In my article Typical Results posted on May 8, 2018, I referenced an important meta-analytic study for the assessment and survey industry. Today, I’d like to draw attention to information in a table within the published article. The subtitle of Table 1 on page 434 (PDF page 4) reads: Examples of Variable Types Used to Classify 147,328 Correlational Effect Size Estimates Reported in Journal of Applied Psychology andPersonnel Psychology, 1980-2010.
The spotlight I am shining on this table and figure number two in the article is to demonstrate the almost unlimited types of variables that can be studied and end up rendering predictive value. Figure 2 in the study is an abbreviated hierarchical variable taxonomy the authors used to classify 30 years of variables. It references 4,869 variable nodes.
When I downloaded the original data in spreadsheet format, there were 207,330 rows of information. In the column titled Variable type I sorted alphabetically and removed all duplicate rows. That left 22,005 rows with unique variable names. Quite a large pool of ideas for research…
What type of data does your company or organization capture?
1. People Attitudes: Supervisor satisfaction, Coworker satisfaction, Leader-member exchange
2. Job Attitudes: Job satisfaction, Autonomy perceptions, Pay satisfaction
3. Organizational Attitudes: Organizational commitment, Perceived organization support, Procedural justice
4. Intentions: Turnover intention, Intention to accept a job offer, Intent to participate in development
5. Behavior: Performance, Absenteeism, Turnover
6. Knowledge, Skills, and Abilities (KSAs): Job Knowledge, Decision-making skills, General mental ability
7. Psychological Characteristics: Traits (Conscientiousness, Core self-evaluation), States (Stress, Burnout)
8. Objective Person Characteristics: Age, Gender, Turnover
9. Movement: Voluntary turnover, Job choice, Involuntary turnover
Scanning through the relationships, I see interesting positive correlations (from 0.0 to 1.0 signifying that when one variable increases, so too does the other). There are thought-provoking negative correlations also (from 0.0 to -1.0 meaning that when one variable increases, the other decreases). Some examples:
· Holland SDS Scale: Enterprising (E) & Holland SDS Scale: Social (S) [correlation: 0.79]
· Alcohol during workday-frequency & Alcohol during workday-quantity [correlation: 0.78]
· O*NET descriptor: Inductive Reasoning & O*NET descriptor: Interpreting the meaning of information for others [correlation: 0.78]
· O*NET descriptor: Deductive Reasoning & O*NET descriptor: Organizing, planning, and prioritizing work [correlation: 0.78]
· Years in work force & Company tenure [correlation: 0.68]
· Team learning & Team performing [correlation: 0.56]
· Performance & Jovial [correlation: 0.55]
· Overall justice & Satisfaction [correlation: 0.54]
· Neuroticism & Self-Esteem [correlation: -0.78, negative relationship]
· Process measures: Team conflict & Process measures: Workload sharing [correlation: -0.78, negative relationship]
· Satisfaction: Subordinate & Intent to Quit [correlation: -0.69, negative relationship]
· Role conflict & Role ambiguity [correlation: -0.56, negative relationship]
The question arises: What aspect of your work environment do you and your colleagues believe can make your employees more satisfied, committed to staying, being productive, and influencing everyone around them in a positive way? Partner with a data scientist and discover if your hypothesis is a golden nugget for future success.
