Domain-Specific Greed

Trait Greed

The construct of greed can be defined as the excessive, insatiable desire and striving for more even at the expense of others (Krekels & Pandelaere, 2015; Lambie & Stickl Haugen, 2019, Merriam-Webster, 2013; Mussel & Hewig, 2016; Seuntjens, Zeelenberg, Van de Ven, et al., 2015). Thus, at the core of the construct is the excessive striving for more, thus indicating that not any striving for more would qualify as greed, but only striving that goes beyond what one needs (Balot, 2001; L. Wang et al., 2011)—even though it is difficult to define how much is needed. In addition, greed is characterized by insatiability, indicating that the desire for more is never satisfied, no matter how much has been obtained (Krekels & Pandelaere, 2015; Seuntjens, Zeelenberg, Van de Ven, et al., 2015). Finally, a person’s greed may come at the expense of others. Although greedy individuals do not necessarily intend to harm others, they accept that their striving for more may have negative consequences for others (Lambie et al., 2022; Mussel & Hewig, 2016; Mussel & Hewig, 2019).

Regarding its nomological net, trait greed has been found to be related to the constructs of acquisitiveness, low agreeableness, meanness, risk-taking propensity, approach motivation, entitlement, materialism, envy, and miserliness (Krekels & Pandelaere, 2015; Lambie & Stickl Haugen, 2019; Mussel & Hewig, 2016; Seuntjens, Zeelenberg, Van de Ven, et al., 2015). While sharing variance with these constructs, greed is distinct with regard to its definitional elements (i.e., excessiveness, insatiability, and acceptance of costs to others).

The construct of greed has received considerable research attention since the 2008–2009 financial crisis, which was allegedly initiated by the greedy behavior of investment bankers and other protagonists in the financial sector (but see Hoyer et al., 2021). Since then, such notions have sparked several lines of research with the goal of reaching more profound knowledge regarding its definition, antecedents, and behavioral as well as neural correlates (Lambie & Stickl Haugen, 2019; Mussel et al., 2015; Mussel & Hewig, 2019; Seuntjens, Zeelenberg, Van de Ven, et al., 2015; Q. Wang et al., 2021; L. Wang & Murnighan, 2011).

Regarding measurement, several self-report instruments have been developed to assess greed: The Greed Trait Measure (Mussel et al., 2015), the subscale Greed from the Vices and Virtues Scales (Veselka et al., 2014), two instruments both labeled Dispositional Greed Scale (Krekels & Pandelaere, 2015; Seuntjens, Zeelenberg, Van de Ven, et al., 2015), the Multidimensional Dispositional Greed Assessment (Lambie et al., 2022), and the GR€€D scale (Mussel & Hewig, 2016). These measures have been shown to exhibit good psychometric properties, including the prediction of theoretically related criteria, and to converge on a single latent factor (Mussel et al., 2018). Interestingly, these scales diverge, nonetheless, regarding assumed and measured domain-specificity.

Some scholars developed the view that greed can be related to any object. Such a generic view of greed is, for instance, reflected in the 6-item Dispositional Greed Scale by Krekels and Pandelaere (2015). In line with their conceptualization, their measure contains only generic items, that is, items that are formulated on an abstract and general level (e.g., “One can never have enough,” p. 226). This is in line with Seuntjens, Zeelenberg, Van de Ven, et al. (2015) and Lambie et al. (2022) who also considered greed to be generic (although their measures of greed, the 7-item Dispositional Greed Scale and the 20-item Multidimensional Dispositional Greed Assessment include only items relating to money and not to other domains). Other authors have explicitly adopted the view that greed mainly reflects a materialistic desire. For instance, Mussel and Hewig (2016) noted that greed is “mainly seen as a materialistic type of desire and associated with the acquisition of money, as a value by itself or for its instrumental value for obtaining material goods” (p. 53). Consequently, their 12-item GR€€D scale, like other greed scales (e.g., the 10-item Greed subscale from the Vices and Virtues Scales; Veselka et al., 2014), contains items such as “I will always try to increase my income and my assets.” A narrow view on the materialistic aspect of greed has also been adopted in the organizational and management literature (Bruhn & Lowrey, 2012; Haynes et al., 2015). Thus, there is a lack of consensus concerning the conceptualization of greed (generic vs. specific for money and material things). Moreover, current operationalizations of greed do not enable researchers to explicitly examine the nature of greed, as the only domains that are explicitly included are money and material things.

It is important to note that even generic items may not necessarily assess the construct in general. Rather, several scholars have argued that generic items leave room for interpretation regarding the frames of reference to consider in the judgment process (e.g., Lievens et al., 2008; Schulze et al., 2021). Lay theories about a trait (e.g., “a greedy person is someone who wants as much money as possible”) influence how generic items (e.g., “I am kind of greedy”) are interpreted and answered. As a result, the score obtained from generic items may differentially reflect certain domains over others. Whether this actually occurs is an empirical question which, thus far, could not be investigated due to a lack of models and measures describing and assessing domains of greed beyond money.

In the present research, and in line with the prototype analysis by Seuntjens, Zeelenberg, Breugelmans, et al. (2015), we started with a broad view, assuming that greed may pertain to a wide array of objects. We investigate whether objects to which greed may pertain can be aggregated into a wieldy number of broad domains. Thus, we propose that the construct of trait greed can be conceptualized as domain-specific. This assumes that individuals have distinct levels of trait greed across a certain number of domains, which would result in a multidimensional latent factor structure according to greed domains. Given that such a latent structure is supported, it would enable us to investigate the degree to which generic greed reflects individual differences in domain-specific greed and, as such, give an empirical answer to the debate regarding the definition of trait greed. Such knowledge contributes to our understanding of this construct in terms of construct validity and advances definitions of trait greed. To advance toward this goal, we developed a working model of domains to which greed may pertain.

Domains of Greed

Domain-specific traits are traits that primarily relate to a certain class of entities, tasks, subjects, or situations (see American Psychological Association, 2022). They are usually assessed by explicating the class in addition to the item (e.g., “at work,” “with friends”). A domain-specific approach has been shown to be fruitful across many areas and applications. For example, domain-specific self-esteem (e.g., relating to academic, social, parent and peer relationships, or physical attractiveness) was found to predict depression (Steiger et al., 2014), aggression among adolescents (Descartes et al., 2019), and childhood overweight (Danielsen et al., 2012). Domains relating to achievement motivation (usually educational subjects such as math, physics, science, English) showed incremental validity over generic indicators for predicting academic success (Michel et al., 2022). Contextualized personality measures with a reference to work-specific behaviors have been found to be better predictors of job performance (Shaffer & Postlethwaite, 2012). Domains relating to identity processes (best friend, education, occupation, and partner) have been shown to have unique links to identity processes in emerging adulthood (Vosylis et al., 2018). Similar approaches can be found for constructs such as optimism-pessimism (Chang et al., 2011), locus of control (Tong & Wang, 2012), knowledge (Muis et al., 2006), ability self-concept (Steinmayr & Spinath, 2007), consideration of future consequences (Murphy et al., 2020), need for cognition (Keller et al., 2019), risk-taking (Blais & Weber, 2006), creativity (Kaufman, 2012), or narcissism (Grosz et al., 2021).

From this discussion it also becomes clear that domains vary strongly from one construct to the other. To date, a systematic investigation of potential domains of greedy behavior and their measurement does not exist (Mussel et al., 2018). However, according to our review of the literature and in line with the first empirical step (expert interviews; see “Scale Development”), there are a number of different domains to which greed may potentially pertain.

As mentioned above, greed has been frequently associated with the concept of money. For example, the Merriam-Webster (2013) dictionary defines greed as “a selfish and excessive desire for more of something (such as money) than is needed.” Seuntjens, Zeelenberg, Breugelmans, et al. (2015) conducted a prototype analysis to investigate lay conceptions of greed. Among the most often mentioned domains were money and materialism, which were also rated as being among the central aspects of greed. Mussel and Hewig (2016) found that individuals high on greed showed more positive affect after winning money and more negative affect after losing money than individuals low on greed. Thus, on one hand, greed can be associated with the acquisition and valuing of money as a value by itself, labeled as “love for money” by Tang (1992). On the other hand, money has instrumental value for obtaining material goods (Bruhn & Lowrey, 2012), which is related to the literature on materialism, defined as the propensity to enjoy buying things above what would be necessary and to appreciate one’s property and belongings (Richins & Dawson, 1992).

However, greed may also pertain to the domains of acknowledgment, power, performance, or knowledge. Greed in acknowledgment describes an excessive striving for recognition and acceptance and may be related to the concept of excessive reassurance seeking (Joiner & Metalsky, 2001; Schwennen & Bierhoff, 2014; Shaver et al., 2005). Greed in the domain of power can be described as an intense striving for influence and is related to the concept of excessive power-seeking (Charny, 1997). Excessive power-seeking has also been related to ruthlessness and a lack of empathy, which corresponds to the striving for more even at the expense for others (Charny, 1997; Mussel & Hewig, 2016). The link between power and greed has also been supported by Machiavellianism research, which has indicated that there is a connection between power and money in the context of self-serving behavior (Christie & Geis, 1970; Rauthmann, 2012). Greed in the domain of performance is related to the construct of achievement motivation (McClelland, 1985) and describes the excessive and insatiable desire for self-achieved success. This excessive striving might be related to concepts such as workaholism (Spence & Robbins, 1992) or perfectionism (Hewitt & Flett, 1991). In sports, excessive striving for success was found to be related to addiction and exercise dependence (McNamara & McCabe, 2012). Zhu et al. (2019) showed that, in workers, higher greed was associated with higher performance levels. These performance levels referred to self-related task performance as well as to other-related contextual performance and were mediated by the striving for a higher social status. This finding is also consistent with the result that greedier people are more productivity-oriented (in addition to having a desire for profits; Krekels & Pandelaere, 2015). Greed in the domain of knowledge is related to an excessive striving for information, comparable to an extreme form of curiosity (Litman, 2019; Mussel, 2010) and is possibly related to constant checking (e.g., of email clients, social networking systems, news websites; Gerlach & Cenfetelli, 2020). Thus, a striving for status, power, performance, or knowledge may take place in the form of excessive, insatiable, and selfish striving (Kim, 1976).

In addition, greedy behavior may also pertain to sex, food, friendship, or substances. One of the first references to sex in relation to greed was noted by Countryman (1988). Against the backdrop of Christian ethics, he related adultery to theft and incest to violating the hierarchy of the family. More generally, greed in the domain of sex can be related to excessive sexual desire and hypersexuality (Kafka, 2010; Kuhn et al., 2014). An excessive striving for food describes eating behavior beyond what is needed and thus pertains to the hedonic rather than metabolic component of appetite (Berthoud, 2011). It can be related to food craving (White et al., 2002), eating habits, eating disorders, and obesity (Drewnowski, 1997) and an excessive consumption of palatable food (Kenny, 2011). Striving for a large number of friends is related to the affiliation motive (Depue & Collins, 1999; McClelland, 1985). It is associated with preferences regarding the size of one’s personal network, both online and offline, and can be related to an excessive form of the need for popularity (Utz et al., 2012). Seuntjens, Zeelenberg, Van de Ven, et al. (2015) explicitly stated that striving for sex, food, and social status pertains to the construct of greed. In line with their reasoning, these authors found positive correlations between trait greed and self-reports of greedy eating behavior, desiring as many casual sex partners as possible, and striving for as many friends as possible on social network sites. Finally, greed in the domain of substances pertains to the excessive desire for intoxication. It may be related to craving (Robinson & Berridge, 1993), substance abuse (Hawkins et al., 1992), and addiction proneness (Flagel et al., 2009). Thus, there is evidence that the domains of sex, food, friendship, and substances may be relevant regarding greedy striving.

As this brief literature review suggests, each of the domains of greed can be expected to show a unique nomological net that overlaps with existing, theoretically related constructs. For example, greedy striving for friendship can be expected to be related to the affiliation motive. At the same time, domains of greed should be distinct from existing constructs regarding the definitional elements of greed (i.e., excessiveness, insatiability, and striving for more even at the expense of others). Thus, domains of greed should be related to generic greed after accounting for the variance that is shared with related constructs. Such evidence is necessary to show that the proposed domains of greed reflect unique constructs that go beyond existing ones or, in other words, to preclude what Block (1995) labeled a jangle fallacy.

To sum up, greedy behavior may be found in different domains, and the extent to which such behavioral propensities generalize from one domain to another is unknown.

Research Questions

In the current research, we investigated the construct of trait greed with regard to domains. In a first step, we investigated whether greed can be conceptualized as a domain-specific construct. Based on domains derived from the literature and from an expert study, (a) we propose a working model of domains to which greed may pertain, (b) operationalize this model by a newly developed scale, (c) test the fit of the model regarding the proposed multidimensional structure, and (d) investigate the construct- and (e) criterion-related validity of the model. We summarize this attempt by the following broad research question:

Our second research question is related to the nature of generic greed. Generic greed is conceptualized as the excessive, insatiable desire and striving for more even at the expense of others in general. Thus, generic greed does not specify the desired object that is strived for, as illustrated in the item “I always want more” (Seuntjens, Zeelenberg, Van de Ven, et al., 2015, p. 921). Based on the results regarding the domain-specific structure of greed according to RQ1, we investigate whether generic greed is equally or differentially related to greed domains. Such knowledge informs about the nature of trait greed including its definition and how to properly interpret scores derived from generic greed scales. We summarize these queries by the following research question:

The Present Research

With the present research, we first investigate whether greed can be conceptualized as a domain-specific construct. Using a literature review and a qualitative expert survey, we propose several domains to which greed may pertain. Next, we developed a measure to assess these greed domains. In a prestudy, we refined the measure according to its psychometric properties and report results on reliability and factor structure. In Study 1, we replicated these results and additionally investigated our measure’s construct-related validity. We related each greed domain, on one hand, to generic greed, and, on the other, to a corresponding measure that reflects preferences for the respective domain. Finally, we investigated potential differences in the extent to which each greed domain is saturated by generic greed to arrive at a deeper understanding of the nature of trait greed regarding its domain-specificity. In Study 2, we addressed the criterion-related validity of greed domains according to self- and peer-ratings of the corresponding criteria from each greed domain. We also investigated the incremental contribution of a domain-specific approach beyond a generic approach. Across all studies, we addressed the issue of whether generic greed is differentially related to different domains by comparing the relation between a generic subscale with the domain-specific subscales of the newly developed greed measure.

Scale Development

We conducted a literature review to identify the domains of greed as summarized above. In addition, we used a bottom-up approach to generate the domains empirically. In this preliminary study, we asked 25 experts to name as many greed domains as possible. Experts were students attending a university course who had received at least 8 h of training on the topic of greed. This procedure resulted in 133 statements which are depicted as a word cloud in Figure 1. We had three independent research assistants’ cluster statements according to similarity. The cluster solutions were compared, and discrepancies were resolved by discussion. As a result, 10 separate domains evolved that adequately captured all 133 statements. The 10 domains were labeled as follows: acknowledgment, food, friendship, knowledge, material things, money, performance, power, sex, and substances. Next, the literature, as cited in the introduction, was scanned for additional domains that were suggested in the context of greedy behavior. We found that all concepts referred to in the literature could be categorized within these 10 domains. Thus, these domains reflect our working model for domains of greed.

OPEN IN VIEWER

Figure 1.

Word Cloud of Domains Associated With the Term “Greed,” as Derived From 133 Terms Provided by 25 Experts (Created With the R Package “Wordcloud”).

We developed our domain-specific greed questionnaire (DOSPEG: DOmain-SPEcific Greed) according to these domains. In sum, we developed 57 German items for the 10 domains with between five and nine items per domain. In addition, we used and rephrased six generic greed items from existing scales (Mussel & Hewig, 2016; Seuntjens, Zeelenberg, Van de Ven, et al., 2015) that lacked any reference to a domain (e.g., “I always want more”). All items can be found on Open Science Framework (OSF; https://osf.io/t98mc).

A prestudy (N = 171; for details see OSF) was conducted to pretest the items. Based on results from item analysis, four items per scale were selected (see Supplemental Material online for details). Items are depicted in Table 1. According to our theorizing, we expected the internal structure of the newly developed DOSPEG questionnaire to be characterized by a general generic greed factor and 10 factors.

OPEN IN VIEWER

Table 1.

Items for the DOSPEG Scale (English Translations).

Generic

1. I have a proverbial hunger for more.

2. I am indeed quite greedy.

3. I always want more.

4. As soon as I have achieved something, I think about what I want next.

Acknowledgment

1. I need recognition like the air I breathe.

2. I strive to get recognition from all sides.

3. I can’t get enough recognition from others.

4. I’m always trying to get more recognition.

Food

1. When food is distributed, I can’t get enough for myself.

2. When I think of a certain food, I absolutely want it immediately.

3. I often put more on my plate than I can eat.

4. Sometimes I eat until I feel sick.

Friendship

1. I always try to expand my circle of friends.

2. I strive to make more and more friends.

3. You can never have enough friends.

4. I am always looking for new friends.

Knowledge

1. My thirst for knowledge cannot be satisfied.

2. When I have acquired new knowledge, I would like to continue learning immediately.

3. I have an insatiable hunger for more knowledge.

4. I always want to expand my knowledge.

Material Things

1. When I have bought something beautiful, I immediately think about what I want to have next.

2. Sometimes I feel the irresistible urge to possess certain things.

3. I can’t own enough things.

4. I always want to buy new things.

Money

1. I would like to have a lot of money, even if I don’t know exactly what for yet.

2. I am driven internally to get more and more money.

3. I strive to earn as much money as possible.

4. Earning more money will always be important to me.

Performance

1. I always try to improve my performance.

2. I will always strive to achieve new successes.

3. As soon as I have achieved something, I would like to continue immediately with new tasks.

4. I push myself from one achievement to the next.

Power

1. I am hungry for power.

2. You can’t be powerful enough.

3. Getting more power is my most important goal.

4. I’m striving to have as much influence as possible.

Sex

1. I can’t have sex often enough.

2. When I had sex, I want to have sex right away.

3. I have a very strong sexual desire.

4. I would like to have sex all the time.

Substances

1. I have a great desire for intoxication.

2. Every now and then I consume more intoxicants than I can handle.

3. I never get enough of certain substances.

4. I am greedy for the “kick” that certain substances give me.

Source. Reproduced with permission of the authors; full item content available on OSF (https://osf.io/t98mc).

Note. DOSPEG = DOmain-SPEcific Greed.

Study 1

Study 1 aimed at estimating the psychometric properties of the DOSPEG scale, including a cross-validation of the results obtained in the prestudy (see Supplemental Material online). This is important because in the prestudy, the scale optimization and the investigation of the scales’ psychometric properties were based on the same data. In addition, we investigated the relations between generic greed and the greed domains, thereby also investigating the difference in correlation coefficients to further contribute to our understanding of the nature of trait greed. Finally, we provided construct-related evidence for the 10 domains by relating each of the domains to a corresponding measure that did not measure greed but referred to the respective domain.

Results

We investigated the reliability and adjusted item-scale correlations within each of the 10 domains of the DOSPEG questionnaire. The scales showed good psychometric properties (ω between .78 and .94, see Table 2).

OPEN IN VIEWER

Table 2.

Results for the Bifactor-(S-1) Models in Study 1 (S1), and Study 2 (S2).

	True scores
	Reliability		Consistency		Specificity
	S1	S2	S1	S2	S1	S2
Generic (reference)	.83	.86	1	1	0	0
Acknowledgment	.90	.91	.34	.34	.66	.66
Food	.78	.79	.29	.39	.71	.61
Friendship	.88	.90	.13	.15	.87	.85
Knowledge	.89	.90	.12	.07	.88	.93
Material things	.84	.87	.50	.66	.50	.34
Money	.90	.91	.46	.58	.54	.42
Performance	.83	.87	.36	.32	.64	.68
Power	.87	.90	.56	.56	.44	.44
Sex	.94	.93	.14	.16	.86	.84
Substances	.87	.97	.17	.34	.83	.66

Note. Generic greed is the reference. The reliabilities (omega; see Dunn et al., 2014) of the scales are displayed as well as the consistency between generic greed and the 10 postulated domains with their respective specificity.

All criteria in the specified bifactor-(S-1) model exhibited at least acceptable model fit ( χ 2 = 1 , 945 , df = 817, p < .001; comparative fit index [CFI] = .93; Tucker–Lewis index [TLI] = .92; root mean square error of approximation [RMSEA] = .05 (90% confidence interval [CI] = [.046, .051]); standardized root mean square residual [SRMR] = .04; factor loadings are depicted in Table S6). The consistency coefficients varied from domain to domain, ranging from .12 to .56. The highest consistency coefficients were found for power, material things, and money (Table 2). Lower consistency coefficients were found for knowledge, friendship, sex, and substances. In other words, generic greed shared more variance with greed toward power, material things, and money compared with greed toward knowledge, friendship, sex, and substances. The partial correlations between domains ranged from r = −.12 to r = .64 (Table 3). This means that the residual factors shared some variance after accounting for the common variance of generic greed but were far from equivalent in terms of discriminant validity.

OPEN IN VIEWER

Table 3.

Partial Correlations for the Specific Factors Obtained From the Bifactor-(S-1) Model in Study 1.

Domain		0	1	2	3	4	5	6	7	8	9	10
Generic	(0)	1
Acknowledgment	(1)	0	1
Food	(2)	0	.26	1
Friendship	(3)	0	.41	.23	1
Knowledge	(4)	0	.09	−.12	.10	1
Material things	(5)	0	.34	.55	.24	−.06	1
Money	(6)	0	.17	.24	.11	−.02	.38	1
Performance	(7)	0	.34	.11	.29	.64	.14	.28	1
Power	(8)	0	.43	.22	.32	.11	.35	.28	.29	1
Sex	(9)	0	.15	.19	.15	.13	.19	.25	.14	.38	1
Substances	(10)	0	.24	.40	.22	.02	.37	.08	.03	.39	.36	1

Model fit for the full ESEM and the confirmatory factor analysis (CFA) are reported in the supplement (see Table S5). The full ESEM showed improved model fit compared with the CFA and the bifactor-(S-1) model (see Table S5). However, there were only three cross-loadings >.30 in the full ESEM, and the median of the cross-loadings was .01. Regarding latent correlations, the median of the absolute difference between the CFA and the ESEM was .07; only one correlation difference was > .30 (power and sex). The median latent correlation between one domain and all other domains (excluding generic greed) in the CFA ranged from .23 (knowledge) to .56 (material things) and in the ESEM from .21 (knowledge) to .41 (material things).

On a manifest level, the DOSPEG domains were correlated with each other, ranging from r = −.01 to r = .63 (see Table S2 in Electronic Supplementary Material 1). The correlations between generic greed and the DOSPEG domains ranged from r = .30 to r = .60 (all ps < .001, see second column in Table 4), indicating that all 10 domains shared substantial variance with generic greed. We used the procedure proposed by Meng et al. (1992) to test whether the domains were differentially related to generic greed. The results indicated that the correlation coefficients were not equal ( χ 2 = 227 , df = 9, p < .001). Therefore, it is concluded that generic greed is differentially reflected by the domains (see also the variability of the consistency coefficients in the bifactor-(S-1) model). In accordance with the bifactor-(S-1) model, material things, money, and power (all rs ≥ .58) showed the highest correlations with generic greed. Lower correlations were found for food, friendship, knowledge, sex, and substances (.32 ≤ r ≤ .40). The biggest difference in the correlation between a domain and generic greed was found between power and friendship: the squared correlation was four times higher for power than for friendship ( R p o w e r 2 . 37 vs. R f r i e n d s h i p 2 = . 09 ).

OPEN IN VIEWER

Table 4.

Correlations Between the DOSPEG Generic Greed Scale and the DOSPEG Domains in Study 1.

DOSPEG	Generic	Convergent validity	Generic (partial)
Acknowledgment	.47	.46	.41
Food	.40	.65	.23
Friendship	.30	.58	.25
Knowledge	.34	.55	.41
Material things	.58	.69	.40
Money	.59	.57	.48
Performance	.52	.64	.39
Power	.60	.64	.46
Sex	.32	.76	.27
Substances	.32	.61	.29

Note. The correlations between the DOSPEG generic greed scale and the DOSPEG domains are displayed in the column labeled generic. The correlations between the DOSPEG greed domain and the corresponding validation questionnaires are presented in the column labeled convergent validity. The partial correlations between generic greed and each domain (controlling for the validation questionnaire) are shown in the column labeled generic (partial). All correlations were significant (all ps ≤ .05). DOSPEG = DOmain-SPEcific Greed.

The convergent validities between the DOSPEG scales and the corresponding validation questionnaires are reported in the third column of Table 4. On average, the correlation between the DOSPEG domains and the corresponding measures was .62. On the level of individual domains, all 10 domains were significantly correlated with their corresponding measure. Conversely, the discriminant validities were .22, on average (see Table S7 in Electronic Supplementary Material 1). Thus, the results provide support for the construct validity of the DOSPEG domains. At the same time, the convergent validities were far from equivalent, indicating that the proposed greed domains are not the same as the corresponding constructs.

To further sound out the uniqueness of the DOSPEG, we investigated whether the domains (e.g., greed for knowledge) were correlated with generic greed after accounting for the variance shared with the corresponding measures (e.g., curiosity). As can be seen from the last column in Table 4, the partial correlations showed that the greed domains were still significantly related to generic greed after accounting for the variance shared with the corresponding measures (all ps ≤ .001). Thus, the DOSPEG domains are domain-specific measures of greed, rather than merely reflecting preferences for the respective domain (e.g., greedy striving for performance, rather than merely reflecting achievement motivation), indicating that the newly developed scale does not provide “old wine in new skins” in terms of the jangle fallacy.

Study 2

In Study 2, we provided a replication of the bifactor-(S-1) model identified in Study 1. Furthermore, we investigated the convergent validity of our generic greed scale by relating it to an established generic measure of greed (Krekels & Pandelaere, 2015). The scale was chosen as it conceptualizes and operationalizes greed as generic, that is, none of the six items refer to a domain such as money or luxury goods. The main purpose in Study 2 was to investigate criterion-related validity for the greed domains. We investigated convergent and divergent validities with corresponding criteria, obtained via self- and other-ratings, thereby providing further evidence of the validity of the postulated 10-dimensional model of domain-specific greed. Building on these results, we investigated the benefit of a domain-specific over a generic approach by testing the incremental validity of the DOSPEG domains with respect to corresponding criteria beyond generic greed.

Statistical Analyses

We reran the bifactor-(S-1) model, as specified in Study 1, to replicate the factor structure of the DOSPEG. We used the same fit indices and cut-offs as before.

To validate the generic greed subscale, we calculated its correlation with the Dispositional Greed Scale (Krekels & Pandelaere, 2015). As in Study 1, we also report correlations between the DOSPEG domains and DOSPEG generic greed.

Regarding the peer-rated criteria, we computed the intraclass coefficients (ICCs) for the participants with more than one peer rating. After averaging multiple peer ratings, we used bivariate correlations to measure the consistency between the self- and peer-ratings on the criterion items.

We calculated bivariate correlations between each of the 10 greed domains and its corresponding criterion to investigate convergent validity. Next, we performed 10 separate multiple regressions; for each of the one-item peer-rated criteria, we first regressed generic greed (Krekels & Pandelaere, 2015) and subsequently investigated the incremental validity of the corresponding domain-specific greed subscale.

Results

All criteria in the specified bifactor-(S-1) model exhibited at least an acceptable model fit ( χ 2 = 1 , 826 , df = 817, p < .001; CFI = .94; TLI = .93; RMSEA = .05 (90% CI = [.046, .053]); SRMR = .04). The consistency coefficients varied from domain to domain, ranging from .07 to .66. The highest consistency coefficients were found for material things, money, and power (see Table 2).

The generic greed subscale from the DOSPEG was strongly correlated with the Dispositional Greed Scale (r = .81, p < .001). Correlations between the DOSPEG subscales (for reliabilities, see Table S1 in Electronic Supplementary Material 1) and the Dispositional Greed Scale were high (on average r = .44, see Table 5).

OPEN IN VIEWER

Table 5.

Results of Study 2 Targeting the Criterion-Related Validity of the DOSPEG.

Domain	Construct-related validity	Interrater consistency		Criterion-related validity		Incremental validity from multiple regression analyses
	DGS	ICC	Self-peer	Self	Peer	Predictor	Model 1β	R²	Model 2β	R²	ΔR²
Ackn	.45	.31	.40	.62	.34	DGS	.44	.083	.27	.135	.053
						Domain	—		.29
Food	.42	.50	.51	.41	.37	DGS	.25	.017	−.05	.137	.122
						Domain	—		.60
Friend	.31	.37	.44	.51	.28	DGS	.31	.040	.20	.091	.052
						Domain	—		.28
Know	.20	.34	.42	.41	.29	DGS	.01	−.002	−.09	.082	.085
						Domain	—		.36
Mat	.66	.48	.50	.55	.35	DGS	.51	.103	.25	.136	.035
						Domain	—		.31
Mon	.67	.36	.46	.56	.34	DGS	.44	.073	.13	.116	.044
						Domain	—		.32
Per	.48	.31	.38	.69	.36	DGS	.39	.060	.15	.136	.078
						Domain	—		.40
Pow	.59	.32	.38	.72	.40	DGS	.48	.086	.15	.163	.078
						Domain	—		.45
Sex	.26	.54	.59	.61	.42	DGS	.38	.049	.21	.189	.141
						Domain	—		.44
Sub	.39	.42	.57	.67	.55	DGS	.47	.083	.14	.311	.229
						Domain	—		.69

Note. Correlations between the DOSPEG subscales and the Dispositional Greed Scale (DGS) from Krekels and Pandelaere (2015; in the column labeled DGS), the self-criterion (self) and the peer-criterion (peer) are displayed, as well as correlations between the self-criterion and peer-criterion (self-peer) and the intraclass coefficient between peers (ICC). Finally, the incremental validity of the DOSPEG subscales (Model 2) in addition to the generic DGS only (Model 1), the R² value, and the increase in R² (ΔR²) are depicted. The DOSPEG subscales are ackn (acknowledgment), foo (food), fri (friendship), kno (knowledge), mat (material things), mon (money), per (performance), pow (power), and sub (substances). All correlations were significant (p ≤ .05). Significant (p ≤ .05) predictors in the model without the specific greed scales (Model 1) and the peer rating as the criterion and with the incremental validity of the specific greed scale (Model 2) are presented in bold. DOSPEG = DOmain-SPEcific Greed; DGS = Dispositional Greed Scale; ICC = intraclass coefficient.

For the peer-rated criteria, the ICC for n ≥ 2 peer ratings was moderate (.31 ≤ ICC ≤ .54), showing that different peers evaluated the same target person somewhat heterogeneously regarding the criterion items. The average correlation between the self-ratings and peer-ratings was r = .46 (.38 ≤ r ≤ .59), indicating moderate agreement on the criteria between the self-perception and peer ratings, see Table 5.

Regarding criterion-related validity, correlations between the DOSPEG subscales and the self-rated criterion were high (.41 ≤ r ≤ .72). The correlations between the DOSPEG and the peer-rated criterion were all significant but moderate (.28 ≤ r ≤ .55), as reported in Table 5.

Finally, we investigated the incremental validity of the DOSPEG domains over DOSPEG generic greed on the corresponding criteria in separate multiple regression analyses. In Model 1, the DOSPEG generic greed subscale was a significant predictor of all peer-rated criteria, apart from knowledge (β = 0.01, p = .94). When comparing the predictive validity of DOSPEG generic greed (column “Model 1 β” in Table 5) with the construct-related validity (column “construct-related v. DGS”), there is a clear correspondence: The stronger a domain is reflected in generic greed, the stronger the validity of generic greed with the corresponding criterion (r = .71 across the 10 coefficients). Note that the criteria were measured using peer-ratings, ruling out the explanation of a common method effect.

Regarding the incremental validity of the DOSPEG subscales, all domain-specific predictors were significant (all ps ≤ .001). Importantly, there was a significant increase in explained variance when we added the domain-specific measure for all 10 criteria (.04 ≤ ΔR² ≤ .23), supporting the additional value of domain-specific greed beyond generic greed.

General Discussion

Personality psychology is moving from the description of broad personality traits (Goldberg, 1990) to the investigation of underlying processes (Dweck, 2017), influences of situational characteristics (Parrigon et al., 2017), personality states (Horstmann & Ziegler, 2020), and differential effects of facets (Judge et al., 2013) and nuances (Mottus et al., 2019), thereby providing a finer-grained understanding of personality. The research described in the current article contributes to this advancement by providing a differentiated approach to the construct of greed. Specifically, we investigated whether the propensity for greedy behavior could be differentiated according to the domain toward which the excessive striving for more is directed.

On the basis of a literature review and an expert survey, we identified different domains to which greedy behavior may pertain. We developed and validated a new personality measure, the DOSPEG scale, which assesses these greed domains and, additionally, generic greed. The scale showed adequate psychometric properties. Furthermore, we found support for the convergent and discriminant validity of the 10 domains in Study 1 and for criterion-related validity in Study 2. Accordingly, we consider the DOSPEG scale a reliable and valid scale for the assessment of domain-specific greed.

A domain-specific approach to the assessment of greed has the potential to provide a fine-grained intraindividual profile of trait greed and to enhance the prediction of related criteria over and above domain-general scales. Based on a domain-specific conceptualization, future research may investigate whether developmental trajectories found for trait greed (Mussel et al., 2022) are different according to domain, maybe reflecting changes in personal values as proposed by the theory of socioemotional selectivity (Carstensen, 1995). In addition, the DOSPEG provides a detailed intraindividual profile of greedy behavior in different domains, which has additional value for feedback, self-reflection, and personal development. Finally, prediction of specific criteria such as trading behavior (Hoyer et al., 2021), leadership (Haynes et al., 2015), hoarding of food (Yoshino et al., 2021), or eating behavior (De Backer et al., 2015) will likely benefit from using a domain-specific predictor that corresponds to the criterion. In line with this reasoning, we showed in Study 2 that for all 10 domains, domain-specific greed had incremental validity in predicting corresponding criteria over generic greed. Given that for such specific predictions only the relevant domain is needed, the DOSPEG, with only four items per scale, is a highly economic measure.

A Structural Model of Domain-Specific Greed

We integrated the findings from our studies into a structural model of domain-specific greed, as depicted in Figure 2. First, this model postulates that in addition to generic greed, 10 domains can be differentiated according to the respective target to which greedy striving may pertain. These domains have been proposed according to a thorough review of the literature, a qualitative study with 25 experts, and structural analyses across samples. Whereas it cannot be ruled out that other greed domains exist, we believe that these 10 domains include the majority of objects that greedy individuals strive for. Exploratory and confirmatory factor analyses across all three studies showed that these 10 domains can be differentiated. In addition, we demonstrated convergent and discriminant validity with corresponding constructs as well as criteria. Moreover, we showed that all 10 domains were significantly correlated with generic greed after accounting for the variance they shared with constructs that reflect a preference for the respective domain (e.g., striving for achievement) but not the excessiveness that is characteristic of greed (e.g., greedy striving for performance). This result was important for showing that all dimensions pertain to the construct domain of trait greed and may be labeled accordingly, rather than reflecting a different construct that erroneously goes under a false name in terms of the jangle fallacy (Block, 1995).

OPEN IN VIEWER

Figure 2.

Onion Model of Domain-Specific Greed.

Second, the model reflects our observation that some domains are more central than others: Generic greed was significantly related to all domains, but the correlations differed significantly from each other. Similarly, the results of the bifactor-(S-1) models have shown that generic greed shares more variance with domains such as money, material things, performance, and power and less variance with domains such as friendship, knowledge, sex, and substances. Differences in centrality may reflect lay conceptions of greed. As such, even though there are many domains to which greed pertains, some might be more important or readily available to lay persons than others. This reasoning is in line with the prototype analyses by Seuntjens, Zeelenberg, Breugelmans, et al. (2015), who found that terms relating to money, material things, and power were often associated with greed, whereas food or sex were only rarely or never associated with greed. Thus, when answering generic greed items (e.g., “I always want more”), individuals seem to relate the statement more strongly to greedy behavior in certain domains over others (Schulze et al., 2021).

We conclude that, as a response to RQ1, greed can be conceptualized as a domain-specific construct. The onion model in Figure 2 shows 10 domains to which greed pertains and additionally illustrates their centrality with regard to generic greed.

Supplemental Material