Hello everyone,

I'm a student and I'm currently having some issues with coding and generating the variables for my research. My research is about the moderating effect of CEO-TMT characteristics on the relationship between TMT faultlines and R&D spending.

I have collected the following data on TMT:

- exective age
- executive gender
- executive tenure
- executive tenure on the top management team
- type of degree of executive

I would like to generate the following variables:

CEO-TMT sociodemographic similarity:

This variable is a composite of a CEO’s similarity to the rest of the TMT in terms of two externally observable socio-demographic attributes: age and gender. These attributes reflect the underlying social-identities of individual executives (Crucke and Knockaert, 2016; Georgakakis and Ruigrok, 2017; Milliken and Martins, 1996; van Knippenberg et al., 2004; Veltrop et al., 2015) and have been used in the extant literature as predictors of similarity attraction inclinations (Bezrukova et al., 2009)

As age is a continuous variable, age similarity between CEO and the TMT was calculated using the distance formula:


Array

(Westphal and Zajac, 1995).


In this formula, Xi represents the age of the CEO I, Xj represents the age of each non-CEO Executive j, and n is the number of TMT members. To turn this into an age similarity measure, we reversed the variable so that higher scores indicate higher similarity. Further, CEO-TMT gender similarity was measured using a modified version of Blau’s (1977) formula expressed as Pi^2, where P is the proportion of TMT members that share the same gender category i with the CEO. To calculate the overall CEO-TMT similarity measure, we rescaled age similarity to take values between 0 and 1, and then aggregated the two components in a composite variable.

CEO-TMT shared experience:

This variable was measured using the procedure by Caroll and Harrison (1998), which averages pairwise overlap of team tenure (TLAP) for all possible pairs in the team. Carroll and Harrison’s (1998) TLAP formula is expressed as:


Array

Where u represents the TMT tenure (in years) of each individual i. Since we are interested in the tenure overlap of the CEO with other TMT members, we measured the average pairwise tenure overlap between the CEO and other senior executives. To account for the diminishing effect of team members’ interaction over time we transformed this variable by taking the natural logarithm (Buyl et al., 2011).


Faultlines:

To define this faultline measure, consider a group containing a total of n members who are measured on p characteristics. A faultline can split this group into two subgroups in a total of S = 2n–1 – 1 ways. For each of these possible subgroupings we measure the percent of total variation in overall group characteristics accounted for by the variation between subgroups by calculating the ratio of the between group sum of squares to the total sum of squares:



Array


where xijk denotes the value of the j th characteristic of the i th member of subgroup k, x - . j . denotes the overall group mean of characteristic j, x - . jk denotes the mean of characteristic j in subgroup k, and ng k denotes the number of members of the kth subgroup (k = 1, 2) under split g (more details on how to incorporate categorical variables by using dummy variables in this calculation are described later in this section). The faultline strength, Fau, is then calculated as the maximum value of Faug over all possible splits g = 1,2, ... S. For small group sizes it is possible to calculate Fau by enumerating all possible splits and calculating the maximum Faug , g = 1, 2, ... S.



For any helpful hints on how to efficiently generate these variables, I would be very grateful.

Thank you in advance.

Dion Sinack