3 July 2018

Background

This data pertains to competency scores for a group of individuals in a certain organization

The scores are given by a group of assessors after observing the individual in a variety of simulated settings.

The competencies are

  • Anlytical thinking score : ATS
  • Planning and Organizing score : POS
  • Performance management score : PMS
  • Custodian of HO systems, processes and values score : HOS
  • Coaching and counselling score : CCS

Histogram of Scores

None of the Histograms looks visually similar. Hence we conclude that these are mostly different distributions

Difference in scores of competencies

Statistical tests to check for difference in scores

Since we have seen visual evidence of a difference in scores, we will check if the scores have any satistically significant difference

## 
##  One-way analysis of means (not assuming equal variances)
## 
## data:  value and variable
## F = 58.102, num df = 4.0, denom df = 354.8, p-value < 2.2e-16

We see that the p value is near zero. Hence we reject the Null hypothesis and infer that there is a statistically significant difference in scores within the average scores of different competencies

To evaluate which score is different, we would have to investigate further with a series of paired-t tests.

Score differences between zones (1/2)

Score differences between zones (2/2)

##         Delhi       Gujarat     Karnataka        Kerala   Maharashtra 
##            29             3            10             1            27 
##     Rajasthan     TamilNadu     Telangana Uttar Pradesh   West Bengal 
##             2            24            12             9            26

while the sample size is not large enough to be statistically relevant, we see that the zones for which sufficient observations exist do not have seem to have too much of a visual difference

Way forward (1/2)

With the current data, We can run some further tests to identify which is the exact competency in which to focus, depending on the score.

However, the current sample has a couple of limitations

  • Is limited to 1 organization limiting generalization for the industry
  • Data is not evenly spread across zones, though this might not be an error and could be representative
  • There are no demographic or work experience attributes available for the assesees making it difficult to build models

Way forward(2/2)

Given more data across industry, organization and geographies, the analysis could become more insightful providing answers to

  • What should the training roadmap of the company on a short term and long term basis, depending on the individual's level, geography and other demographic attributes
  • What should the promotion and succession planning ploicy and strategy of the organization be
  • Which should the recruitment strategy of the organization be in terms of previous experience, education and previous sector.