Study Of CSR with IRT AnaLytics SCORES
[Corporate Social Responsibility] [Item Response Theory]


This website hosts the Study Of Corporate Social Responsibility (CSR) with Item Response Theory (IRT) AnaLyitcs--a project initiated by Robert J. Carroll, David M. Primo, and Brian Kelleher Richter.  We plan to host here research, data, and underlying programs that represent innovations in the measurement of CSR using IRT tools.  The ultimate goal is to enhance our understanding of CSR and the implications of this growing area both in the business world and as a field of academic inquiry.  

D-SOCIAL-KLD Scores introduced in 2016 Strategic Management Journal Paper 

In our paper in the January 2016 issue of the Strategic Management Journal (SMJ) titled "Using Item Response Theory to Improve Measurement in Strategic Management Research: An Application to Corporate Social Responsibility," we introduced the idea of how IRT models could help us better understand and measure CSR.  In it we use the underlying or raw CSR data from the KLD STATS database (now owned by MSCI) to construct an improved firm-level measure of latent CSR.  Details on the methodology can be found in the paper linked below--and technical details can be found in its appendix linked below.  

|Pre-publication Version of 2016 SMJ Paper (PDF)|
|Appendix to Pre-publication Version of 2016 SMJ paper (PDF)|
|Strategic Management Journal Version (Link to publisher)|

D-SOCIAL-KLD Scores Dataset


We create a new CSR dataset in the 2016 SMJ paper and a measure which we term D-SOCIAL-KLD scores. In the acronym D-SOCIAL-KLD, (i) the “D” before the first dash indicates the approach is “Dynamic” in this instance, and (ii) the “KLD” after the second dash indicates the approach is applied to “KLD STATS data” in this instance. (Future applications of the approach might yield SOCIAL scores but have different modifiers indicating the methodology yields scores than have some feature of interest other than that they are dynamic or indicating that they have been applied to different underlying datasets. )

To enable other researchers to use our data, we have posted it below.  Please cite the 2016 SMJ article if you use our data.  We're also happy to hear about your work and link to it if you let us know about it by e-mailing us.   

|D-SOCIAL-KLD Scores Dataset (ZIP)|



SOCIALscores.org