Procedures

In the present study, we choose the computer manufacturing industry as the target industry in examining the relationships among corporate image dimensions and organizational attractiveness. In contrast to other industries, manufacturing industry companies normally tend to release more information about themselves to the public owing to these companies’ need either to accelerate the transfer of

technology or to meet government regulations. The applicants may acquire relevant information easily. Moreover, considering the highly related nature of corporate product image and corporate service image in certain industries (e.g., service industry), it is likely that choosing a target industry with clearer corporate product image and service image may facilitate the manipulation in the experiments. We therefore chose the computer manufacturing industry as our target. Referring to the information disclosed in the mass media of several famous computer manufacturing companies in Taiwan, we formed the corporate image stimuli and each of the scenarios.

Questionnaires were distributed to final-year undergraduate and graduate students in class at eight universities and to experienced employees from eleven companies in Taiwan. After filling out the first section of the questionnaire containing items related to individual difference variables and demographic information, the participants were randomly assigned to one of sixteen recruitment scenarios. Participants were instructed to adopt the role of an applicant. The instructions provided general information about the conditions of organizations and job vacancy. Then, the participant would read a half-page featuring written comments purported to have been made in a famous business magazine regarding a computer manufacturing company. After reading the comments, the participants completed the designated questionnaire including items related to organizational attractiveness and the manipulation checks. Participation was voluntary and anonymous

To minimize concerns about the effects of different participant-group characteristics, we conducted a series of analyses to ensure there were no statistical differences in the demographic profiles between students and experienced

employees. The results showed that, except for age and willingness to apply (i.e., applicant willingness to apply for jobs in the near future), there was no significant difference in either gender (χ² = .67, p > .05) or another application willingness variable, “willingness to apply for jobs in the computer manufacturing industry” (F

= .71, p > .05). Therefore, the two variables (i.e., age and willingness to apply) were statistically controlled in this study.

3.3 Independent Variable Manipulations

We manipulated four types of corporate images (i.e., corporate product image, corporate service image, corporate citizenship image, and corporate credibility image) by changing the information attributable to the magazine’s commentary reports on the images of corporations at two levels (i.e., high or low). Referring to previous research, we took an approach similar to the approach for testing the effects of different organizational characteristics on individuals (e.g., Aquino, Tripp, & Bies, 2006; Lievens et al., 2005).

We used a 2*2*2*2 between-subject factorial design, which thus formed sixteen scenarios. We created corporate image stimuli based on the original framework and definitions proposed by Keller (2002) and also referred to real company image information disclosed in mass media. The stimuli were further revised by three senior marketing and HR managers and two I-O psychology professors. It should be noted that for the manipulated corporate image variables, four dummy variable was created (coded 0 = “low image,” 1 = “high image”). The following is a sample of image manipulation (the level of the manipulated image stimuli is in brackets).

This company has committed itself to providing nothing but world-class

high-quality products to its customers. It also keeps developing a broad range of innovative and high-performance products boasting all the latest features and practical functionality [high corporate product image]. In serving and supporting customer needs, this company maintains quality services and also sustains a culture of continuous improvement to ensure customer satisfaction.

Well-established training programs and communication channels for employees greatly improve and strengthen the company’s services [high corporate service image]. We found that this company uses resources and handles waste efficiently. Its present facilities are certified for ISO-9001 and ISO-14001 international standards to ensure workplace safety and environmental protection. The company has also devoted itself to community activities, helping people in need, and is dedicated to becoming a leading corporate citizen through active sponsorship of educational, charitable, and cultural activities in Taiwan [high corporate citizenship image]. Committed to premium professional performance, the company has received industry-wide recognition and is considered trustworthy and reliable by customers. It has also been awarded the National Excellence Award for expertise in operations. These accomplishments elevate the company’s reputation among customers and vendors alike [high corporate credibility image].

3.4 Measures

3.4.1 Need for affiliation

We used a 5-item scale adapted from Steers and Braunstein (1976) to assess this construct. Reviewing past research, we found that a comparatively low alpha for

the five-item need for affiliation measure indicated the need to increase the number of items. Thus, we added two items based on the calculation of Spearman-Brown formula and the conceptual definition of the variable. Two sample items used here are “I enjoy belonging to groups and organizations” and “I tend to build close relationship with others.” Subjects were asked to indicate their agreement using a 6-point Likert scale (1 = strongly disagree; 6 = strongly agree). The coefficient alpha was .81 for this seven-item measure.

3.4.2 Environmental sensitivity

We measured this construct by using a 15-item scale adopted from Berger and Kanetkar (1995). The scale used by Berger and Kanetkar (1995) was adapted from a part of Antil and Bennett’s (1979) Socially Responsible Consumer Behavior scale (SRCB). The original scale in Berger and Kanetkar (1995) contained 20 items.

However, four items concern specific consumer behavior (e.g., consumer interests and environmentally sound product characteristics). Moreover, the content of another item has been emphasized in the environmental protection laws in Taiwan and is well known as the prerequisite for the operation of each company. In order to prevent no variation in response, these items were removed. Two sample items used in the present study are “Pollution is presently one of the most critical problems facing this nation” and “Natural resources must be preserved even if people must do without some products.” Subjects were asked to indicate their levels of agreement using a 6-point Likert scale (1 = strongly disagree; 6 = strongly agree). The Cronbach’s alpha for this fifteen-item measure was .84 in the present study.

3.4.3 Materialism

Following the recommendation of Richins (2004), we used a revised shorter version of the Material Values Scale (MVS; Richins and Dawson, 1992) to measure

how respondents rated the degree to which they rated both possessions and acquisition of possessions as necessary or desirable in reaching goals. According to Richins (2004), the shorter version of the MVS is designed to assess the construct at a general level and has good psychometric properties. Therefore, we adopted the 9-item scale to measure materialism. Two sample items are “I admire people who own expensive homes, cars, and clothes” and “The things I own say a lot about how well I’m doing in life.” Respondents were asked to indicate their levels of agreement using a 6-point Likert scale (1 = strongly disagree; 6 = strongly agree). The Cronbach’s alpha for this measure was .79 in the present study.

3.4.4 Organizational attractiveness

Organizational attractiveness was measured at the individual level, in line with the measurements used by Collins and Stevens (2002) and Lievens et al. (2005).

Three items from Turban and Keon (1993) and Highhouse et al. (1999) were used to measure an applicant’s willingness to pursue jobs in an organization. Items include:

“I’d prefer a job there over a job in most other companies,” “If I were looking for a job, a job there would be very appealing,” and “If I were looking for a job, I would have strong motivation to apply for a job there.” Subjects were asked to base their responses on 6-point Likert scale (1 = strongly disagree, 6 = strongly agree). The Cronbach’s alpha for this measure was .90 in the present study.

3.4.5 Control variables

“Willingness to apply for jobs in the near future” was controlled. Subjects were asked to state whether or not they were going to apply for a job in the near future (1

= Yes, I plan to apply for a job in six months, 4 = No, I’m not planning on applying for a job in the near future). Age was also controlled based on the results described in the “Participant” section.

3.5 Image Manipulation Check Items

Following the definitions proposed by Keller (2000) and Goldberg and Hartwick (1990), ten items were constructed to assess whether or not the participants differed from one another regarding their perceptions of the described four corporate image dimensions. Each corporate image dimension was measured with a semantic differential scale with 6 points (please see Appendix 1 for details regarding 6-point semantic differential scale format and anchor). On a 6-point semantic differential scale, individuals rated the extent to which these adjectives described the manipulated company. The similar approach has been adopted in previous studies (e.g., Smidt, Pruyn, & Van Riel, 2001). The samples of semantic differential item are listed in Table 1.

Table 1. The Semantic Differential Scale for Corporate Image Manipulation Check Corporate image

dimensions Low High

values low-quality product values high-quality product Corporate

Product Image doesn’t value

innovative product values innovative product is low customer oriented is high customer oriented Corporate

Service Image is not employee focused is employee focused is not community concerned is community concerned

is not philanthropic is philanthropic Corporate

Citizenship

Image doesn’t value

environmental production values environmental protection is unreliable is reliable

is deceitful is trustworthy Corporate

Credibility

Image is amateurish is expertise

IV. RESULTS 4.1 Manipulation Check

To test whether or not the four corporate images were successfully manipulated, we conducted a series of t-tests. The results showed that the company possessing a high rating for corporate product image received significantly higher ratings from the participants (mean = 4.56, s.d. = .91) than did the company possessing a low rating for corporate product image (mean = 3.65, s.d. = .96; t[536]

= 10.07, p < .01). Second, the company possessing a high rating for corporate service image received significantly higher ratings from the participants (mean = 4.50, s.d. = .96) than did the company possessing a low rating for corporate service image (mean = 4.04, s.d. = .94; t[536] = 5.60, p < .01). Third, the company possessing a high rating for corporate citizenship image received significantly higher ratings from participants (mean = 4.68, s.d. = .83) than did the company possessing a low rating for corporate citizenship image (mean = 3.98, s.d. = .94; t[536] = 9.14, p < .01). Finally, the company possessing a high rating for corporate credibility image received significantly higher ratings from participants (mean = 4.50, s.d. = .81) than did the company possessing low ratings for corporate credibility image (mean = 4.13, s.d. = .87; t[536] = 4.42, p < .01). On the basis of these results, we deemed that the manipulations of corporate image variables had been successful.

4.2 Hypothesis Testing

The descriptive statistics and the correlations among variables are reported in Table 2.

A confirmatory factor analysis (CFA) using maximum likelihood estimation showed that the moderators and dependent variable did not fit well with the 4-factor model (chi-square = 908.96, df = 521, p < .01, chi-square/df = 1.74, CFI = .80,

NNFI = .78, and RMSR = .10). However, the fit of this 4-factor model was also compared with the fit of twelve alternative models (i.e., a null model, a 1-factor model, six different 2-factor models, and four 3-factor models). The fit of the current 4-factor model was significantly better than the fit of any alternative model, which indicates that the proposed 4-factor structure should be acceptable. The factor loadings ranged from .08 to .88 with an average loading of .50. This provided some evidence of convergent validity of the constructs (Bagozzi & Yi, 1988). Furthermore, we assessed discriminant validity by constraining inter-construct correlations in the measurement model to unity one at a time and by measuring the difference in the chi-square statistic (Anderson & Gerbing, 1988). The results show that all changes in chi-square (ranging from 134.84 to 596.51, Δdf = 1) were significant at the .01 level; hence, discriminant validity was achieved.

To test the hypotheses, we regressed organizational attractiveness on four corporate image dimensions and three moderators, after controlling for applicant age and willingness to apply for jobs. The results are presented in Table 3.

In model 2, four corporate image variables were entered into the regression. As shown in model 2, we found that corporate product image (β = .12, p < .01), corporate citizenship image (β = .08, p < .05), and corporate credibility image (β

= .08, p < .1) were positively associated with organizational attractiveness. However, the results showed that corporate service image was not significantly associated with organizational attractiveness (β = .04, p > .05). With respect to corporate product image (X1), we found that organizations possessing high ratings for corporate product image (X1) were more likely to attract the potential applicants (β = .13, p

< .01), providing support for Hypothesis 1. We also found that organizations

Table 2. Means, Standard Deviations, and Correlations for the Variables^a

Variables Mean s.d. 1 2 3 4 5 6

1. Organizational attractiveness 3.92 1.03 (.90)

2. Need for affiliation 4.95 .52 .10**

(.81)

3. Environmental sensitivity 4.67 .54 .12** .21** (.84)

4. Materialism 3.77 .71 .07 .12** -.11** (.79)

5. Applicant age 25.32 3.37 .01 -.00 .05 -.09*

- -

6. Willingness to apply 1.59 .76 .02 .07 .01 -.04 .23** - -

a n = 538. Alpha coefficients appear on the diagonal in parentheses.

* p < .05

** p < .01

possessing high ratings for corporate citizenship image (X3) were more likely to attract the potential applicants (β = .09, p < .05), providing support for Hypothesis 3.

As shown in model 3, corporate credibility image (X4) was positively associated with organizational attractiveness (β = .08, p = .06), providing marginal support for Hypothesis 4. Corporate service image (X2) was not significantly associated with organizational attractiveness (β = .04, p > .10), failing to support Hypothesis 2.

In model 3, four corporate image variables and three moderators—need for affiliation (Naff), environmental sensitivity (En), and materialism (Ma)—were added to the regression. As shown in Table 3, the standardized regression weight was significant for environmental sensitivity (β = .14, p < .01) and materialism (β = .08, p

< .1).

In model 4, we performed a moderated regression in which we added four possible two-way interaction terms about need for affiliation (i.e., X2 x Naff), environmental sensitivity (i.e., X3 x En), and materialism (i.e., X1 x Ma and X4 x Ma). To counter multicollinearity problems in our testing of the interaction terms, we centered all moderating variables before creating the interaction terms (Jaccard, Turrisi, & Wan, 1990). The interaction terms accounted for a significant amount of unique variability in organizational attractiveness (ΔR² = .05, p < .05). As shown in Table 3, the interaction of X2 x Naff was not significant in relation to organizational attractiveness, thereby failing to support Hypothesis 5. Furthermore, there were insignificant two-way interactions between X1 x Ma and X4 x Ma, thereby failing to support Hypothesis 7a and Hypothesis 7b. There was a significant two-way interaction between environmental sensitivity and corporate citizenship image (β

= .18, p < .01). To better understand the form of the moderating effect, we followed Aiken and West (1991) and Cohen and Cohen (1983) and drew an interaction-effect

plot. As shown in Figure 1, the positive relationship between corporate citizenship image and organizational attractiveness was stronger when applicants rated high on the environmental sensitivity scale, providing support for Hypothesis 6.

Figure 2. The interactive effect of environmental sensitivity and corporate citizenship image on organizational attractiveness.

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

Low High

Corporate Citizenship Image

Organizational Attractiveness

En (High) En (low)

Table 3. Results of Regression Analysis of Organizational Attractiveness^a

Variables Model 1 Model 2 Model 3 Model 4

Control

Age .00 -.02 -.02 -.02

Willingness to apply .02 .02 .03 .04

Corporate image

Corporate product image (X1) .12** .13** .14**

Corporate service image (X2) .04 .04 .05

Corporate citizenship image (X3) .08^* .09* .09*

Corporate credibility image (X4) .08⁺ .08⁺ .07

Moderators

Need for Affiliation (Naff) .06 .10⁺

Environmental sensitivity (En) .14** .00

Materialism (Ma) .08⁺ .02

Interactions

X2*Naff -.08

X3*En .18**

X1*Ma .07

X4*Ma .02

R² .00 .02 .06** .08**

△R² .03** .04** .05*

a n = 538.

+ p < .10

* p < .05

** p < .01

V. DISCUSIONS

5.1 Theoretical Implications

In the present paper, we examined the effects of corporate image dimensions on organizational attractiveness, specifically taking into account the effects of applicant individual differences. As expected, we found that corporate images relevant to product, social and environmental responsibility, and credibility are important determinants of organizational attractiveness, and this is generally true in the manufacturing industry contexts. Moreover, the results also showed that applicant environmental sensitivity moderates the relationship between corporate citizenship image and organizational attractiveness.

In the present study, we uncovered an interesting finding concerning the effect of corporate service image. Originally, corporate service image and corporate product image were treated separately in Keller’s (2000) model of corporate image. The results of the present study showed that corporate service image cannot predict organizational attractiveness. One possible reason is that individuals tend to embellish service-relevant performances when they have pre-established positive images toward the organizations (Martinez & Pina, 2005). The effect of corporate service image may be embedded in other corporate image dimensions and may, for this reason, have exhibited no unique influence on attractiveness.

Another possible explanation is rooted in the prior research on corporate communication and advertising. For example, research has shown that individuals considered prints and ads, placed by organizations, to be the least credible source of information (Cable & Yu, 2006). And according to Van Hoye and Lievens (2005), applicants place relatively little importance on the recruitment-information sources that organizations strongly influence. In contrast to other corporate images, service

image seems to involve perceptions of many regulatory behaviors and work-required performances that organizations control relatively easily. In the manufacturing industry, where most product production and customer service operations are separate, applicants may rather clearly discern information relevant to after-sales or maintenance services. While comparing multiple information sources in order to choose a future employer, applicants may naturally ignore information thought to be lacking in credibility. As a result, applicants may not consider corporate service image to be important to their assessments of organizational attractiveness.

In addition, our data did not support the hypotheses regarding the moderating effects of materialism. One possible explanation is that the applicants who rated high on the materialism scale may have perceived high quality and innovativeness (corporate product image) or trustworthiness and expertise (corporate credibility image) as being subtle symbolic attributes, not as being obvious evaluable attributes.

According to Lievens and Highhouse (2003), symbolic attribute to attractiveness reflects an individual’s self-identity or the individual’s desire to express deeply personal thoughts. Even though materialists may be inclined to assign symbolic meanings to these images, symbolic meanings do not guarantee the success of organizations. A similar rationale may apply to the insignificant finding in the moderating effect on corporate credibility image. Applicants who rate low on the materialism scale may also notice the influences of a strong corporate credibility image, whereas applicants who rate high on the materialism scale may appreciate the professional and reliable performance of the organization but not relate such performance to their material preferences.

5.2 Limitations

One limitation of the present research is that much of the data derived from

student participants. As the research focus was to examine how corporate image influenced potential applicants’ levels of attractiveness, the data should have reflected an authentic applicant population. While much past research has used the student sample to examine how various organizational characteristics affected job applicants (e.g., Cable & Judge, 1996), it has been suggested that more research should use real applicant samples or experienced workers in studying the effects of organizational characteristics on recruitment outcomes (Ployhart, 2006).While we made an effort to collect data from experienced workers in this research, we finally combined student samples and experienced employee samples owing to the relatively small sample size on employees. However, while combining the student participants and the experienced employee participants, we performed statistical analyses to ensure that there was no significant difference between the characteristics of each of the two groups.

Another limitation of the present research concerns our decision to use a relatively lenient standard (i.e., α = .10) when testing the main effects of corporate image in model 2 of the regression analysis. According to Murphy and Myors (2003), if we make it very difficult to reject the null hypothesis in order to minimize Type I errors, the possibility of the occurrence of Type II errors may increase at the same time. Considering the comparatively smaller effect sizes (d values ranged from .03~ .14¹) of the corporate image dimensions, we believed that the use of a more lenient alpha may help ensure the statistical power of the study.

Besides the limitations addressed above, one important concern should be noted ____________________________

1According to Cohen (1988), d values below .20 are considered to be relatively small effect sizes.

here. The current research focuses on addressing the important role of organizational characteristics in predicting organizational attractiveness. As a result, we neither considered nor controlled for job specific variables in the present studies. One may

here. The current research focuses on addressing the important role of organizational characteristics in predicting organizational attractiveness. As a result, we neither considered nor controlled for job specific variables in the present studies. One may