that people behave morally when they assess a specific moral trait as essential for their self-concept. Moral identity should, therefore, be a motivational driver for acting consistently (Aquino & Reed, 2002) and is the basis for moral motivation (Erikson, 1964; Nisan, 1990). To measure moral identity actively, Aquino and Reed (2002) asked participants to rate how important it is for L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984970 them to possess these traits (Internalization) and if they participate in activities (e.g., hobbies), wear clothes or buy products that identify them as having these characteristics (Symbolization) (Aquino & Reed, 2002). 2.2.2. Moral Self-Image Jordan et al. (2015) introduced the concept of the moral self-image to explain how the self-perception of the individual’s morality fluctuates. They define the moral self-image as the malleable and dynamic moral self-concept. A person’s moral self-image can be described as the answer to the question “’How moral am I?’” (Monin & Jordan, 2009, p. 347). This reflects exactly how morally individuals see themselves at any point in time. The moral self-image is part of the dynamic working self-concept, the malleable part of the self (Jordan et al., 2015). It is completely subjective and only measures how moral persons perceive themselves (Jordan et al., 2015). Monin and Jordan (2009) highlight that individuals can constantly show differences in their moral self-image, as it can be lowered or bolstered through previous actions, which motivates subsequent behavior. The researchers agreed that the moral self-image has a behaviorgenerating power (Jordan et al., 2015; Monin & Jordan, 2009). Due to the lack of previous empirical measurement methods, Jordan et al. (2015) introduced an explicit nine-point Likert scale to measure the moral self-image as highly connected to the traits of a typical moral person (based on the traits introduced by Aquino and Reed (2002)). This elaborated scale has been used as an explicit moral self-image measure in previous research (Ferguson, 2018) to investigate the convergent validity, which “reflects the extent to which two measures capture a common construct.” (Carlson & Herdman, 2012, p. 18). 2.3. Measurement Methods 2.3.1. Explicit vs. Implicit Measures In order to understand, predict, and control human behavior, psychologists have been trying to measure people’s cognitive processes, attitudes, and self-image (de Houwer, 2006). A straightforward approach is to conduct a survey and actively ask participants about their opinions toward a situation or an object. This explicit method is easy to conduct, comprehensible, and easily measured (de Houwer, 2006). The most common approach for measuring the moral identity or the moral self-image is letting participants rate different personality traits on a Likert scale. This approach assesses how important these personality traits are for them (moral identity) (Aquino & Reed, 2002) or how much they are already fulfilling some characteristics compared to the person they want to be (moral self-image) (Jordan et al., 2015). However, despite their wide use (Asendorpf et al., 2002), these surveys might be subject to impression management (Paulhus, 1984), which means that participants include answers to be seen in a favorable light. When being asked, people are influenced by concerns about their self-presentation (Doherty & Schlenker, 1991; Schnabel et al., 2007) and social desirability (Crowne & Marlowe, 1960), which could incentivize them to give socially conform answers to the interviewer. Additionally, these surveys are limited to the introspective personality and might not reflect a person’s entire personality (Schnabel et al., 2007). Because of these disadvantages, new implicit measures have been developed. Initially introduced in social psychology, implicit measures are now widely applied across different disciplines and commonly used in psychology (de Houwer et al., 2009). But what is an implicit measure exactly? de Houwer (2006) suggests using the synonym automatic when explaining implicit effects. A process can be called automatic when it still operates, although the participants are unaware of results, stimulus, or procedure, do not have a specific goal, or do not invest many cognitive resources (de Houwer, 2006). Following this argumentation, the same should apply to an implicit measure. This measure intends to get an immediate (automatic) response from people without them being aware of it or involving their cognitive thinking. Such an implicit or indirect measurement method could be used to measure a person’s unconscious attitude (Bartels & Schoenrade, 2022; Schnabel et al., 2007). 2.3.2. Implicit Association Test (IAT) As a way to avoid biases of explicit measurement, Greenwald et al. developed the Implicit Association Test (IAT) in 1998. This test aims to measure the relative implicit association strength of two contrasting concepts (target categories) (e.g., FLOWER-INSECT) and PLEASANT-UNPLEASANT3 (evaluation attribute) (Greenwald et al., 1998)). In their initial experiment, all participants should react by pressing an assigned key on the left or right. In different blocks, a target category and an attribute are assigned to one key. For example, the left key is assigned to FLOWER +PLEASANT, whereas the right key is assigned to INSECT +UNPLEASANT. Whenever a stimulus (either a FLOWER, an INSECT, a PLEASANT, or an UNPLEASANT word) appears on the screen, the participant should press the assigned key (Greenwald et al., 1998). In other words, the stimuli should be classified into four mutually exclusive categories (FLOWER, INSECT, PLEASANT, or UNPLEASANT) (Greenwald et al., 1998; Schimmack, 2021). Participants are required to distinguish between words referring to INSECT +FLOWER and words referring to PLEASANT +UNPLEASANT words. For instance, with the assigned keys described above, the stimuli tulip or happy should be assigned to the left key (FLOWER +PLEASANT), whereas wasp or rotten should be assigned to the right key (INSECT +UNPLEASANT) (Greenwald et al., 1998). After the first combined task of discrimination between the target categories and evaluation attributes, a second block with a reversed combined 3The categories (target categories and evaluative attributes) are written in capital letters. L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984 971 task was conducted. In this reversed combined task, one key was assigned to INSECT +PLEASANT, and the other key was assigned to FLOWER +UNPLEASANT. It is fundamentally assumed that participants’ response time is faster when the association between the target category and the evaluative attribute is stronger (de Houwer, 2006; de Houwer, 2001; Greenwald et al., 1998; Johnston et al., 2013). This means that the pairing FLOWER +PLEASANT should be easier compared to the INSECT +PLEASANT pairing if the association between FLOWER and PLEASANT words is stronger (de Houwer, 2006; de Houwer, 2001; Greenwald et al., 1998; Johnston et al., 2013). With this experiment, Greenwald et al. (1998) provided significant results demonstrating that the incompatible combination of INSECT +PLEASANT was more challenging to confirm, and participants had longer response times compared to the compatible combination of FLOWER +PLEASANT. The authors explain this effect with a stronger association and familiarity between FLOWER +PLEASANT words and than between INSECT +PLEASANT words, indicating a more positive attitude toward FLOWERS than INSECTS (Greenwald et al., 1998). While the IAT was quite revolutionary, set new standards, and offered new opportunities, criticism about the IAT and implicit measures, in general, needs to be addressed. There are concerns about its construct validity (Schimmack, 2021), its capability to predict behavior (Bartels & Schoenrade, 2022; Brownstein et al., 2020), and its temporal instability (Brownstein et al., 2020; Schimmack, 2021). Schimmack (2021) raises concerns that there is no consensus about what the IAT measures and that it is difficult to compare if it measures something different than explicit measures. This problem has been recognized by the dual attitudes model (Wilson et al., 2000) (also known as the double dissociation model (Perugini, 2005)), which clearly distinguishes implicit and explicit attitudes into two systems (Wilson et al., 2000). According to this model, implicit measures predict impulsive, spontaneous, and automatic behavior, while explicit measures predict controlled and conscious behavior (Wilson et al., 2000). In agreement with the double dissociation model, Johnston et al. (2013) suggest that implicit and explicit attitudes can only be measured with implicit or explicit measurement methods, respectively. The contrasting perspective describes an additive view, where both types of attitude describe a “different portion of variance in the same criterion” (Perugini, 2005, p. 29). Fazio and Olson (2003) argue that both measures assess the same construct and explain a potential difference between the measurement methods with participants’ deliberative control strategies. The IAT and other implicit measurement methods can add predictive insights to self-report measures (Brownstein et al., 2020) and investigate the implicit moral self-image. Considering the concerns about the behavior predictability of the IAT, there are several studies about predictions of voting behavior with the IAT. For example, Friese et al. (2007) successfully predicted the voting behavior and attitudes for the German parliamentary elections in 2002. They used a singletarget IAT, where one key was assigned to an evaluative attribute and the target category, while the second key was assigned only to the opposing attribute. This single-target IAT yielded excellent validity in predicting voting behavior (Friese et al., 2007). 2.3.3. Go/No-Go Association Task (GNAT) To expand the use of implicit measurement methods and the Implicit Association Test (IAT), a new method was developed by Nosek and Banaji (2001). They introduced the Go/No-Go Association Task (GNAT), which mainly focuses on the error rate as the dependent variable to measure the strength of implicit associations (Greenwald et al., 1998; Nosek & Banaji, 2001). Unlike the previously known implicit methods, in the GNAT, only a single concept (target category /e.g., ME) is evaluated considering one attribute dimension (evaluative attribute /e.g., GOOD) (Bassett & Dabbs, 2005; Ferguson, 2018; Nosek & Banaji, 2001). The GNAT does not need two contrasting concepts (two target categories); hence, it is more flexible and can reveal new aspects of social cognition. Another difference is that only one response (key) is required for the GNAT (Nosek & Banaji, 2001), simplifying the experimental setup. During the task, a target stimulus (signal item) or a distracter stimulus (noise item) is presented on the screen for some milliseconds. Following the experimental design and example of Ferguson (2018) for the GNAT, when a stimulus word that is similar to either the attribute (e.g., GOOD) or the target category (e.g., ME) is shown, the participant should press the space bar (or any key) to give a Go response. On the contrary, when the word on the center of the screen does not match the attribute or the target category (=distractor), no response (No-Go) is required, and the participant should refrain from pressing any key. According to Nosek and Banaji (2001), the strength of association in the GNAT is determined by how well stimuli words associated with the target category and the attribute (for example, ME +GOOD) are distinguished from distractor items unrelated to these concepts. The authors suggest that the sensitivity between the pairing conditions (in this example, ME +GOOD or ME +BAD) illustrates the strength of the association between the target category and the evaluative attribute (Greenwald et al., 1998; Nosek & Banaji, 2001). In general, the faster and/or the fewer errors (and therefore easier) the response, the stronger the association. Greenwald et al. (1998) and Nosek and Banaji (2001) argue that both error rates and average response times can provide information about task performance due to a speed-accuracy tradeoff. Nevertheless, most implicit measures focus solely on response times “as the dependent variable and therefore may lose relevant information contained in error rates.” (Nosek & Banaji, 2001, p. 628). Previous research in psychology has used the GNAT to investigate different implicit attitudes. For example, implicit spider fear associations (Teachman, 2007), implicit bias in L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984972 phrasing drug addiction (Ashford et al., 2019), and implicit attractiveness beliefs of people who are constantly worrying about their physical appearance (Buhlmann et al., 2011). Those studies provide significant insights into the reliability and validity of the GNAT. In those applications, the researchers suggest that the GNAT is an effective tool for measuring involuntary associations and might help measure implicit associations, especially since it does not require a comparison category on a second key (Buhlmann et al., 2011; Teachman, 2007; Williams & Kaufmann, 2012). Williams and Kaufmann (2012) specifically investigated the reliability of the GNAT. They recommend a minimum of 40 trials per block for minimally acceptable reliability and at least 80 trials per block for good reliability. They argue that the GNAT is a valuable tool with many advantages and should be used in further research. (Williams & Kaufmann, 2012). By omitting a comparison concept (unlike the IAT), the GNAT can use distractor items more flexibly, allowing for a direct assessment of the attitude (Nosek & Banaji, 2001). “In addition, the GNAT may be less susceptible to errors introduced by term valence and less biased by response criteria than reaction time-based techniques.” (Boldero et al., 2007, p. 354). The convergence between implicit and explicit personality traits was examined by Boldero et al. (2007). They support the reliability and convergent validity of the GNAT when controlling the systematic variance of the GNAT. However, as their explicit measure was conducted before the implicit GNAT, it is possible that this inflated their associations and, thus, the correlation between both measures (Boldero et al., 2007). Unfortunately, sufficient research in the moral domain, including the GNAT, has not yet been conducted. Previous studies focused on predictions about moral behavior with implicit measurement methods, like the IAT (Perugini & Leone, 2009). Another study tried to measure the moral identity with the IAT (Johnston et al., 2013). The only known publication of an application of the GNAT to measure the implicit moral self-image is a dissertation by Ferguson (2018), who failed to show moral balancing effects. Implicit measures, such as the IAT and GNAT, can easily be implemented into software and be used as a portable version on mobile devices, thus overcoming the limitation of requiring a local computer (Bassett & Dabbs, 2005; Dabbs et al., 2003). This provides several advantages. Firstly, it tests participants in a more natural setting outside a laboratory (Bassett & Dabbs, 2005; Dabbs et al., 2003). This real-life setting could lower the feeling of being observed during the experiment and lead to more honest and impulsive answers. Secondly, Bassett and Dabbs (2005) argue that portable versions could be used to measure malleable attitudes at different times and important events. Thirdly, having a mobile and portable version of these tests could help to reach populations that would usually not participate in laboratory experiments. Fourthly, more people could participate because the effort needed is much less, as they are no longer required to go to the laboratory (Bassett & Dabbs, 2005). 3. Methodology 3.1. Research Design This bachelor’s thesis aims to apply an existing implicit measurement method, the Go/No-Go Association Task (GNAT), to measure the participants’ implicit moral selfimage and analyze these results for a correlation with their explicit moral self-image. It intends to investigate the effectiveness of the GNAT with the convergent validity between the implicit and explicit measures. As explained in Chapter 2, most of the previous research to measure the moral self-image is not based on the GNAT (e.g., Johnston et al., 2013; Perugini and Leone, 2009). The underlying parameters of the performed experiment and the test setup were as follows: 3.1.1. Experimental Design Objective & Material & Groups & Variables Objective: Measure the implicit moral self-image and analyze the correlation with the explicit moral self-image (convergent validity). Material and Groups: Inspired by Ferguson (2018), who failed to provide significant evidence for moral balancing using the GNAT, her experiment was reproduced in group A using the exact same stimuli (six words for GOOD/BAD and four words for ME/OTHER) to be consistent with her research method. In group B, the list of words was extended for potentially stronger effects. Group A: These stimuli words were replicated from Ferguson (2018): ME: me, I, my, myself OTHER: other, others, them, they GOOD: good, honest, faithful, modest, sincere, altruist BAD: bad, dishonest, deceptive, pretentious, arrogant, cheater Group B: The extended stimuli words (attributes) were selected from different scientific papers and measured in a pre-study according to their evaluative intensity (see Appendix V.I Pre-Study GOOD & BAD). This group used 20 stimuli words for each attribute and 15 for each concept category (ME & OTHER). To be consistent with Ferguson (2018) and the previous group A, the 40 attribute stimuli (57.14%) and 30 concept category stimuli (42.86%) in relation to the total stimuli of 70 in group B was almost equal to the 12 attribute stimuli (60%) and 8 concept category stimuli (40%) initially introduced by Ferguson (2018). The used stimuli words in group B were: ME: me, I, my, myself, mine, self, personally, oneself, person, intrinsic, own, individual, ego, inner essence, inner self L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984 973 Table 1: Possible outcomes of the Go/No-Go Association Task, inspired by “Trial of yes-no experiment” in Macmillan (2002) Response Go No-Go Stimulus Signal (target item) Hit ◦Miss × Noise (distractor item) False Alarm ×Correct Rejection ◦ OTHER: other, others, them, they, their, themselves, theirs, his, him, her, anybody, anyone, those people, persons, the individuals GOOD: caring, fair, compassionate, friendly, hardworking, generous, helpful, kind, honest, faithful, altruist, modest, sincere, genuine, joyful, patient, grateful, loyal, forgiving, respectful BAD: hostile, unfair, lazy, unhelpful, ruthless, selfish, evil, brutal, hateful, angry, impatient, bad, dishonest, deceptive, pretentious, arrogant, cheater, disrespectful, disloyal, egocentric Variables: The stimuli word lists (group A or B) served as independent variables. The dependent variables were Hit- /False-Alarm rates. Blocks & Trials & Stimuli The GNAT was divided into two blocks. The target category ME was paired with the attribute GOOD in the starting block. In the second block, the same target category, ME, was paired with the opposite attribute, BAD. Each block comprised a total of 96 trials for group A or 86 trials for group B. Both blocks started with 16 practice trials (not considered in the analysis), followed by a reminder screen, before proceeding to the 80 critical trials (considered in the analysis) for group A or 70 critical trials for group B. A trial started when a stimulus word from one of the categories (ME, OTHER, GOOD, BAD) emerged on the screen. It ended when the word disappeared. As a constant reminder of the current combination (pairing) in each block, labels for the target category (ME) and the attribute (GOOD or BAD) remained on the screen’s upper left and right corners. The labels and stimuli items were displayed in black font against a white screen. The participants were advised to either (1) give a Go response by quickly pressing the space bar if the stimulus word displayed could be categorized into one of the two labeled categories (signal item) or (2) refrain from pressing any key (No-Go response) for words that could not be categorized (noise items). The stimulus word appeared in the center of the screen and remained visible until the response deadline was reached or a key was pressed. The subsequent trial started when the participant pressed the space bar or after the response time ran out. Like Nosek and Banaji (2001), the opposing category (OTHER) or the alternate attribute served as distracter trials (noise). For example, when GOOD was the signal, BAD was the noise, and vice versa. A signal-to-noise ratio of 1:1 was held constant for all trials and both groups. The 20 stimuli words Ferguson (2018) used for the critical trials in group A were selected randomly. Each word was repeated four times for a total of 80 trials, which was in the range of 50 to 80, yielding sufficient and good reliability, as Williams and Kaufmann (2012) recommended. For group B, the stimuli items were chosen randomly and selected without repetition from the four categories (ME, OTHER, GOOD, BAD) to reach 70 trials. Within both groups, each block (pairing) consisted of an equal number of words in order to minimize learning effects, as was again advised by Williams and Kaufmann (2012). The critical trials used the complete set of stimuli words, which were selected randomly and appeared in random order. Additionally, for the word list in group B, a small-scale pre-study (Appendix V.I Pre-Study GOOD & BAD) on the evaluative intensity of these words was conducted beforehand. This ensured that the words’ evaluative intensity was a) strong enough to yield sufficient results and b) similar between the words, which was strongly suggested by Nosek and Banaji (2001). Response Deadline & Feedback The participants had to categorize the stimuli words as quickly and accurately as possible during the short time displayed on the screen. The response deadline was constant at 700 milliseconds (ms) across all trials and blocks, following the recommended range of 500ms to 850ms by Nosek and Banaji (2001). The interstimulus interval between two trials was held constant at 500ms. During this interstimulus interval, immediate feedback on performance accuracy was provided. Trials where signal items were accurately identified (Hit) or noise items were correctly ignored (Correct Rejection) were recorded as correct responses, indicated by a green “O” appearing. Trials were marked as errors when noise items were mistakenly identified as signals (False Alarm) or signal items were overlooked (Miss). For these trials, a red “X” was displayed in the center of the screen after the stimulus item disappeared. The possible outcomes and corresponding feedback are visualized in Table 1. 3.1.2. Statistical Analysis The statistical analysis of the GNAT was based on the signal detection theory, first introduced by Green and Swets (1966), as cited in Macmillan (2002), and as previous experiments using the GNAT already have done (Ferguson, 2018; L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984980 to measure moral self-image, and is there a correlation between the outcomes of this method and the explicit moral self-image?” should be answered separately for both groups. Group A has shown little, even negative, to no correlation between the two measures, which indicates an ineffective application of the GNAT in measuring the moral self-image. For group B, there is, in fact, a significant positive correlation between the implicit and explicit measures, indicating an effective application because learning effects were avoided by using various stimuli without repetitions. This thesis has closed the gap in previous research, as it has proven that the GNAT is an effective instrument for measuring the moral self-image if learning effects are considered in the research design. It suggests utilizing various stimuli words without repetition instead of repeating fewer stimuli words. This finding has highlighted the extraordinary importance of the experimental design for yielding significant results. However, the GNAT still has some limitations, and further research investigating the GNAT’s ability to predict behavior and convergent validity is highly recommended, whereas the first results were already quite promising. This thesis was the first study conducted to assess the convergent validity of the GNAT for capturing the moral selfimage. 5.2. Limitation of Results & Suggestions for Future Research It is important to emphasize the limitations of this research, especially to enable future research. First of all, investigating the correlation between explicit and implicit measures could be misleading and ambiguous due to various explanations for both attitudes and missing consensus about its significance. According to Perugini (2005), a low correlation could suggest insufficient convergence validity between the two measurement methods. However, it could also be seen as evidence for the double dissociation, supporting the theory that a dual system of attitudes exists and that those attitudes are unrelated. Instead, it might be more expressive to separately investigate the capability of both measures to predict behavior (Perugini, 2005) and explore this in further experiments. Validating the GNAT as a moral self-image measure according to its ability to predict behavior would provide further insights and could support its usefulness (Perugini, 2005). Nevertheless, Carlson and Herdman (2012) generally advised using measures with convergent validities above 0.7 (r>0.7) and avoiding those below 0.5 (r<0.5). In the context of implicit measures and especially regarding the moral self-image, one might have to deviate from this suggestion, as there are various reasons for the low correlation. More research should be conducted to review these recommendations. Future research could extend the experiment over a longer timeframe and ask participants about their moral self-image before and after specific moral or immoral actions. Since the measured moral self-image is a snapshot (in daily life) and can deviate in time (Jordan et al., 2015; Monin & Jordan, 2009), tests could be conducted at regular intervals with the identical test group to analyze the ranges of variation around their fixed personal reference, potentially initiating moral balancing behavior (Jordan et al., 2015; Nisan, 1990). Regarding the experimental design, further investigation is needed about the effect of the block order on the moral self-image, as previous research has already suggested that it might have an impact (Greenwald et al., 1998,2003; Nosek & Banaji, 2001). It would be interesting to see if a reversed order of the blocks yields the same results and convergent validity. An extensive experiment should investigate the occurrence of learning effects due to increased familiarity with the procedure and stimuli words in the second block. This could be done by changing the order of the blocks ME +GOOD and ME +BAD in different groups. When investigating the convergent validity, an additional aspect is changing the order of the explicit questionnaire. Asking participants explicitly about their moral self-image before and after the GNAT could examine potential priming effects (of both explicit questionnaire and implicit GNAT), eventually inflating the correlation (Boldero et al., 2007). Further, including the questionnaire twice (before and after the GNAT) could provide insights into how the individual’s performance in the GNAT changes the explicit moral selfimage. Additionally, it would make sense to expand this sample and conduct the experiment with more participants. Simple deviations already greatly impacted the coefficients obtained in the current sample of 30 (group A) to 38 participants (group B). Additionally, the uneven size of both groups raised concerns about the comparison. Most participants were native German speakers (76.67% in group A and 76.32% in group B), which differentiated from previous experiments, where only native English speakers were asked to perform the GNAT in English (Ferguson, 2018; Nosek & Banaji, 2001; Williams & Kaufmann, 2012). Repeating the same experiment in the first language of each participant could eventually provide more robust results, as the implicit association could be much stronger, investigating more “natural” and more accessible associations. In addition, the portable nature of the GNAT (Bassett & Dabbs, 2005), as conducted in this online experiment, allowed for measuring the moral self-image of participants in everyday life. This setting enabled participants to perform the experiment in a familiar environment, without mandatory seminars, away from an unpleasant and stressful (graded) laboratory atmosphere, and whenever they felt ready and comfortable to process the test. Participants were only instructed to perform the test to support a bachelor’s thesis in assessing error rates and reaction times when categorizing English words without telling them the true purpose of this experiment. This more relaxing environment might have supported less biased, unconscious responses and yielded real-life insights into participants’ moral selfimage without them being aware of it. However, this could also have led to participants not taking the test seriously and giving less concentrated responses than in a controlled L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984 981 laboratory environment, which made it harder to compare the results with previous studies. Furthermore, the differences in results obtained through mobile and stationary devices were not investigated due to the limited scope of this bachelor’s thesis and the small sample size. Examining this further with a larger sample is highly recommended, as the experimental Go response differed for both versions. Participants using touch screens had to tap exactly on the word, whereas participants using laptops had to press the space bar. Moreover, due to the limited scope of this bachelor’s thesis and the partially experienced technical discrepancies, a detailed analysis of the reaction times (response latencies), according to Cohen (2013), was omitted. An analysis applying the signal detection theory based on reaction times is proposed in Appendix V.IV Implicit Moral Self-Image Score Based on Reaction Times. However, an analysis based on Cohen (2013) and the often-used algorithm developed by Greenwald et al. (2003), relying on mean latencies and standard deviation of all reaction times, is strongly recommended in the future as it could hold important insights and support the GNAT’s validity by confirming the results and correlations obtained. 5.3. Future Implications After introducing the IAT and the GNAT, a new field of research opened up for assessing people’s implicit attitudes. This could not only be used for psychological counseling (Asendorpf et al., 2002) but can also provide valuable and hidden insights into behavioral economics and for different economic stakeholders, e.g., companies, employees, investors, and policymakers. Companies could use implicit measures to gain deeper insights into their stakeholders and to predict consumer behavior, especially when bringing this into the context of the moral balancing theory. They could exploit the knowledge about the compensatory balancing behavior of people after an initial immoral act. For example, companies could offer customers the option to donate to a prosocial charity after a self-centered purchase (e.g., a carbon-intensive flight for vacation), as adapted from Schlegelmilch and Simbrunner (2019). This allows customers to compensate for a decreased moral account with a subsequent good action (moral cleansing). Moral licensing effects could be utilized, for instance, by allowing people to demonstrate moral behavior (e.g., donating to a charity) before offering them a carbonintensive flight (adapted from Schlegelmilch and Simbrunner (2019)). Both effects influence the customer’s purchase decision, giving them a better feeling after balancing or justifying their self-centered purchase, potentially leading to a higher demand in the future. Additionally, “[i]mplementing donation options in a web shop is an easy way for a company to signal that it is socially responsible.“ (Schlegelmilch & Simbrunner, 2019, p. 551). In Marketing, companies could cluster their target group, as people with higher importance on their morality have a higher incentive to maintain their moral self-image and are more likely to be influenced by (negative) deviations (Aquino & Reed, 2002; Monin & Jordan, 2009). If a person with a high importance on morality performs an immoral act, it has a larger negative impact on his moral selfimage, and he is more likely to compensate for it afterward (Monin & Jordan, 2009). This could be utilized for personalized advertisements, pricing, and influencing people. As the GNAT reveals deeper, unconscious attitudes and feelings toward brands or products, it could be applied to more generic and strategic decisions, such as product development process, brand positioning, pricing, and improved market research. Assessing the implicit moral self-image may serve as an efficient tool in the recruitment process (Asendorpf et al., 2002), potentially revealing the implicit biases and attitudes of both recruiters and prospective employees. This could be a helpful instrument when selecting employees as it, for example, could examine the accordance with company values. Kim (2003) discovered that participants could successfully develop strategies to fake the results in an IAT when instructed to react slower. Remarkably, this was not done spontaneously. This can affect the argument for using implicit measures, such as the IAT or GNAT, when selecting employees, as they could improve their performance if they understood the underlying mechanism and did this test more than once (Kim, 2003). Asendorpf et al. (2002) raised concerns about the ethical use of implicit measures as they reveal participants’ involuntary answers and are not under their own control. This data, therefore, needs to be handled carefully, and sufficient data security should be implemented. Investors could use this measure to uncover the implicit attitudes and morality traits of founders or companies’ top management (e.g., CEO). They could assess the morality of these managers and draw conclusions for their investment decisions. For example, investing only in honest (or moral) founder teams potentially implies a transparent and more honest environment, exchange of information, and prospective company success. Policymakers could benefit from an effective tool to measure peoples’ moral self-image and implicit attitudes to understand their concerns and reasonings. This could be applied to effectively nudge them toward more sustainable behavior by implementing reasonable restrictions and policies. Additionally, this could affect elections and politics, as voters feel better understood by politicians. 6. Conclusion An effective tool to measure people’s moral self-image provides valuable insights, extending the research in behavioral economics and moral balancing (Mazar & Zhong, 2010). This thesis explored the effectiveness of the GNAT by examining the convergent validity between the GNAT and the explicit moral self-image questionnaire. It discovered that a lower correlation between implicit and explicit measures, thus a lower convergent validity, was obtained if a set of only 20 stimuli words were repeated four times in each block L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984982 (group A). This was explained by the participants’ learning effects throughout the experiment, decreasing the convergent validity and effectiveness of the GNAT. For this reason, this thesis postulates using different stimuli without repetition instead, to avoid learning effects. This approach yielded a higher convergent validity (correlation) between implicit and explicit measures of the moral self-image, indicating higher effectiveness of the GNAT (group B). These findings are crucial to effectively measuring the moral self-image, which is essential later on to investigate and understand its changes when aiming to predict moral behavior. This knowledge can be used for practical applications in behavioral economics, such as corporate management or policymaking. Furthermore, this thesis contributes to research with its experimental design and analysis of the GNAT, including the adapted exclusion criteria, a sample correction (of all perfect response rates), and a first attempt to analyze the results based on reaction times (see Appendix V.IV Implicit Moral Self-Image Score Based on Reaction Times). This offers promising approaches for future research, which should particularly focus on a joint analysis of reaction times and response rates, experimental adjustments of the GNAT (e.g., changing the block order), and measuring the participant’s moral self-image over a longer timeframe. Further research is necessary to validate the power of the GNAT to predict (e.g., consistent or balancing) behavior before bringing it into practical use in consumer behavior, for instance, by nudging people toward purchasing a more expensive but sustainable product. However, the GNAT yields promising validity and insights into understanding people’s implicit moral self-image and actions. This can be a powerful tool in the future, especially since it can be conducted entirely online, which helps assess the moral self-image in daily life (outside a laboratory environment). Next time you order something online, use non-recyclable packaging, choose non-organic products, do not separate food waste appropriately, or book a plane ticket, remember the effect of these decisions on the environment and ultimately on your moral self-image, which influences future behavior. References Aquino, K., & Reed, A. (2002). The self-importance of moral identity. Journal of Personality and Social Psychology,83(6), 1423–1440. https://d oi.org/10.1037/0022-3514.83.6.1423 Asendorpf, J. B., Banse, R., & Mücke, D. (2002). Double dissociation between implicit and explicit personality self-concept: The case of shy behavior. Journal of Personality and Social Psychology,83(2), 380–393. https://doi.org/10.1037/0022-3514.83.2.380 Ashford, R. D., Brown, A. M., & Curtis, B. (2019). The Language of Substance Use and Recovery: Novel Use of the Go/No–Go Association Task to Measure Implicit Bias. Health Communication,34(11), 1296– 1302. https://doi.org/10.1080/10410236.2018.1481709 Asuero, A. G., Sayago, A., & González, A. G. (2006). The Correlation Coefficient: An Overview. Critical Reviews in Analytical Chemistry, 36(1), 41–59. https://doi.org/10.1080/10408340500526766 Barkan, R., Ayal, S., & Ariely, D. (2015). Ethical dissonance, justifications, and moral behavior. Current Opinion in Psychology,6, 157–161. https://doi.org/10.1016/j.copsyc.2015.08.001 Bartels, J. M., & Schoenrade, P. (2022). The Implicit Association Test in Introductory Psychology Textbooks: Blind Spot for Controversy. Psychology Learning & Teaching,21(2), 113–125. https://doi.org/1 0.1177/14757257211055200 Bassett, J. F., & Dabbs, J. M. (2005). A portable version of the go/no-go association task (GNAT). Behavior Research Methods,37(3), 506– 512. https://doi.org/10.3758/BF03192721 Becker, G. S. (1968). Crime and Punishment: An Economic Approach. Journal of Political Economy,76(2), 169–217. https://doi.org/10.10 86/259394 Blasi, A. (1993). The Development of Identity: Some Implications for Moral Functioning. In G. G. Noam & T. E. Wren (Eds.), Studies in contemporary German social thought. Moral self: Working conference : Revised papers (pp. 99–122). MIT Press. Boldero, J. M., Rawlings, D., & Haslam, N. (2007). Convergence between GNAT-assessed implicit and explicit personality. European Journal of Personality,21(3), 341–358. https://doi.org/10.1002/per.622 Brownstein, M., Madva, A., & Gawronski, B. (2020). Understanding Implicit Bias: Putting the Criticism into Perspective. Pacific Philosophical Quarterly,101(2), 276–307. https://doi.org/10.1111/papq.123 02 Buhlmann, U., Teachman, B. A., & Kathmann, N. (2011). Evaluating implicit attractiveness beliefs in body dysmorphic disorder using the Go/No-go Association Task. Journal of Behavior Therapy and Experimental Psychiatry,42(2), 192–197. https://doi.org/10.1016 /j.jbtep.2010.10.003 Carlson, K. D., & Herdman, A. O. (2012). Understanding the Impact of Convergent Validity on Research Results. Organizational Research Methods,15(1), 17–32. https://doi.org/10.1177/10944281103 92383 Chen, D. L., Schonger, M., & Wickens, C. (2016). oTree—An open-source platform for laboratory, online, and field experiments. Journal of Behavioral and Experimental Finance,9, 88–97. https://doi.org /10.1016/j.jbef.2015.12.001 Cohen, J. (2013). Statistical Power Analysis for the Behavioral Sciences. Routledge. https://doi.org/10.4324/9780203771587 Cornelissen, G., Bashshur, M. R., Rode, J., & Le Menestrel, M. (2013). Rules or consequences? The role of ethical mind-sets in moral dynamics. Psychological Science,24(4), 482–488. https://doi.org/10.11 77/0956797612457376 Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349–354. https://doi.org/10.1037/h0047358 Cunningham, W. A., Preacher, K. J., & Banaji, M. R. (2001). Implicit attitude measures: Consistency, stability, and convergent validity. Psychological Science,12(2), 163–170. https://doi.org/10.1111/14679280.00328 Dabbs, J. M., Bassett, J. F., & Dyomina, N. V. (2003). The palm IAT: A portable version of the implicit association task. Behavior Research Methods, Instruments, & Computers : A Journal of the Psychonomic Society, Inc,35(1), 90–95. https://doi.org/10.3758/BF03195500 de Houwer, J. (2006). What are Implicit Measures and Why are We Using Them? In R. Wiers & A. Stacy (Eds.), Handbook of Implicit Cognition and Addiction (pp. 11–28). SAGE Publications, Inc. https: //doi.org/10.4135/9781412976237.n2 de Houwer, J., Teige-Mocigemba, S., Spruyt, A., & Moors, A. (2009). Implicit measures: A normative analysis and review. Psychological Bulletin,135(3), 347–368. https://doi.org/10.1037/a0014211 de Houwer, J. (2001). A Structural and Process Analysis of the Implicit Association Test. Journal of Experimental Social Psychology,37(6), 443–451. https://doi.org/10.1006/jesp.2000.1464 Doherty, K., & Schlenker, B. R. (1991). Self-Consciousness and Strategic Self-Presentation. Journal of Personality,59(1), 1–18. https://do i.org/10.1111/j.1467-6494.1991.tb00765.x Effron, D. A., & Monin, B. (2010). Letting people off the hook: When do good deeds excuse transgressions? Personality & Social Psychology Bulletin,36(12), 1618–1634. https://doi.org/10.1177/0146167 210385922 L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984 983 Erat, S., & Gneezy, U. (2012). White Lies. Management Science,58(4), 723– 733. https://doi.org/10.1287/mnsc.1110.1449 Erikson, E. H. (1964). Insight and responsibility: Lectures on the ethical implications of psychoanalytic insight ([Norton pbk. ed.]). W.W. Norton. Fahrmeir, L., Heumann, C., Künstler, R., Pigeot, I., & Tutz, G. (2016). Statistik. Springer Berlin Heidelberg. https://doi.org/10.1007/978-3662-50372-0 Fazio, R. H., & Olson, M. A. (2003). Implicit measures in social cognition. Research: Their meaning and use. Annual Review of Psychology, 54, 297–327. https://doi.org/10.1146/annurev.psych.54.10160 1.145225 Ferguson, R. (2018). How flexible is morality? A test of the moral credits model of moral balancing [Doctoral dissertation, Australian Catholic University]. Festinger, L. (1957). A theory of cognitive dissonance ([Renewed 1985 by author]). Stanford University Press. Fischer, J., Dyball, R., Fazey, I., Gross, C., Dovers, S., Ehrlich, P. R., Brulle, R. J., Christensen, C., & Borden, R. J. (2012). Human behavior and sustainability. Frontiers in Ecology and the Environment, 10(3), 153–160. https://doi.org/10.1890/110079 Freedman, J. L., & Fraser, S. C. (1966). Compliance without pressure: The foot-in-the-door technique. Journal of Personality and Social Psychology,4(2), 195–202. https://doi.org/10.1037/h0023552 Friese, M., Bluemke, M., & Wänke, M. (2007). Predicting voting behavior with implicit attitude measures: The 2002 German parliamentary election. Experimental Psychology,54(4), 247–255. https://doi.o rg/10.1027/1618-3169.54.4.247 Green, D. M., & Swets, J. A. (1966). Signal detection theory and psychophysics. Wiley New York. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology,74(6), 1464– 1480. https://doi.org/10.1037//0022-3514.74.6.1464 Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality and Social Psychology,85(2), 197– 216. https://doi.org/10.1037/0022-3514.85.2.197 Jacobsen, C., Fosgaard, T. R., & Pascual-Ezama, D. (2018). WHY DO WE LIE? A PRACTICAL GUIDE TO THE DISHONESTY LITERATURE. Journal of Economic Surveys,32(2), 357–387. https://doi.org/1 0.1111/joes.12204 Johnston, M. E., Sherman, A., & Grusec, J. E. (2013). Predicting moral outrage and religiosity with an implicit measure of moral identity. Journal of Research in Personality,47(3), 209–217. https://doi.o rg/10.1016/j.jrp.2013.01.006 Jordan, J., Leliveld, M. C., & Tenbrunsel, A. E. (2015). The Moral SelfImage Scale: Measuring and Understanding the Malleability of the Moral Self. Frontiers in Psychology,6, 1878. https://doi.org /10.3389/fpsyg.2015.01878 Jordan, J., Mullen, E., & Murnighan, J. K. (2011). Striving for the moral self: The effects of recalling past moral actions on future moral behavior. Personality & Social Psychology Bulletin,37(5), 701–713. http s://doi.org/10.1177/0146167211400208 Kadlec, H. (1999). Statistical properties of d′and βestimates of signal detection theory. Psychological Methods,4(1), 22–43. https://doi.o rg/10.1037/1082-989X.4.1.22 Kim, D. Y. (2003). Voluntary Controllability of the Implicit Association Test (IAT). Social Psychology Quarterly,66(1), 83. https://doi.org/10 .2307/3090143 Lee, Y. H., & Hsieh, G. (2013). Does slacktivism hurt activism? Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 811–820. https://doi.org/10.1145/2470654.2470770 Macmillan, N. A. (2002). Signal Detection Theory. In H. Pashler (Ed.), Stevens’ Handbook of Experimental Psychology: Volume 4: Methodology in Experimental Psychology. THIRD EDITION (pp. 43–90). John Wiley & Sons, Inc. https://doi.org/10.1002/0471214426 .pas0402 Mazar, N., Amir, O., & Ariely, D. (2008). The Dishonesty of Honest People: A Theory of Self-Concept Maintenance. Journal of Marketing Research,45(6), 633–644. https://doi.org/10.1509/jmkr.45.6.633 Mazar, N., & Zhong, C. B. (2010). Do green products make us better people? Psychological Science,21(4), 494–498. https://doi.org/10.1177 /0956797610363538 Melé, D., & Cantón, C. G. (2014). The Homo Economicus Model. In Human Foundations of Management (pp. 9–29). Palgrave Macmillan UK. https://doi.org/10.1057/9781137462619_2 Merritt, A. C., Effron, D. A., & Monin, B. (2010). Moral Self-Licensing: When Being Good Frees Us to Be Bad. Social and Personality Psychology Compass,4(5), 344–357. https://doi.org/10.1111/j.1751-9004 .2010.00263.x Miller, D. T., & Effron, D. A. (2010). Chapter Three - Psychological License: When it is Needed and How it Functions. In Advances in Experimental Social Psychology (pp. 115–155, Vol. 43). Academic Press. https://doi.org/10.1016/S0065-2601(10)43003-8 Monin, B., & Jordan, A. H. (2009). The Dynamic Moral Self: A Social Psychological Perspective. In Personality, Identity, and Character (pp. 341–354). Cambridge University Press. https://doi.org/10 .1017/CBO9780511627125.016 Monin, B., & Miller, D. T. (2001). Moral credentials and the expression of prejudice. Journal of Personality and Social Psychology,81(1), 33– 43. https://doi.org/10.1037/0022-3514.81.1.33 Mullen, E., & Monin, B. (2016). Consistency Versus Licensing Effects of Past Moral Behavior. Annual Review of Psychology,67, 363–385. https ://doi.org/10.1146/annurev-psych-010213-115120 Nisan, M. (1990). Moral Balance: A Model of How People Arrive at Moral Decisions. In T. E. Wren (Ed.), The Moral Domain: Essays in the Ongoing Discussion between Philosophy and the Social Sciences (pp. 283– 314). The MIT Press. Nosek, B. A., & Banaji, M. R. (2001). The Go/No-Go Association Task. Social Cognition,19(6), 625–666. https://doi.org/10.1521/soco.19.6 .625.20886 Paulhus, D. L. (1984). Two-component models of socially desirable responding. Journal of Personality and Social Psychology,46(3), 598–609. https://doi.org/10.1037/0022-3514.46.3.598 Perkins, B. G., Podsakoff, N. P., & Welsh, D. T. (2024). Variance in Virtue: An Integrative Review of Intraindividual (Un)Ethical Behavior Research. Academy of Management Annals,18(1), 210–250. https: //doi.org/10.5465/annals.2022.0057 Perugini, M. (2005). Predictive models of implicit and explicit attitudes. The British Journal of Social Psychology,44(Pt 1), 29–45. https://doi .org/10.1348/014466604X23491 Perugini, M., & Leone, L. (2009). Implicit self-concept and moral action. Journal of Research in Personality,43(5), 747–754. https://doi.o rg/10.1016/j.jrp.2009.03.015 Ploner, M., & Regner, T. (2013). Self-image and moral balancing: An experimental analysis. Journal of Economic Behavior & Organization, 93, 374–383. https://doi.org/10.1016/j.jebo.2013.03.030 Sachdeva, S., Iliev, R., & Medin, D. L. (2009). Sinning saints and saintly sinners: The paradox of moral self-regulation. Psychological Science, 20(4), 523–528. https://doi.org/10.1111/j.1467-9280.2009.02 326.x Sachdeva, S., Jordan, J., & Mazar, N. (2015). Green consumerism: moral motivations to a sustainable future. Current Opinion in Psychology,6, 60–65. https://doi.org/10.1016/j.copsyc.2015.03.029 Schimmack, U. (2021). The Implicit Association Test: A Method in Search of a Construct. Perspectives on Psychological Science : A Journal of the Association for Psychological Science,16(2), 396–414. https://do i.org/10.1177/1745691619863798 Schlegelmilch, B. B., & Simbrunner, P. (2019). Moral licensing and moral cleansing applied to company-NGO collaborations in an online context. Journal of Business Research,95, 544–552. https://doi.o rg/10.1016/j.jbusres.2018.07.040 Schnabel, K., Asendorpf, J., & Greenwald, A. (2007). Using Implicit Association Tests for the Assessment of Implicit Personality Self-Concept. In G. J. Boyle, G. Matthews, & H. Saklofske (Eds.), Handbook of Personality Theory and Testing. SAGE Publications, Inc. https://d oi.org/10.4135/9781849200479.n24 Shalvi, S., Gino, F., Barkan, R., & Ayal, S. (2015). Self-Serving Justifications. Current Directions in Psychological Science,24(2), 125–130. https ://doi.org/10.1177/0963721414553264 L. F. Bläßer /Junior Management Science 10(4) (2025) 966-984984 Teachman, B. A. (2007). Evaluating implicit spider fear associations using the Go/No-go Association Task. Journal of Behavior Therapy and Experimental Psychiatry,38(2), 156–167. https://doi.org/10.10 16/j.jbtep.2006.10.006 Williams, B. J., & Kaufmann, L. M. (2012). Reliability of the Go/No Go Association Task. Journal of Experimental Social Psychology,48(4), 879–891. https://doi.org/10.1016/j.jesp.2012.03.001 Wilson, T. D., Lindsey, S., & Schooler, T. Y. (2000). A model of dual attitudes. Psychological Review,107(1), 101–126. https://doi.org/10.1037 /0033-295X.107.1.101 Zhong, C. B., Ku, G., Lount, R. B., & Murnighan, J. K. (2010). Compensatory Ethics. Journal of Business Ethics,92(3), 323–339. https://doi.or g/10.1007/s10551-009-0161-6 Zhong, C. B., & Liljenquist, K. (2006). Washing away your sins: Threatened morality and physical cleansing. Science (New York, N.Y.), 313(5792), 1451–1452. https://doi.org/10.1126/science.1130 726