Performance Bonuses and the Potential Signaling Void

An additional sorting mechanism concerning attraction to nonprofit organizations could potentially be the effects of “industry culture stereotypes” posited by De Goede,

98 Ibid., 1641-1642.

99 Cadsby et al., 387.

100 Kuhn, 1635.

Van Vianen and Khele. Person-industry (P-I) fit is an extension of

person-environment cross-level analyses in which job-seekers are not thought of as tabula rosa before undertaking a job search, but instead have preconceptions about individual companies and the sectors in which these organizations operate. Once again invoking Spence’s (1974) signaling theory, when lacking familiarity with an organization, job-seekers often make assumptions about organizations they encounter by defaulting to industry stereotypes.¹⁰¹ This mental ‘schemata’ anchors a

job-seeker’s perception while searching,¹⁰² influencing the subsequent information learned throughout the process. As nonprofits in the modern era become more professionalized and increasingly consider variable pay of one sort or another, applicants may be surprised to learn that nonprofit positions entail PFP plans.

Clarifying the preconceptions held by nonprofit applicants is also a worthwhile pursuit for recruitment and management professionals wanting to avoid low or even nonexistent levels of fit. Chatman and Barsade (1995) define “low fit” as a situation in which an individual feels that an organization does not share the same values. An even more incongruous matching might constitute “misfit” between individuals and the organization–a situation where a set of values strongly endorsed by the individual are thought to be diametrically opposed to those of the organization.¹⁰³ Therefore, avoiding situations in which the values of the organization may be misrepresented or concealed from the applicant would seem to be in the best interest of fit.

Understanding applicant preconceptions about PFPs between sectors is one way to minimize the potential for situations of low fit or misfit. If PFP plans are not explicitly mentioned in a nonprofit job listing, and assuming nonprofit sector

101 De Goede et al., 51.

102 Ibid., 52.

103 Robert and Wasti, 548.

applicants begin job searching with the industry stereotype that nonprofit

organizations do not (or are not allowed to) offer bonus incentives, an uncomfortable clash of values may ensue once the applicant is informed of the incentive

arrangements. The fourth hypothesis aims to find evidence of industry stereotypes in the preconceptions about bonus incentives held by nonprofit and for-profit applicants:

Hypothesis 4: Assuming job applicants are searching within their preferred sector, if a job posting does not explicitly mention pay-for-performance bonuses, nonprofit applicants are less likely to believe that the position will entail performance bonuses than for-profit applicants.

Failing to mention PFP arrangements in the ads likely removes a demonstrated sorting mechanism and could have important practical implications for the growing number of nonprofits attempting to induce higher employee performance.

Much of the pay-for-performance literature stresses the differences between the intrinsic motivation, extrinsic motivation and amotivation of workers,¹⁰⁴ which is highly correlated with the employee’s fit with an organization.¹⁰⁵ In fact, the

proliferation of PFP measures within the nonprofit sector has produced documented incongruences between the majority intrinsically motivated nonprofit workforce and relatively new extrinsic measures meant to stoke motivation. A growing number of studies have already explored these budding dynamics with mixed results,¹⁰⁶ some of which have found a negative correlation between the introduction of extrinsic reward initiatives and employees favoring intrinsic motivation.¹⁰⁷ Given that these sorts of tensions are known to surface in the third sector, explicitly signaling to individuals of a certain kind—namely those driven largely by extrinsic motivations—should be

104 Ryan and Deci, 61.

105 Frey, B. S., and R. Jegen. "Motivation Crowding Theory: A Survey of Empirical Evidence." Journal of Economic Surveys 15, no. 5 (2001): 589.

106 Speckbacher, 1010.

107 Speckbacher, 1006-025

prioritized by certain nonprofits. Since PFP arrangements are meant to attract more extrinsically motivated people, unattached workers of this sort might be missing essential information that would facilitate the kind of self-selection. Many of those organizations may be missing out on exactly the type of individuals their policies are designed to attract and motivate.

Chapter 3 Methodology

Despite promising findings inside and outside of the laboratory, there remains only a handful of empirical cases for PFP sorting effects on the for-profit workforce,¹ and the phenomenon has been left virtually untouched regarding the nonprofit sector.

In the private sector, support sorting comes from a 1997 study of corporate human resources records concluding that highly motivated individuals were more likely to leave the firm if their exceptional was not met with a corresponding increase in pay.² Harrison et al. (1996) demonstrated complementary findings in which employees with lower motivation were more likely to stay when performance and pay were weakly associated.³ The previously mentioned study on sorting effects in the

Safelight Glass Corporation is understood as “exceptionally important”⁴ partially due to Lazear’s ability to tease out conclusive findings in a “real world” context.⁵ The quantitative data under examination in this instance proceeds in a quasi-laboratory setting, in which potential applicants rate the attractiveness of a fictional job

1 Cadsby Song Tapon, 388.

2 Trevor, Charlie O., Barry Gerhart, and John W. Boudreau. "Voluntary Turnover and Job Performance:

Curvilinearity and the Moderating Influences of Salary Growth and Promotions." Journal of Applied Psychology82, no. 1 (1997): 44-61. doi:10.1037//0021-9010.82.1.44.

3 Harrison, David A., Meghna Virick, and Sonja William. "Working without a Net: Time, Performance, and Turnover under Maximally Contingent Rewards." Journal of Applied Psychology 81, no. 4 (1996): 331-45.

doi:10.1037//0021-9010.81.4.331.

4 Cadsby et al., 388.

5 Lazear, Edward P. "Performance Pay and Productivity," 1347.

description similar to those found at a typical e-recruitment website.⁶ As other scholars in attraction and searching have readily conceded, many of the conclusions reached in controlled settings are difficult to generalize.⁷ Although no method is without faults, obtaining quantitative data inspired by real job postings from

e-recruitment websites may translate into more generalizable implications for nonprofit organizations and job seekers.

3.1 Participants

In order to detect an effect size of 0.25 (a medium effect size for 2 ANOVA factors) it was determined that at least 158 respondents would be needed.⁸ Participants were 173 potential job applicants divided into two subsamples–the nonprofit sector and for-profit sector (86 nonprofit and 87 for-profit respondents). A convenience sample of participants that self-identified as job searching applicants or future applicants in their respective sector were informed that the online survey, which offered a chance to win a random drawing of $20 US ($600 NTD) and was distributed primarily by email and Facebook, investigated motivational tendencies and organizational

attraction in the workplace. Among the 173 participants, 123 identified as female, 48 identified as male and 2 chose not to disclose their gender. All age ranges were represented from 18-24 to 65-74, with most participants (77) coming from the 18-24 category. The second-most represented age range was 25-34 (66). The majority of participants self-identified nation of origin was either the U.S. or Taiwan (30.06%

and 32.27%, respectively), with a host of other countries comprising rather minute percentages of the sample set. An overall majority of participants (51.45%) currently hold at least a bachelor’s degree, followed by master’s degree graduates (23.70%)

6 Kuhn, 1638.

7 Cadsby et al., 401.

8 Cohen, Jacob. Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: L. Erlbaum Associates, 1988. <http://www.statmethods.net/stats/power.html>.

coming in at a distant second. An overall majority of participants were also students (52.60%) with a mean of 7.6 years of work experience.

3.2 Manipulation and Procedure

Participants were initially directed to either the for-profit or non-profit portion of the survey according to their preference, and then completed socio-demographic

questions. The individual characteristics section of the survey included forward-looking and hypothetically phrased items from the Gangé et al. (2010) MAWS scale, items from the Triandis and Gelfand (1998) allocentrism and idiocentrism scale, and the items chosen by Gomes and Neves (2011) on attraction from Highhouse (2003).

The survey closed with a question on the likelihood of expecting a performance bonus plan if not explicitly mentioned in the job posting (Appendix 1).

Participants in both survey types were randomly assigned one of three fictitious job descriptions (meaning that participants could receive one of six different advertised job descriptions in a 2x3 survey design), all of which were nearly identical with the sole exception being the experimental manipulation of compensation offers. Because of the unassailable dominance of web-based recruiting in recent years,⁹ and to avoid the well-documented effects of individual corporate websites as a recruitment medium,¹⁰ this survey asked the participants to imagine the job listing as if it were posted on a widely used e-recruitment website rather than a website hosted an organization. Wording from Kuhn and Yockey (2003) and Kuhn (2009) was

9 Backhaus, 117.

10 Cober, R. T., D. J. Brown, L. M. Keeping, and P. E. Levy. "Recruitment on the Net: How Do Organizational Web Site Characteristics Influence Applicant Attraction?" Journal of Management 30, no. 5 (2004): 623-46.

doi:10.1016/j.jm.2004.03.001. as cited in De Goede et al. "Attracting Applicants on the Web: PO Fit, Industry Culture Stereotypes, and Website Design."

incorporated for the fictitious job description along with additional information meant to make the advertisement more realistic.¹¹ The exact question wording is as follows:

Imagine that you are using an e-recruitment website (e.g. Monster.com, Indeed.com, idealist.org) to search for a full-time job in the your chosen (for-profit) nonprofit field. During this time, you come across the job description of Organization X, a (for-profit) nonprofit organization located in a relatively convenient location for you. The duties and responsibilities required of the position match your expectations and you estimate that your chances for

promotion and regular cost of-living pay increases would be reasonable–roughly the equivalent of the market standard. Upon reading the compensation section provided by the company, please indicate your level of attraction and intent to apply to the position according to the scale below:

At Organization X our vision is clear – to be the best. We anticipate community needs and deliver superior products and services that genuinely improve people’s lives. This is the place for people who want to be center stage in one of the world’s most fascinating and dynamic industries. We want extraordinary people who share our passion for the industry and our vision for success. Organization X offers stimulating and challenging careers, and [compensation manipulation].

There are many fantastic benefits to a career at Organization X. In addition to working to affect positive change in the health and lives of thousands, Organization X offers its employees a competitive salary

11 Kuhn, 1638. and Kuhn and Yockey, "Variable Pay as a Risky Choice: Determinants of the Relative

Attractiveness of Incentive Plans." Organizational Behavior and Human Decision Processes 90, no. 2 (2003): 326.

doi:10.1016/S0749-5978(02)00526-5.

and a stimulating work environment with comprehensive benefits including medical, dental, life insurance, flexible spending accounts, 403(b) and 401(k), generous paid time off, onsite Wellness facilities, and educational assistance.

If you want to make a difference in the lives of thousands of people and if the position speaks to your capabilities, experience and commitment to improve our mission, this is the place for you!

The three compensation manipulation versions spread across the two sectors

consisted of an individual performance bonus, a collective performance bonus, and a description with a fixed salary. Participants randomly assigned to these versions would have seen one of these three possibilities:

Individual performance: ...and we reward our top-performing employees at each level with bonuses of up to twenty percent of salary.

Collective performance: ...and we reward employee teams that demonstrate excellent performance at each level with bonuses of up to twenty percent of salary.

Fixed Salary: ...Organization X offers its employees a competitive salary...

The percentage bonus of 20% offered in the simulated job descriptions is deliberately on the high end of average findings from an unofficial survey of more than 60

nonprofit organizations conducted by a major HR consulting firm. Performance bonuses ranged from 5-10% of annual income for staff and non-management to 30-50% for CEOs. Supervisory staff typically received 5-15%, middle management 10-20%, senior management 15-30%, and executive management 20-40%.¹² Additional support for the hypothetical amount was further informed by another informal poll of

12"Astron Solutions HR Interview." Online interview. 31 Mar. 2015.

more than 20 nonprofit organizations by a website that regularly polls for salary comparisons on an estimated 15,000 jobs in 3,000 different industries.¹³

Participants read one of the randomized descriptions, then evaluated the five items from Highhouse (2003) on attraction to the organization based on the fictitious description.

3.3 Measures

The heterogeneous spread of Likert scales used in the previous studies (Gomes and Neves five-point scale, Gagné et al. seven-point scale and Triandis and Gelfand nine-point scale) were converted into a more traditional seven-nine-point scale across the entire survey. Unless otherwise indicated, items were measured using a seven-point scale ranging from 1= Strongly disagree to 7 = Strongly agree.

Psychometric scales from three separate journal publications comprised the work motivation, horizontal allocentrism and idiocentrism, organizational attractiveness and intention to apply to a job vacancy measures.

Work motivation. Twelve items from Gagné et al. (2010) known as the

Motivation at Work Scale (MAWS) split between four subscales (Intrinsic, Identified, Introjected and Extrinsic) were used to measure the orientation and level of intrinsic and extrinsic motivation possessed by the hypothetical potential applicants. MAWS was designed by Gagne et al. as a psychometric incarnation of Deci and Ryan’s self-determination theory (SDT).¹⁴ Gagne et al. (2010) found these statistical constructs held after polling 1,644 workers across two different languages,¹⁵ in which the reliabilities ranged in their English study from α = .69 to .89 across the four

13 "Non-Profit Organization Salaries - Non-Profit Organization Salary Survey - PayScale." Non-Profit Organization Industry Salary, Average Salaries. Accessed 2015.

http://www.payscale.com/research/US/Industry=Non-Profit_Organization/Salary.

14 Gagné et al., 642.

15 Ibid., 629.

subscales.¹⁶ While α =.69 may be adjudged as not particularly robust, lower reliabilities could be considered acceptable when it comes to measures of abstract broad social values.¹⁷ Items include: “Because this job would fit my personal values.”

Horizontal allocentrism and idiocentrism. Eight items from Triandis and Gelfand (1998) were used to measure both the horizontal allocentric and horizontal idiocentric dispositions possessed by the respondents.¹⁸ A more expansive selection of the Triandis and Gelfand (1998) scale was used by Kuhn (2009), in which the allocentrism scale achieved a reliability of α = .78 and idiocentrism achieved a scale reliability of α = .72 in that study.¹⁹ Items include: “I'd rather depend on myself than others.”

Organizational attractiveness. Three items from Highhouse (2003) were used to measure organizational attractiveness. The items selected were those of the highest factor loading as determined by Gomes and Neves (2011). Items include: “I find this a very attractive company.”

Intention to apply to a job vacancy. Two items from Highhouse (2003) were used to measure organizational attractiveness. The items selected were those of the highest factor loading as determined by Gomes and Neves (2011). Items include: “If I were searching for a job, there would be a strong probability of applying to this offer.”

16 Alpha coefficients for Intrinsic (.89), Identified (.83), Introjected (.75), and Extrinsic (.69)

17 Singelis, T. M., H. C. Triandis, D. P. S. Bhawuk, and M. J. Gelfand. "Horizontal and Vertical Dimensions of Individualism and Collectivism: A Theoretical and Measurement Refinement." Cross-Cultural Research 29, no. 3 (1995): 240-75. doi:10.1177/106939719502900302.

18 Triandis and Gelfand, 120.

19 Kuhn, 1639.

Chapter 4 Results and Discussion 4.1 Socio-demographic Descriptive Statistics

The four hypotheses were tested via several statistical analysis methods using

STATA, including chi-square, t-test, Mann-Whitney U, Kruskal-Wallis and Analysis of Variance (ANOVA). For the socio-demographic results that could be compared, Pearson’s chi-square analysis of independence was conducted.¹ In comparing both sectors, the relation between age groups χ² (5, N = 173) = 5.10, p = .40, educational levels χ² (7, N = 173) = 6.23, p = .072, and work experience t (171, N = 173) = 8.40, p = .30 were not found to be significant at the 95% level. The only statistically significant difference found was in gender distribution χ² (2, N = 173) = 11.26, p

< .001, in which women far outnumbered men in the non-profit group (83%) compared to the for-profit group (60%). The disproportionate amount of women in the combined sample population applicants and the overwhelming percentage (83%) in the nonprofit subsample is a source for potential bias to be discussed subsequently in Chapter 5. Alpha coefficients measuring for internal consistency among the integrated subscales are provided in Table 1. Out of seven coefficients spread among three subscales, only two (identification and allocentrism) fall below .80, but remain above the .70 standard understood as “acceptable” for the social sciences.² Two of the subscales (intrinsic/extrinsic and allocentrism/idiocentrism) were also deemed acceptable above the .70 standard.

Table 1. Cronbach’s Alpha of the Measures

Scale N Cronbach’s alpha

1 Kremelberg, David. "Pearson's R, Chi-square, T-Test, and ANOVA." In Practical Statistics: A Quick and Easy Guide to IBM SPSS Statistics, STATA, and Other Statistical Software, 120-28. Los Angeles: SAGE Publications, 2011.

2 George, Darren, and Paul Mallery. SPSS for Windows Step by Step: A Simple Guide and Reference, 11.0 Update. 4th ed. Boston: Allyn and Bacon, 2003.

Intrinsic/Extrinsic Scale 12 .783

Intrinsic 3 .882

Identification 3 .777

Introjected 3 .848

Extrinsic 3 .912

Allocentrism/Idiocentrism 8 .728

Allocentrism 4 .778

Idiocentrism 4 .805

Attraction 5 .966

As seen in Table 1 above, the Cronbach’s alpha values for the study are at least acceptable. The intrinsic/extrinsic scale composed of the four subscales recorded a Cronbach’s alpha of .78. The allocentrism and idiocentrism scale recorded a Cronbach’s alpha of .73, while attraction scored the highest of the psychometric scales with a .97 value.

4.2 Hypothesis 1

Hypothesis 1: Intrinsically satisfying work is more important for nonprofit job applicants than for-profit job applicants.

The intrinsic and extrinsic motivational profiles of nonprofit and for-profit applicants were assessed by the complete Gagné et al. (2010) MAWS scale, which subdivides intrinsic and extrinsic measures into four subscales according to the Ryan and Deci SDT (1985) model. Each subscale was analyzed using an independent t-test or the more conservative non-parametric equivalent (the Mann-Whitney U test) depending on conditions of normality,³ beginning with the intrinsic motivation subscale.

Nonprofit applicants (N = 86) generally scored higher on psychometric measures of intrinsic motivation M = 16.91 (SD = 2.84). By comparison, for-profit applicants generally scored lower on psychometric measures of intrinsic motivation M = 14.06 (SD = 3.24). To discern the statistical significance of the discrepant mean values, an

3 Billiet, Paul. The Mann-Whitney U-test -- Analysis of 2-Between-Group Data with a Quantitative Response Variable. 2003. http://psych.unl.edu/psycrs/handcomp/hcmann.PDF.

independent t-test assuming equal variance was performed as to whether intrinsic enjoyment in the workplace is more important for nonprofit applicants when

compared to for-profit sector applicants. Prior to conducting the t-test, the assumption of normality was evaluated and confirmed with distributions associated with a

respective skew and kurtosis of .02 and .35, p = .05. The assumption of homogeneity of variances was tested and satisfied via Levene’s F test, F(171) = 1.75, p = 0.19.⁴ The independent samples t-test was associated with a statistically significant effect t(171) = 6.15, p = 0.00. The difference in means between nonprofit and for-profit applicants is statistically significant, with nonprofit applicants definitively preferring intrinsic enjoyment in the workplace as compared to for-profit applicants. A

graphical representation of the means and the 95% confidence intervals can be found in Table 2.

Table 2. Student’s t-test comparing nonprofit and for-profit applicant intrinsic motivation levels

Group n M SD t df p

Nonprofit 86 16.90698 2.8351902 6.1494 171 p < .001 For-profit 87 14.05747 3.2434418

Nonprofit applicants (N = 86) likewise scored higher in the identification subscales M = 17.26 (SD = 2.77). For-profit applicants generally scored lower in the

identification subscales M = 15.26 (SD = 3.43). To discern the statistical significance of the discrepant mean values, a Mann-Whitney U test (a non-parametric t-test analogue) was performed as to whether identification aspects of motivation were stronger in nonprofit applicants in comparison with for-profit applicants. The Mann-Whitney U test was preferred to a standard t-test because assumptions of normality were rejected with respective skew and kurtosis levels of .00 and .40, p = .01. As

4 Schmider, Emanuel, Matthias Ziegler, Erik Danay, Luzi Beyer, and Markus Bühner. "Is It Really

Robust?" Methodology: European Journal of Research Methods for the Behavioral and Social Sciences 6, no. 4 (2010): 147-51. doi:10.1027/1614-2241/a000016.

illustrated in Table 3, the Mann-Whitney U test was associated with a statistically significant effect (U = 2462.50, p < .001). The difference between nonprofit and

illustrated in Table 3, the Mann-Whitney U test was associated with a statistically significant effect (U = 2462.50, p < .001). The difference between nonprofit and