Section B. There are two questions included in this section. Please answer all of them.

Question 3.

You have conducted a research project to investigate the determinants of global knowledge sourcing of overseas R&D subsidiaries (Dependent variable). Three factors are considered:

   Vertical administrative embeddedness (measured by joint involvement of HQ and subsidiaries

in key administrative decision-making),

   Vertical  knowledge  embeddedness  (measured  by  joint  research  projects,  transfer  of

technology/knowledge, and movement of researchers and engineers between overseas R&D subsidiaries and MNC HQs),

    Geographic  distance  (measured  by  the  distance  between  the  cities  in  which  overseas

subsidiary and HQs were located).

Four hypotheses have been developed and tested based on a sample of 497 Japanese firms:

Hypothesis 1 (H1). Vertical administrative embeddedness (VAE) negatively affects global knowledge sourcing (GKS) by overseas subsidiaries.

Hypothesis  2  (H2).  Vertical knowledge  embeddedness  (VKE) positively affects  global knowledge sourcing (GKS) by overseas subsidiaries.

Hypothesis 3a (H3a).  Geographic distance weakens the negative relationship between vertical  administrative  embeddedness  (VAE)  and global knowledge  sourcing  (GKS)  of overseas subsidiaries.

Hypothesis 3b  (H3b).  Geographic distance  weakens the positive relationship between vertical knowledge embeddedness (VKE) and global knowledge sourcing (GKS) of overseas subsidiaries.

Also, two control variables are included in the project.

    Citation (the total number of citations made by a patent)

    Inventor (the number of inventors)

The linear regression outcomes are presented in Table 1 (next page).

Table 1 Results

Note. Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001.

You are required to answer the following four questions:

a)  Write down the statistical model for Model 3.

[3 marks]

b)  Suggest whether the results in Table 1 are consistent with any of the four hypotheses. Please support your answer with the  detailed  analysis  (in terms  of the  direction  and  statistical significance of the key variables).

[7 marks]

c)  Define R-square and the adjusted R-square values and interpret those values in the three models.

[7 marks]

d)  The managers want to know how the results of your project can help them increase their global knowledge sourcing of overseas R&D subsidiaries. How would you respond?

[5 marks]

Question 4:

A research project is conducted to examine the impact of perceived labour regulation obstacle (“LaborReg”) on labour productivity growth (“LPgrowth”) of Emerging Market Enterprises (EMEs). The labour regulation obstacle is measured by a 5-point Likert scale survey question, where 5 suggests the highest level of perceived obstacle.

The  researchers consider three moderators that may affect the  relationship between labour regulation obstacle and labour productivity growth:

    Formal training (“Training”, measured by the years of formal training received by employees),     Education (“Education”, measured by the average years of education of employees),                 Top Management Experience (“TMExperience”, measured by the years of top managers’

experience in the sector).

Two control variables are identified:

   Whether the firm was formally registered when it began operation (“Registration”, measured

by a binary variable, where 1 = Yes and 0 = No),

   Whether the firm competes against unregistered or informal firms (“InfCompetition”, measured

by a binary variable, where 1 = Yes and 0 = No).

The researchers conduct a linear regression based on 7524 EMEs. The results are illustrated as follows.

Table 2 Descriptive statistics

Table 3 Correlation matrix

Table 4 Model Summary

Table 5 Coefficients

where

      LaborReg_Training is the interaction between LaborReg and Training (i.e., LaboTReg × TTaining);

      LaborReg_Education is the interaction between LaborReg and Education (i.e., LaboTReg × Education);

      LaborReg_TMExperience is the interaction between LaborReg and TMExperience (i.e., LaboTReg × TMExpeTience).

Figure 1 Histogram of residuals

Figure 2 Scatter plot of residuals

Figure 3 P-P plot

Based on the given information, you are required to answer the following question:

a)  Evaluate the use of linear regression model by applying the five linear regression assumptions. And what managerial implications can we get from the results?

[34 marks]

The researchers further explore this project by categorizing the dependent variable (i.e., Labour productivity growth) into a binary variable, where LPgrowth=1 if its original value is positive and LPgrowth=0 otherwise. Accordingly, a logistic regression (i.e., logit model) is applied. The SPSS outputs are as follows.