In my opinion, regardless of the levels of the variables, you can run regression analysis and test your hypotheses. For the following steps, no need to define variables class wise. Keep the data file simple.

I am replying point wise:

For Action point 1: Test these regressions:

(i) Humble Relationships (H) --> Creativity (C)

(ii)

Psychological Safety (Mediating Variable)- P --> Creativity (C)

(iii) P + H --> C (Since P is a mediating variable, P+H is needed. For moderating, it should be a multiplication, as explained below)

check for the "R2 changed" for (ii) and (iii) along with its "F statistics". which will show you whether the mediating variable has a significant effect on the relationship between H and C

For Action point 2: Run these regressions:

(i) H --> C

(ii) H*S --> C

Since S is a moderating variable, no direct link between S and C, hence no regression is needed.

Again, check R squared change and its significance, based on which you will be able to deduce whether Knowledge Sharing really works as a moderating variable.

For Action point 3: Not sure whether Monte Carlo would fit here

For Action point 4: If you find all three (H, P and S) are correlated with each other, then it would be difficult to analyse mediating and moderating effects. In that case, check how to reduce the multicollinearity.

for Action point 5: Not sure whether CFA would fit here.

Just a curiosity: Are you going to test the whole model in AMOS? If possible, do it. It may show you more meaningful results.

------------------------------

Vandit Desai

------------------------------

Original Message:

Sent: Thu August 12, 2021 12:01 PM

From: Brendan Meany

Subject: Organizing Data for Multilevel Path Analysis

Good Morning-

I am preparing data for a multilevel path analysis. I am stuck trying to organize my data into levels on the individual and team level. Should I create a separate category to denote the variable as a Level 1 and Level 2 or is that done on the backend (variable view) of SPSS?

Your help would be greatly appreciated with this matter!

Cheers,

Dr. Meany

Below are the details of the variables, levels, and actions required:

**Purpose of the Study**

The purpose of this study is to explore the positively mediating role of psychological on the relationship between humble leadership and follower creativity, and how knowledge sharing positively moderates psychological safety and creativity in a less high-power distance and less collectivistic culture.

Variable Levels:

Level 1 (Individual Level) -

Gender- G

Age- A

Tenure- T

Creativity (Dependent Variable)- C

Level 2 (Team Level) -

Team Size- TS

Humble Leadership (Independent Variable)- H

Psychological Safety (Mediating Variable)- P

Knowledge Sharing (Moderating Variable)- S

Actions:

1. Depict relationships between team-level variables (Level 2) and individual-level variable (Level 1), through a multilevel analysis- multilevel path analysis to test this multilevel mediation hypothesis.

2. Additionally a simultaneous multilevel regression procedure within multilevel path analysis to determine if knowledge sharing indeed moderates the relationship between leader humility and follower creativity as mediated by psychological safety

3. A Monte Carlo simulation procedure using open-source software R will test the significance of the influences of mediation and moderated mediation

4. Within-group agreement (depicted as rWG) and reliability (intraclass correlations ICC1 and ICC2) will need to be tested to determine if the aggregation of humble leadership, psychological safety, and knowledge sharing as appropriate as team variables

5. Confirmatory factor analyses (CFAs) to examine discriminant validity.

------------------------------

Brendan Meany

------------------------------