Results

Forty-three percent of respondents occupied the positions 'project manager' and 'project coordinator'; approximately 30% occupied the positions 'product engineer', and 'sales manager'; and the remainder held executive positions like 'director' or 'quality engineer'. The most common type of project cited by respondents was 'new product development' (37%), followed by improvement projects for production/operations (16%). Concerning company size, more than half of the companies (about 58%) had more than 500 employees (large-sized firms), 25% were medium-sized firms; 12% were small-sized, and 4% were micro-sized firms.

The ANOVA test was used to verify the effect of the size of the company on the success of projects. As a result, a p-value = 0.11 (greater than 0.05) was obtained, which indicated that, on average, the agreement with successful projects in the perception of the respondents of the different companies is approximately similar. Therefore, company size has statistically no effect on project success response.

Table 1 shows Cronbach's alpha, composite reliability, and average variance extracted.

Table 1 Results of reliability and constructs validity tests.

Constructs Cronbach's Alpha Composite reliability Average variance extracted
Top Management Support (TM) 0.889 0.891 0.671
Communication (CO) 0.936 0.938 0.791
Change Management (CM) 0.907 0.912 0.723
Organizational Culture (OC) 0.903 0.902 0.699
Training (TR) 0.954 0.957 0.846
Project Success (PS) 0.845 0.843 0.522

As can be seen in Table 1, Cronbach's alpha values were above 0.7, which can be considered as an acceptable degree of construct reliability. The composite reliability values were above 0.6, thus also regarded as acceptable. The average variance extracted values were above 0.5, providing evidence of the construct's convergent validity. Table 2 shows the results of adjustment measures with the complete model.

Table 2 Adjustment measures.

Setting Value Reference Model
Chi-Square 454.623 -
Degrees of Freedom 260 -
Normed Chi-Square 1.749 < 2.00
Non-Normed Fit Index (NNFI) 0.875 > 0.90
Comparative Fit Index (CFI) 0.892 > 0.90
Standardized Root Mean Square Residual (SRMR) 0.068 < 0.08

Table 2 also showed that some of the adjustment measures of the model did not present reasonable values. Therefore, some variables were eliminated for model improvement.

The adjustment measures should adequately represent the full model. For this, the model was enhanced by eliminating variables with a high modification index (MI). The following variables had the highest modification indices and were thus excluded from the model: "the organization's vision, strategy and policy are openly discussed". (MI = 17.76); "top management provides enough resources to support the effort in project implementation". (MI = 13.28); "there is an open organization culture and trust in the organization". (MI = 12.63); and "there are appropriate guidance procedures in the company". (MI = 12.44).
The adjusted final model was satisfactory, since the modification indices stayed below 10 and the factor loadings were above 0.634. Table 3 shows the factor loadings results.

Table 3 Results of factorial loads.

Construct Variable Estimates of loads Squared Error z-Value Pr(>|z|)
Top Management Support TM1 0.806 0.100 8.037 9.189e-16
TM2 0.954 0.091 10.464 1.269e-25
TM4 0.789 0.101 7.791 6.626e-15
Communication CO1 0.925 0.091 10.194 2.106e-24
CO2 0.799 0.099 8.042 8.848e-16
CO3 0.917 0.091 10.037 1.043e-23
CO4 0.899 0.093 9.700 2.999e-22
Change Management CM1 0.837 0.098 8.572 1.013e-17
CM2 0.823 0.099 8.355 6.537e-17
CM3 0.929 0.091 10.173 2.617e-24
CM4 0.822 0.099 8.326 8.329e-17
Organizational Culture OC3 0.927 0.092 10.112 4.899e-24
OC4 0.908 0.093 9.771 1.499e-22
Training TR1 0.956 0.088 10.867 1.658e-27
TR2 0.948 0.089 10.701 1.011e-26
TR3 0.946 0.089 10.663 1.518e-26
Project Success PS1 0.657 0.109 6,027 1.669e-09
PS2 0.634 0.110 5.761 8.366e-09
PS3 0.913 0.094 9.693 3.226e-22
PS4 0.704 0.107 6.597 4.182e-11
PS5 0.845 0.098 8.582 9.290e-18

The hypotheses were tested to check the statistical significance of the constructs (Table 4).

Table 4 Hypotheses results.

Hypothesis Path Load estimates (β)  Squared
Error 		z-Value Pr(>|z|)  Significance
Level
H1 TM -> PS 0.599 0.087 6.859 6.957e-12 0.000
H2 CO -> PS 0.445 0.105 4.243 2.206e-05 0.000
H3 CM -> PS 0.603 0.087 6.976 3.046e-12 0.000
H4 OC -> PS 0.651 0.081 8.086 6.142e-16 0.000
H5 TR -> PS 0.504 0.097 5.209 1.894e-07 0.000


As can be seen in Table 4, project success is positively related with top management support (β = 0.599). Therefore, hypothesis 1 is supported. The results indicate that project success is positively related to communication (β = 0.445), thus hypothesis 2 is supported. The results also indicate positive relationships between project success and change management (β = 0.603), and, the strongest relationship, between project success and organizational culture (β = 0.651). Consequently, hypothesis 3 and hypothesis 4 are supported. According to the results, there is a positive relationship between project success and training (β = 0.504), therefore hypothesis 5 is also supported.

The final model including hypotheses and factor loadings is shown in Figure 1. All exogenous organizational constructs positively influence the success of different kinds of projects. The constructs OC, CM, and TM have the most statistical significance in this sample for project success in the context of the Brazilian automotive sector.

figure 1

Figure 1 Path diagram for organizational factors and project success.