Despite an increasing need for creativity in all corners of business, the spotlight of most recruitment and selection procedures has not shifted accordingly. This article examines the performance of different assessment methods of creative ability in two small-scale hiring contexts. With prospective marketing employees, a combination of test ratings and student CV ratings of creative ability shows high operational validity.
Despite an increasing need for creativity in all corners of business, the spotlight of most recruitment and selection procedures has not shifted accordingly. This article examines the performance of different assessment methods of creative ability in two small-scale hiring contexts. With prospective marketing employees, a combination of test ratings and student CV ratings of creative ability shows high operational validity.
Despite an increasing need for creativity in all corners of business, the spotlight of most recruitment and selection procedures has not shifted accordingly. This article examines the performance of different assessment methods of creative ability in two small-scale hiring contexts. With prospective marketing employees, a combination of test ratings and student CV ratings of creative ability shows high operational validity.
Niek Althuizen Published online: 21 July 2012 # Springer Science+Business Media, LLC 2012 Abstract Despite an increasing need for creativity in all corners of business, the spotlight of most recruitment and selection procedures has not shifted accordingly. Measures of creative ability that are to be used in practice should preferably be brief and operationally valid in the small and relatively homogenous pools of subjects that companies typically have to deal with. This article examines the performance of different assessment methods of creative ability in two small-scale hiring contexts, i.e., with prospective marketing employees (marketing students) and with current employees of a creative marketing agency. With prospective marketing employees, a combination of test ratings and student CV ratings of creative ability shows high operational validity. Supervisory ratings and self-ratings of creative ability are rea- sonable alternatives to test ratings for senior and long-time employees, but should be used with caution when junior employees and recent recruits are concerned. Keywords Creative ability . Measurement . Recruitment . Selection . Marketing employees 1 Introduction Creativity enables organizations to develop innovative strategies, new products, and novel ways of working that are crucial for survival in highly competitive and dynamic business environments (Andrews and Smith 1996; Woodman et al. 1993). A study by IBM (2010) revealed that 60 % of the surveyed chief executive officers worldwide consider creativity as their top priority. A survey by Patterson et al. (2009) among UK businesses showed that 78 % views creativity and innovation as key to survival and future success. One would expect that organizations have shifted the spotlight of their recruitment and selection procedures accordingly. However, of the aforementioned Mark Lett (2012) 23:973985 DOI 10.1007/s11002-012-9198-x N. Althuizen (*) ESSEC Business School, Avenue Bernard Hirsch, 95201 Cergy-Pontoise, France e-mail: althuizen@essec.edu UK businesses, only 29 % listed creativity among their personnel selection criteria and most appraisal processes still favor conscientiousness over innovativeness (Pat- terson et al. 2009). Measures of creative ability do not enjoy the same reputation as general mental ability and personality tests or other methods to gauge an applicants aptitude for the job, such as the evaluation of CVs, interviews, and assessment centers (see Robertson and Smith 2001). The marketing literature has also paid scant attention to the measurement of creative ability and individual differences in creative ability are largely neglected as an explanatory variable in marketing research (e.g., Goldenberg et al. 1999; Sellier and Dahl 2011). According to a review of applied creativity research in business by Kabanoff and Rossiter (1994), however, the largest factor contributing to the creative output of the firm is the creative ability of the individual. The present article is the first to explore the operational validity of different assessment methods of creative ability in the typically small and relatively homogenous pools of subjects that companies have to deal with (Scullen and Meyer 2012). An additional requirement for use in marketing research and practice is that measures of creative ability should be preferably brief and easy to administer (Scratchley and Hakstian 2001). Most studies that investigate the validity of selection instruments are based on the (implicit) assumption that companies have an infinite pool of subjects to choose from. From this theoretical perspective, it makes sense to correct the observed validity coef- ficients for range restriction and unreliability in the measures to get at the true population validity coefficient (Schmitt 2007). However, companies have to select new recruits or incumbents from a finite pool of subjects, with sample sizes of less than 20 being the rule rather than the exception (Scullen and Meyer 2012). Moreover, the subjects are likely to be similar on a number of background variables (e.g., job qualifications). Because of differ- ences between theoretical and real-world hiring situations (Scullen and Meyer 2012), scholars have urged to examine validity coefficients in small-scale hiring contexts (Heneman et al. 2000). Such studies usually fail to attract attention from researchers because of limited generalizability and reduced statistical power (Scherbaum 2005). This article presents the results of two validation studies in small-scale hiring contexts. Study 1 (n=20) examines the performance of three brief assessment methods of creative ability, viz. test ratings, CV ratings and self-ratings, in a laboratory setting with prospective marketing employees (i.e., marketing students). Study 2 (n=24) further investigates the convergent validity of the self-ratings and test ratings together with supervisory ratings of creative ability in a field setting with current employees of a marketing agency. Small n studies are often criticized for being merely descriptive and a-theoretical. However, small n studies do rely, albeit often implicitly, on interpretive theory to provide meaning to the facts that these studies generate (Coppedge 2007). The studies presented in this article are grounded in the creative cognition literature (e.g., Finke et al. 1992; Plucker and Renzulli 1999), which will be briefly discussed next. 2 The construct and measurement of creative ability Creativity can be defined as the generation of ideas or solutions that are both novel and useful (Amabile 1983) and is generally considered to be the result of an interplay 974 Mark Lett (2012) 23:973985 between personal factors, such as knowledge, skills, abilities and motivation, and contextual factors, such as creativity training and incentives (Amabile 1983; Kilgour and Koslow 2009). Goldenberg et al. (1999), for example, trained subjects in the use of creative templates derived from an analysis of patterns in previous innovations in order to boost their creative performance. One of the templates advises people to link previously unconnected attributes, e.g., making the color of a glass dependent on the temperature of the content. The use of templates influences the creative process by channeling thinking along predefined inven- tive routes (Goldenberg et al. 1999). The interactionist perspective on creativity (e.g., Woodman et al. 1993) argues that process- or context-related variables, such as providing subjects with templates (Goldenberg et al. 1999) or adding task constraints (e.g., Sellier and Dahl 2011), may interact with a persons innate creative ability. The use of creative templates could, for example, boost the performance of noncreative individuals, but may prove less effective for highly creative individuals. To not confound the validity coefficients with such interaction effects, this article focuses exclusively on the relationship between the creative ability of the person and creative outcomes, without interfering in the creative process or context. Now what does a persons creative ability consist of? Creative tasks are typically ill-defined with many parameters left unspecified, which opens up a vast solution space (Finke et al. 1992). According to creative cognition theories, people have a natural tendency to search for ideas within bounded areas of their knowledge network, i.e., they follow a path of least resistance (Ward 1994). The first ideas that come to mind are usually the most obvious ones based on easily accessible knowl- edge structures (Rietzschel et al. 2007). For example, when asked to draw a new type of alien, people tend to incorporate features of known earth animals (Marsh et al. 1999). Divergent thinking is needed for thinking outside the box and to generate more creative ideas (Guilford 1967; Torrance 1974). Divergent thinking comprises the following cognitive subskills: fluency, flexibility, and originality. Fluency is the ability to generate a large number of ideas in response to a problem. Originality is the ability to come up with novel, unusual ideas. Flexibility refers to the ability to produce ideas in different categories. Divergent thinking is necessary, but not suffi- cient. Turning a novel idea into something meaningful requires convergent thinking (Kilgour and Koslow 2009). Elaboration, which is the ability to work out an idea and embellish it with details (Torrance 1974), is a more convergent cognitive subskill (Kris 1952). For a comparative evaluation of different assessment methods of creative ability, it is imperative to clearly define the construct of interest (Robertson and Smith 2001). A good construct definition (see Rossiter 2002) includes the focal object (i.e., the person) and the attribute on which the focal object is to be evaluated (i.e., creative ability, comprising fluency, originality, flexibility, and elaboration). The rater entity, i.e., the person who provides the object-on-attribute judgments is an often neglected, yet important aspect. For example, customer ratings of IBMs service quality do not have to produce the same estimates as manager ratings of IBMs service quality (Rossiter 2002). For the validation studies that are presented next, measures were sought that rate the same object on the same attribute, but deploy different assessment methods or rater entities. Mark Lett (2012) 23:973985 975 3 Study 1: Prospective marketing employees This study involved prospective marketing employees, i.e., marketing students who were close to graduation. They usually do not have a portfolio of past creative work. Recruiters thus have to resort to other indicators of creative ability. Three brief assessment methods with different rater entities were scrutinized: a creative ability test scored by trained raters (test ratings), CVs rated by recruitment and selection professionals (CV ratings), and a questionnaire completed by the individual (self- ratings). Cognitive ability tests have been found to be good, if not the best, predictors of job performance across a wide range of occupations (e.g., Robertson and Smith 2001). CVs are a widely used method of selection that has received surprisingly little attention from researchers (Robertson and Smith 2001). And finally, self-rated ques- tionnaires can convey information about the person that may be difficult to observe by external raters. 3.1 Study framework Figure 1 depicts the framework for this study with the three predictor variables (left- hand side), viz. test ratings, student CV ratings, and self-ratings, and two criterion variables (right-hand side). The first criterion measure, viz. creative performance, concerns the subjects performance on a real-life creative marketing task as rated by marketing experts (work sample ratings). The second criterion measure, viz. early career creativity, concerns an assessment of the subjects creative ability as rated by recruitment and selection professionals based on early career CVs (i.e., 5 years after the collection of the other measures displayed in Fig. 1, indicated by T2 and T1, respectively). Criterion-related validity coefficients are commonly reported for indi- vidual methods (Robertson and Smith 2001). In practice, however, multiple methods are often used to assess the candidates likely performance. It is therefore worthwhile to also investigate the incremental validity of the methods, i.e., the extent to which they provide non-overlapping information (Robertson and Smith 2001). 3.2 Hypotheses The hypotheses will be largely based on the vast body of research on personnel selection in general, which has been summarized in meta-analytical studies (e.g., Schmidt and Hunter 2004). Empirical studies on the relationship between assessment methods of creative ability and creative performance in business are not as ubiquitous as for other cognitive abilities, such as general mental ability, with overall job performance as the criterion. In the absence of evidence for creative ability, these findings will be used to formulate the hypotheses. First, with work sample ratings as the criterion, one may expect higher operational validity for the test ratings of creative ability than for the student CV ratings or the self-ratings, since both the test ratings and the work sample ratings are derived from the subjects actual performance on a creative task, one being general in nature and the other one domain-specific. Cognitive ability tests, in general, have demonstrated high validity coefficients (of about 0.5) with overall job performance as the criterion (see Robertson and Smith 2001; Schmidt and Hunter 2004). As for the self-ratings, 976 Mark Lett (2012) 23:973985 there is some evidence that suggests that the validity coefficient with supervisory ratings of creative performance as the criterion is about 0.3 (e.g., Tierney and Farmer 2004). Thus: H1 Test ratings of creative ability have higher operational validity with creative performance on a marketing task (work sample ratings) as the criterion than student CV ratings and self-ratings. For the second criterion measure, viz. the expert-rated creative ability based on early career CVs, one may expect a different pattern. The student CV ratings are likely to show high operational validity, as past behavior is usually a good predictor of future behavior. Moreover, the information listed in the early career CVs and the student CVs partially overlaps. Self-ratings of creative ability represent a belief in ones ability to perform well on creative tasks (Tierney and Farmer 2004) or in ones aptitude for a creative job and may therefore steer career decisions. The extent to which a persons innate (test-rated) creative ability eventually translates into a creative career, however, will depend on many other personal and contextual factors (Amabile 1983). Work samples, such as a real-life marketing task, have generally demonstrated high validity coefficients (of about 0.5) with overall job performance as the criterion (see Robertson and Smith 2001). Thus: H2 Student CV and self-ratings of creative ability and work sample ratings of creative performance have higher operational validity with early career creativ- ity (CV ratings) as the criterion than test ratings. The question as to whether the three brief assessment methods of creative ability are able to pick up unique, non-overlapping information that is relevant to the .85 p<.01 Test rating (T1) Student CV rating (T1) Self-rating (T1) Creative Ability (T1) Early Career Creativity (T2) Creative Performance (T1) Work sample rating (T1) Early career CV rating (T2) Indicator Latent Construct Legend: T1 = 2005 T2 = 2010 .69 p<.01 .39 p<.01 R 2 = .48 R 2 = .35 .90 p<.01 .56 p<.01 .25 p<.05 Fig. 1 Conceptual framework study 1 Mark Lett (2012) 23:973985 977 criterion measures will be answered empirically. Personnel selection research has shown that combining methods (e.g., cognitive ability tests and work samples) can lead to higher validity coefficients (see Robertson and Smith 2001). But it is unclear whether these findings also translate to the narrower concept of creative performance with predictor measures that intend to assess the focal object on exactly the same attribute. 3.3 Participants and procedure Twenty marketing students (age, 1928; gender, 35 % female; education level, 70 % master level) from a Dutch university participated in return for a monetary reward (30 Euros). The participants were asked to design a creative marketing campaign for a well-known beer brand based on a real-life campaign brief (including marketing objectives, the desired theme, and some practical constraints). The participants first completed a creative ability questionnaire. In the universitys laboratory, they then collectively took a creative ability test after which they worked on the task individually (maximum, 3 h). They were also asked to submit their student and early career CVs (5 years later). Three students could not be tracked down 5 years later. 3.4 Measures Test ratings of creative ability (predictor) The Torrance Tests of Creative Thinking (TTCT; Torrance 1974) that measure the four cognitive subskills of creative ability, viz. fluency, flexibility, originality and elaboration, are by far the most commonly used tests of creative ability (Plucker and Renzulli 1999, p.39). There is also a 15- min version of the TTCT available, namely the Abbreviated Torrance Test for Adults (ATTA; Goff and Torrance 2002), which is more suitable for use in practice (see Althuizen et al. 2010). This test consists of one verbal and two figural tasks each with a 3-min time limit. The verbal task asks respondents to imagine a hypothetical situation (e.g., that you could walk on air) and to list as many problems as possible that might occur. The other tasks ask for figural responses in the form of pictures drawn by the testee, prompted by stimuli. Two trained raters independently scored all test booklets using a detailed scoring manual. The cognitive subskill scores are summed across tasks and then converted into normalized scores. The average across subskills forms the subjects test rating of creative ability. The inter-rater reliability coefficient (intraclass correlation (ICC) requiring absolute agreement, see Shrout and Fleiss 1979) was 0.96. Hence, the scores of the two raters were averaged. The mean test rating of creative ability was 16.0 (SD01.3) on a scale from 11 (0low) to 19 (0high). Self-ratings of creative ability (predictor) The Abedi Test of Creativity (ATC; see Auzmendi et al. 1996) is a 56-item, multiple-choice questionnaire that measures the same four cognitive subskills by means of self-ratings. An example of an item for fluency is: If you had to participate in a contest in which you were asked to come up with as many words as possible which began with the letter J, how would you do? (10poorly, 20okay, 30very well). An example for 978 Mark Lett (2012) 23:973985 originality is: Do people think that you come up with unique ideas? (10no, 20sometimes, 30often). Completing the ATC takes about 15 min and scoring the answers does not require expert judgment. The average of the item scores, up to two decimal places, forms the subjects self-rating of creative ability (M0 2.37, SD00.20). CV ratings of creative ability (predictor and criterion) The CVs contained informa- tion on education, work experience, activities, interests, skills and hobbies. As it is difficult to assess cognitive subskills based on a CV only, the following two items were used: Based on your reading of the CV, how would you rate the creative ability of this person (10minimal to 70substantial), How likely is it that this person receives an invitation for an interview for a creative marketing job (10not likely at all to 70very likely). Four recruitment and selection professionals agreed to rate the subjects creative ability and were randomly assigned to either the student CVs or the early career CVs. After checking the reliability of the two items within judges (Cronbach s >0.88), the item scores were averaged into a single CV rating of creative ability. The inter-rater reliability coefficient (ICC requiring consistency, see Shrout and Fleiss 1979) was 0.69 for the student CVs and 0.82 for the early career CVs. Hence, the scores were averaged across raters (student CV ratings: M04.56, SD01.35; early career CV ratings: M03.94, SD01.27). Work sample rating of creative performance (criterion) The campaign proposals were rated on two commonly used dimension of creativity (see Amabile 1983; Rubera et al. 2010), viz. novelty (three items: uniqueness, originality, and surprise) and usefulness (four items: effectiveness, fit with the brand, appeal, and usefulness) on 11-point scales (00lowest to 100highest; cf. Kilgour and Koslow 2009). Two agency directors and three brand managers, of whom two had served as chairman for the annual Marketing Campaign Awards, agreed to independently rate all proposals. Following a dimensionality and reliability check within judges (all Cronbach s>0.83), the item scores were combined into single novelty and useful- ness scores. To calculate a creative performance score, the mean of the novelty and usefulness scores was taken (Shalley et al. 2004). The inter-rater reliability coefficient (ICC requiring consistency) was 0.72. Hence, the creative performance scores were averaged across judges. The mean work sample rating of creative performance was 5.80 (SD00.68). Other potential indicators of creative ability Recruiters commonly use grade point averages (GPAs) to screen college applicants. Permission was therefore obtained to access the students grade records (after they had graduated). All course grades are reported on an 11-point, decimal scale ranging from 0 (0lowest) to 10 (0highest). The records were incomplete for four participants. The overall GPAs were broken down into a bachelor GPA (M06.88, SD00.47) and a master GPA (M07.15, SD00.43), as partial measures of GPAs, such as in- major GPAs, have been found to be more predictively valid for job performance (McKinney et al. 2003). The grade for the thesis (M07.47, SD00.61) was also included, since writing a thesis arguably allows for more creative expression than taking an exam. Mark Lett (2012) 23:973985 979 3.5 Results Table 1 shows the bivariate correlation coefficients between the predictor and the criterion measures, which is the most straightforward statistic for assessing the operational validity of the assessment methods. Performance of the predictor variables The correlation coefficient between the test ratings and the self-ratings is significant (r00.46, p00.044) but, in line with H1, the test ratings predict the students creative performance on the marketing task (work sample ratings) more accurately (r00.59, p00.006 versus r00.36, p00.117). The operational validity of the test ratings is quite impressive and comparable to the average validity coefficient of about 0.5 for general mental ability tests with overall job performance as the criterion. However, contrary to H1, the coefficient for the student CV ratings is even higher (r00.66, p00.004), although not significantly different from the coefficient for the test ratings (p00.749, based on a Fisher r-to-z transformation). Surprisingly, the student CVs contain sufficient information for professionals to accurately discriminate between individuals with low and high test ratings of creative ability, despite the rather homogeneous sample. As expected (H2), student CV ratings (r00.54, p00.025) and work sample ratings (r00.57, p00.009) have higher operational validity with early career CV ratings as the criterion than the test ratings (r00.40, p00.109) but also higher than the self-ratings (r00.25, p0 0.338). GPAs are not significantly related to the criterion measures (see Table 1). Thus, they should not be used as indicators of creative ability (GPAs are more likely to reflect general mental ability), although the correlation between the master thesis grade and the test ratings is marginally significant (r00.42, p00.075). When combined, the test ratings, student CV ratings and self-ratings are able to explain 48 % of the variance in the work sample ratings of creative performance Table 1 Correlation coefficients between predictor and criterion measures in study 1 1 2 3 4 5 6 7 Creative ability measures (predictors) 1. Test ratings 2. Student CV ratings 0.67* 3. Self-ratings 0.46* 0.16 Other potential indicators 4. Bachelor GPA 0.15 0.04 0.33 5. Master GPA 0.15 0.14 0.05 0.40 6. Thesis grade 0.42** 0.39 0.24 0.34 0.64* Criterion measures 7. Work sample ratings (creative performance) 0.59* 0.66* 0.36 0.14 0.13 0.09 8. Early career CV ratings (early career creativity) 0.40 0.54* 0.25 0.13 0.09 0.20 0.57* N020. Missing values for student and early career CVs (n017), bachelor GPA (n018), master GPA (n016) and thesis grade (n019) are replaced by the mean value *p<0.05, **p<0.10 980 Mark Lett (2012) 23:973985 based on a partial least squares analysis (smartPLS) of the model depicted in Fig. 1. This is 7 % (for the student CV ratings) to 36 % (for the self-ratings) more than the assessment methods can explain individually. Using a combination of test ratings and student CV ratings seems however sufficient, as they already explain 47 % of the variance in the work sample ratings. In combination with the work sample ratings, the three methods are able to explain 35 % of the variance in the early career CV ratings, which is 6 % (student CV ratings) to 28 % (self-ratings) more than the methods can explain individually. It should be noted that the work sample ratings alone already explain 32 % of the variance in the early career CVratings. However, obtaining work sample ratings is more cumbersome. 4 Study 2: Current employees of a marketing agency Study 1 revealed that self-ratings of creative ability have lower operational validity than test ratings and student CV ratings. Hunter and Hunter (1984) argue that the validity of self-ratings drops drastically for individuals who have not received sufficient feedback on the ability of interest, such as the marketing students in study 1. Study 2 therefore involves current marketing employees and investigates the conver- gent validity of test ratings, self-ratings, and supervisory ratings of creative ability. One would expect that supervisors, who are responsible for recruiting and selecting creative marketing personnel, are better able to assess their employees creative ability when they have had more time to observe them. Thus: H3 Test ratings, self-ratings, and supervisory ratings of creative ability show higher convergent validity for senior and long-time employees than for juniors and recent recruits. 4.1 Participants, procedure and measures Twenty-four employees of a Dutch marketing agency (annual gross income, five million euro), varying from art directors and designers to IT specialists (age, 2037; 50 % female; tenure, from a few months up to 8 years), agreed to complete the self- rated ATC and the ATTA test. Five years later, information was collected on the participants current jobs. The same raters as in study 1 independently scored the ATTA test booklets. The inter-rater reliability (ICC requiring absolute agreement) for the test ratings was 0.99. Hence, the test ratings were averaged across raters (M016.5, SD01.5). The mean for the self-ratings of creative ability was 2.32 (SD00.20). The CEO and the creative director agreed to independently rate their employees overall creative ability as well as the four cognitive subskills, viz. fluency, flexibility, originality, and elaboration. For this purpose, a six-item measurement scale was devised based on the subskill definitions provided in the ATTA manual (Goff and Torrance 2002). A sample item for flexibility is: How would you rate the ability of this person to approach a task from different points of view (10minimal to 70 substantial). The inter-rater reliability coefficient (ICC requiring consistency) was 0.78 for the single-item supervisory ratings and 0.88 for the composite supervisory ratings (i.e., the average of the six subskill items). The scores were thus averaged Mark Lett (2012) 23:973985 981 across supervisors. The mean supervisory rating of creative ability was 4.6 (SD0 0.99) for the single-item measure and 4.3 (SD00.70) for the composite measure. 4.2 Results Table 2 shows the correlation coefficients. The test ratings and self-ratings are highly correlated (r00.70, p<0.001), but they are only moderately correlated with the single-item supervisory ratings (test ratings: r00.34, p00.107; self-ratings: r00.48, p00.017) and the composite supervisory ratings (test ratings: r00.31, p00.141; self- ratings: r00.36, p00.087). The remainder of Table 2 shows the correlation coeffi- cients for employees who were less than 2 years with the company (recent recruits, n010) versus 5 years or more (long-time employees, n09) and for juniors (25 years and younger, n09) versus seniors (30 years and older, n08). The differences are as expected (H3), but the magnitude is striking. Overall, the three assessment methods Table 2 Means, standard deviations, and correlation coefficients for the creative ability measures in study 2 Variables Mean SD 1 2 3 All employees (n024) 1. Test ratings (ATTA) 16.5 1.5 2. Self-ratings (ATC) 2.32 0.20 0.70* 3. Supervisory ratings (composite) 4.3 0.70 0.31 0.36** 4. Supervisory ratings (single item) 4.6 0.99 0.34 0.48* 0.77* Recent recruits (2 years; n010) 1. Test ratings (ATTA) 16.9 1.1 2. Self-ratings (ATC) 2.41 0.11 0.35 3. Supervisory ratings (composite) 4.3 0.74 0.15 0.39 4. Supervisor ratings (single item) 4.8 1.2 0.02 0.61** 0.83* Long-time employees (5 years; n09) 1. Test ratings (ATTA) 16.3 2.0 2. Self-ratings (ATC) 2.29 0.22 0.82* 3. Supervisory rated (composite) 4.4 0.71 0.58 0.64* 4. Supervisory rated (single item) 4.5 0.87 0.79* 0.70* 0.69* Juniors (25 years; n09) 1. Test ratings (ATTA) 16.7 1.2 - 2. Self-ratings (ATC) 2.36 0.15 0.48 3. Supervisory ratings (composite) 4.1 0.62 0.01 0.01 4. Supervisory ratings (single item) 4.6 0.98 0.02 0.49 0.74* Seniors (30 years; n08) 1. Test ratings (ATTA) 16.2 2.2 2. Self-ratings (ATC) 2.34 0.19 0.96* 3. Supervisory ratings (composite) 4.4 0.75 0.61 0.65** 4. Supervisory ratings (single item) 4.6 0.83 0.69** 0.60 0.66** *p<0.05, **p<0.10 982 Mark Lett (2012) 23:973985 are highly convergent for seniors and long-time employees, but not for juniors and recent recruits (see Table 2). If the test ratings are operationally valid, high scores should be overrepresented among art directors and senior designers (n06), as they are hired to perform creative tasks (Kilgour and Koslow 2009). Sixty-seven percent (four out of six) of the art directors and designers scored a six or a seven on the test, which is significantly higher than the 16 % in Goff and Torrances (2002) sample of the general population ( 2 (1)011.123, p<0.001). The other two scored a 4, which is below the level of creativity required for the job (as assessed by the two supervisors). Five years later, they had left the agency. Five employees had a higher test rating of creativity than strictly needed for the job. Five years later, two of them were promoted and two had started their own company, which is arguably a creative endeavor. 5 Conclusions and discussion In a small-scale hiring context with prospective marketing employees, the combination of test ratings and student CV ratings of creative ability demonstrated high operational validity with creative performance and early career creativity as the criterion measures. Self-ratings should be approached with more caution. Self-ratings of creative ability are often used to investigate interaction effects between the employee and the work environment (see Shalley et al. 2004). This article shows that self-ratings as well as supervisory ratings may only provide reasonable and easy-to-administer alternatives to test ratings when more senior and long-time employees are concerned, but not for juniors and recent recruits. For studies on employee creativity, it may thus be worthwhile to check whether the results hold for different groups of employees. The use of self-ratings is even more questionable in high-stakes hiring contexts (Hunter and Hunter 1984). The unexplained variance in the criterion measures in study 1 indicates that creative ability alone is not a guarantee for creative performance or early career creativity. Other selection instruments, such as personality tests and assessment centers, may still prove valuable. Last but not least, establishing a fit between the creative demands of a job and the employees creative ability may lead to more job satisfaction and less stress (Livingstone et al. 1997). Because sampling errors are explicitly included in the calculation of statistical tests, there should not be a bias against statistically significant results obtained fromproperly selected small samples (Sawyer and Peter 1983, p.124). One would expect that large effect sizes obtained with small samples can be replicated easily (Sawyer and Peter 1983). Additional studies in other small-scale hiring contexts are nonetheless desirable to further strengthen the confidence in the reported findings. A priori, there is however no reason to expect that the operational validity of the investigated assessment methods will be different for other jobs for which creativity is an important component (see Scherbaum 2005). To stress the importance of finding good selection instruments, Hunter and Hunter (1984, p.7273) argue that any instrument of lower validity would incur very high costs, because even minute differences in validity translate into large dollar amounts. Creativity is important as it can provide a competitive edge (1) to compa- nies in an increasingly global marketplace and (2) to employees facing the automa- tion of relatively structured tasks by computers, robotics and other technologies. Mark Lett (2012) 23:973985 983 Acknowledgments The author would like to thank the editor and two anonymous reviewers for their valuable comments and suggestions. Timothy Heath, Ayse nculer, John Rossiter, and Berend Wierenga are thanked for their feedback and comments on earlier versions of the manuscript. The author is also grateful to Ed van Eunen, Ren Hendriks, Tom Wilms, Pieter Brusman, Eric Welles, Sebastiaan Bongers, Natasja Deelen, Jeanette Bisschop, Ragnhild Savenije, and Willem Royaards for their contributions to this research project. References Althuizen, N., Wierenga, B., & Rossiter, J. R. (2010). The validity of two brief measures of creative ability. Creativity Research Journal, 22, 5361. Amabile, T. M. (1983). The social psychology of creativity. New York: Springer. Andrews, J., & Smith, D. C. (1996). In search of the marketing imagination: factors affecting the creativity of marketing programs for mature products. Journal of Marketing Research, 33, 174187. Auzmendi, E., Villa, A., & Abedi, J. (1996). Reliability and validity of a newly constructed multiple-choice creativity instrument. Creativity Research Journal, 9, 8995. Coppedge, M. (2007). Theory building and hypothesis testing: large- vs. small-N research on democrati- zation. In G. Munck (Ed.), Regimes and democracy in Latin America: theories and findings. Oxford: Oxford University Press. Finke, R. A., Ward, T., & Smith, S. (1992). Creative cognition: theory, research, and applications. Cambridge: MIT Press. Goff, K., & Torrance, E. P. (2002). Abbreviated Torrance test for adults manual. Bensenville: Scholastic Testing Service. Goldenberg, J., Mazursky, D., & Solomon, S. (1999). Toward identifying the inventive templates of new products: a channeled ideation approach. Journal of Marketing Research, 36, 200210. Guilford, J. P. (1967). The nature of human intelligence. New York: McGraw-Hill. Heneman, R. L., Tanksy, J. W., & Camp, S. M. (2000). Human resource practice in small and medium-sized enterprises. Entrepreneurship Theory and Practice, 25, 1116. Hunter, J. E., & Hunter, R. F. (1984). Validity and utility of alternative predictors of job performance. Psychological Bulletin, 96, 7298. IBM. (2010). Capitalizing on complexity: insights from the global chief executive officer study. Somers: IBM. Kabanoff, B., & Rossiter, J. R. (1994). Recent developments in applied creativity. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (pp. 283324). London: Wiley. Kilgour, M., & Koslow, S. (2009). Why and how do creative thinking techniques work? Trading off originality and appropriateness to make more creative advertising. Journal of the Academy of Market- ing Science, 37, 298309. Kris, E. (1952). Psychoanalytic exploration in art. New York: International Universities Press. Livingstone, L. P., Nelson, D. L., & Barr, S. H. (1997). Person-environment fit and creativity: an examination of supply-value and demand-ability versions of fit. Journal of Management, 23, 119146. Marsh, R. L., Ward, T. B., & Landau, J. (1999). The inadvertent use of prior knowledge in a generative cognitive task. Memory and Cognition, 27, 94105. McKinney, A. P., Carlson, K. D., Mecham, R. L., DAngelo, N. C., & Connerley, M. L. (2003). Recruiters use of GPA in initial screening decisions: higher GPAs dont always make the cut. Personnel Psychology, 56, 823845. Patterson, F., Kerrin, M., Gatto-Roissard, G., & Coan, P. (2009). Everyday innovation: how to enhance innovative working in employees and organisations. London: NESTA. Plucker, J. A., & Renzulli, J. S. (1999). Psychometric approaches to the study of human creativity. In R. J. Sternberg (Ed.), Handbook of creativity (pp. 3561). Cambridge: Cambridge University Press. Rietzschel, E. F., Nijstad, B. A., & Stroebe, W. (2007). Relative accessibility of domain knowledge and creativity: the effects of knowledge activation on the quantity and originality of generated ideas. Journal of Experimental Social Psychology, 43, 933946. Robertson, I. T., & Smith, M. (2001). Personnel selection. Journal of Occupational and Organizational Psychology, 74, 441472. Rossiter, J. R. (2002). The C-OAR-SE procedure for scale development in marketing. International Journal of Research in Marketing, 19, 305335. 984 Mark Lett (2012) 23:973985 Rubera, G., Ordanini, A., & Mazursky, D. (2010). Toward a contingency view of new product creativity: assessing the interactive effects of consumers. Marketing Letters, 21, 191206. Sawyer, A., & Peter, J. P. (1983). The significance of statistical significance tests in marketing research. Journal of Marketing Research, 20, 122133. Scherbaum, C. A. (2005). Synthetic validity: past, present and future. Personnel Psychology, 58, 481515. Schmidt, F. L., & Hunter, J. E. (2004). General mental ability in the world of work: occupational attainment and job performance. Journal of Personality and Social Psychology, 86, 162173. Schmitt, N. (2007). The value of personnel selection: reflections on some remarkable claims. Academy of Management Perspectives, 21, 1923. Scratchley, L. S., & Hakstian, A. R. (2001). The measurement and prediction of managerial creativity. Creativity Research Journal, 13, 367384. Scullen, S. E., & Meyer, B. C. (2012). More applicants or more applications per applicant? A big question when pools are small. Journal of Management. doi:10.1177/0149206312438774. Sellier, A.-L., & Dahl, D. W. (2011). Focus! Creative success is enjoyed through restricted choice. Journal of Marketing Research, 48, 9961007. Shalley, C. E., Zhou, J., & Oldham, G. R. (2004). The effects of personal and contextual characteristics on creativity: where should we go from here? Journal of Management, 30, 933958. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: uses in assessing rater reliability. Psychological Bulletin, 86, 420428. Tierney, P., & Farmer, S. M. (2004). The Pygmalion process and employee creativity. Journal of Manage- ment, 30, 413432. Torrance, E. P. (1974). The Torrance tests of creative thinking: norms-technical manual. Lexington: Personnel Press. Ward, T. B. (1994). Structured imagination: the role of category structure in exemplar generation. Cognitive Psychology, 27, 140. Woodman, R. W., Sawyer, J. E., & Griffin, R. W. (1993). Toward a theory of organizational creativity. Academy of Management Review, 18, 293321. Mark Lett (2012) 23:973985 985