PERMANENT POSITION
4.4 Hypothesis testing
4.4.2 Indirect effects
To further zoom in on the exact effects of the moderators, hierarchical regression was made for each effect and the variables were examined using PROCESS.
Starting with the regression. The dependent variable Behavioural Intention is used in the first model and the control variables are added to it. Then a second model is made with the
independent variables (PE, Tech and SN), as with the direct effects discussed earlier. Finally, the moderator is added as the third model and the interaction between the moderator and the dependent variable, for example, Transparency (Trans, PE_Trans, Tech, Trans and
SN_Trans).
After this, the moderation effects were checked using the PROCESS method of Andrew F.
Hayes (2013). Model 1 was used for this.
Moderating effect of Understanding AI
For the first model, we looked at the effect with the moderator Transparency. The third model is with the addition of the interaction between the independent variable and Transparency. This analysis shows that the second model has a significantly R² of 49%. The
40 third model has no significant change, the R² is 50%. The added interactions have no
significant effect, therefore no relationships can be established.
Also from the analysis with Process no significant relationship emerges (c3 = -.03, P > .05).
So there is no moderation effect.
It is concluded from this that Transparency does not play a significant role in the model and H4a, H5a and H6a all are rejected.
Table 7: Summary of Hierarchical Regression Analysis for Variables Transparency (N = 153)
Model 1 Model 2 Model 3
Variable B SE B 𝛽
Age .09 .10 .06
Gender .02 .15 .01
PE .53 .08 .49**
Tech .09 .06 .10
SN .26 .08 .24**
Trans -.10 .07 -.09
PE x Trans -.02 .07 -.02
Tech x Trans -.04 .05 -.05
SN x Trans .08 .08 .08
R²
.01 .49 .50
F for change in R² .76 45.36** 1.06
Note: Dependent Variable: BI
**. Regression is significant at the .01 level (2-tailed) *. Regression is significant at the .05 level (2-tailed)
Moderating effect of Explainability
For the second model, we looked at the effect of the moderator Explainability. This analysis shows that the second model has an R² of 49% and an F change of 45.36 at P < .01.
The third model has a small significant change, the R² is 56%. Explainability itself had a significant effect in the model (𝛽 =.20, P <.00). And a significant effect has been found for the interaction between Performance Expectancy and Explainability (𝛽 =-.16, P <.00).
41 This is a negative effect which indicates that the higher the Explainability, the lower the relationship between Performance Expectancy and Behavioural Intention.
A significant effect also emerges when the PROCESS method is used (c3 = -.11, P
<.01). This solution explains R² = 49.4% of the variance at P <.01). c1 = .22 tells us the expected difference in Behavioural Intention between two persons who differ 1 unit in the Performance Expectancy.
This leads to the conclusion not to support H5b and H6b but to support H4b.
Table 8: Summary of Hierarchical Regression Analysis for Variables Explainability (N = 153)
Model 1 Model 2 Model 3
Variable B SE B 𝛽
Age .70 .09 .04
Gender .01 .14 .00
PE .54 .07 .49**
Tech .05 .06 .05
SN .24 .08 .22**
Ex .19 .06 .20**
PE x Ex -.12 .05 -.16*
Tech x Ex -.07 .04 -.10
SN x Ex .02 .06 .03
R² .01 .49 .56
F for change in R² .76 45.36** 5.66**
Note: Dependent Variable: BI
**. Regression is significant at the .01 level (2-tailed) *. Regression is significant at the .05 level (2-tailed)
Moderating effect of job security (Position)
For the third moderator, we looked at the effect of Job Position. There are two options for this variable namely, whether to have a permanent contract or not. This analysis shows that the second model has an R² of 49%. The third model has no significant change, the R² is 50%. A significant effect has been found for the independent variable Technicality (𝛽 =- .30, P <
42 .05). However, the model is not supported by the data and therefore, no conclusion can be drawn from these data. Also, no significant moderating relationships were found between the independent variables and the Job Position. No significant relationship emerges from the PROCESS method either. With this, we reject H4c, H5c and H6c. There are no significant relationships.
Table 9: Summary of Hierarchical Regression Analysis for Variables Position (N = 153)
Model 1 Model 2 Model 3
Variable B SE B 𝛽
Age .07 .10 .05
Gender .02 .15 .01
PE .51 .17 .47*
Tech .27 .11 .30*
SN .08 .16 .07
Position .23 .17 .09
PE x Position .03 .19 .02
Tech x Position -.21 .11 -.20
SN x Position .21 .18 .17
R² .01 .49 .50
F for change in R² .76 45.36** 1.18
Note: Dependent Variable: BI
**. Regression is significant at the .01 level (2-tailed) *. Regression is significant at the .05 level (2-tailed)
Effect of all variables
For the fourth model, we looked at the effect of all variables. The third model has no significant change, the R² is 50%. A significant effect has been found for the independent variable Technicality (𝛽 =- .30, P < .05). However, the model is not supported by the data and therefore, no conclusion can be drawn from these data. The total result can be found in Table 12.
43 Table 10: Summary of Hierarchical Regression Analysis for all Variables (N = 153)
Model 1 Model 2 Model 3 Model 4
Variable B SE B 𝛽 B SE B 𝛽 B SE B 𝛽 B SE B 𝛽
Age .16 .13 .10 .12 .10 .07 -.00 .10 -.00 -.03 .10 -.02
Gender -.02 .20 -.01 .01 .15 .00 .02 .14 .02 .06 .15 .03
PE .54 .08 .50** .53 .07 .49* .51 .16 .48*
Tech .10 .06 .12 .03 .06 .04 .14 .11 .16
SN .26 .08 .24** .25 .07 .23** .07 .15 .07
Trans -.16 .07 -.14* -.14 .07 -.12*
Ex .23 .06 .24** .22 .06 .23**
Position .25 .02 .10 .27 .14 .11
PE x Trans -.05 .07 -.05
PE x Ex -.11 .05 -.16*
PE x Position .02 .18 .02
Tech x Trans -.03 .05 -.04
Tech x Ex -.06 .04 -.09
Tech x Position -.15 .13 -.15
SN x Trans .04 .08 .40
SN x Ex .05 .06 .08
SN x position .21 .18 .17
R² .01 .49 .55 .59
F for change in R²
.76 45.36** 7.26** 1.17
Note: Dependent Variable: BI
**. Regression is significant at the .01 level (2-tailed) *. Regression is significant at the .05 level (2-tailed)
None of the analyses with the control variable Age, Gender, Industry and Function found a significant relationship with the variables.
44 4.5 Summary of results
Below an overview 11 can be found in Table, with all the outcomes listed. Followed by a schematic representation of the significant outcomes in the tested model. Finally, Table 12 contains the conclusion per hypothesis.
Table 11: summary outcome regression and PROCESS analysis
Figure 7: overview significant results
Hypothesis Interaction effect (c3) P-value 𝛽-value
H1 - - .50**
H2 - - .12
H3 - - .24**
H4a -.03 .66 -.02
H4b
-.11 .01
-.16*
H4c .12 .51 .02
H5a -.06 .31 -.05
H5b -.00 .98 -.10
H5c -.17 .26 -.20
H6a .05 .48 .08
H6b -.07 .18 .03
H6c .26 .13 .17
45 Table 12: an overview of the results per hypothesis
Descriptive Number Hypothesis Result
Performance Expectancy
H1 The experience with AI as a consumer contributes to the behavioural intention to accept AI in a work environment.
Supported Subjective
Norms
H2 The complex technicality of AI plays a negative role in behavioural intention.
Supported Technicality H3 Subjective norms negatively influence the intentional behaviour of
accepting AI.
Not supported Transparency H4a The transparency of AI positively influences the performance
expectancy towards the behavioural intention to accept AI in a work environment.
Not supported
H5a The relationship between the technicality of AI towards the behavioural intention of using it in a work environment is moderated by the transparency of AI.
Not supported
H6a The transparency of AI moderates the relation between the subjective norms and the intentional behaviour towards accepting AI.
Not supported
Explainability H4b The explainability of AI positively influences the performance expectancy towards the behavioural intention to accept AI in a work environment.
Supported
H5b The relationship between the technicality of AI towards the behavioural intention of using it in a work environment is moderated by the explainability of AI.
Not supported
H6b The explainability of AI moderates the relation between the subjective norms and the intentional behaviour towards accepting AI.
Not supported
Position H4c The type of contract an employee has, has an impact on the relation between the performance expectancy and the behavioural intention to accept AI in a work environment.
Not supported
H5c The relationship between the technicality of AI towards the behavioural intention of using it in a work environment is moderated by the type of contract an employee has.
Not supported
H6c The type of contract an employee has moderated the relation between the subjective norms and the intentional behaviour towards accepting AI.
Not supported
46