Could Monitoring of Emotions in Electronic Communications Be Used to Predict the Well-Being of Organizations?

By Carrie Zapka, SIOP Blogger 

Social Scientists Wade Into the Tweet Stream, an article in the September 30, 2011 issue of Science, is a stimulating read (subscription required. Republished by University of Vermont here). The author shared recent studies that mined Twitter data for insights into human behavior.

Peter Dodds and colleagues from the University of Vermont, for example, applied a novel method to analyze years of tweets from around the world. The Twitter data suggest a global decay in mood, a decrease in “average happiness” level, which they determined by comparing the relative usage of common words associated with positive or negative moods. Their results indicate that the world is not as happy today as it was in 2009!

The focus of the Science article was on the potential for these approaches to help us study global trends and communication habits. I wonder, though, is anyone trying these approaches at the organizational level? What if we applied the same algorithm that the University of Vermont researchers used to study global Twitter posts to monitor the language used by members in an organization? In a company setting we would have the advantage of access to more information than these researchers had. In addition to the open social media content we would also be able to monitor employees’ email, texts, and other electronic documents. But, at what point would constant monitoring of all employee communications cross a legal or ethical line?

Perhaps it is too early to tell how useful these innovative methods are going to be for I-O psychologists, but it sure is exciting to think about the possibilities.

What might differences in the average happiness levels between workgroups within one organization indicate? What would it mean if one organization had a higher level than another? If we monitored several companies for years would we discover correlations with standard I-O measures of organizational well-being? Could we learn to predict a future dip in customer satisfaction, turnover rates or financial success by observing a preceding dip in the collective employee mood?

Please reply. Share what you think about the Science article, the University of Vermont study, or the potential I-O applications of monitoring electronic communications within organizations for emotions.

Comments

[Josh Denton]

Hi Carrie,

Thanks for the reply. I wonder if knowing the day-to-day mood changes of employees is worth anything to supervisors. Foremost, supervisors are traditionally poor at being therapists and handling the fluctuation of subordinates' daily interferences.

It is tough to extract sarcasm; plus, it is impossible to determine tone and infliction in emails. Perhaps these can be mitigated in data collection.

As for research, I think it is a great idea. Perhaps something will come of it later.

Josh

[Nuts4ideas]

Hi Josh,
You make some great points about the limitations and challenges in this approach. Any method of data gathering has limitations, though, so we just need to study it and identify what they are. Then later we can consider the results of this new method of data gathering in context with other source of information, like we do today with many methods in I-O and HRM.
You asked; “what is your opinion of organizational use to electronically monitor employee well-being?” If it could be done in an ethical and legal way, then I can imagine many great uses. What if an algorithm was developed that monitored employees’ “mood” or “attitude” (as expressed in communications) over time. Then imagine that research studies revealed that the results correlated well with future performance as documented in standard performance management methods that we use and trust today. If that were true, then electronic communication monitoring may have an advantage to predict future performance and allow for intervention sooner (training, greater supervisor involvement/support, etc.). Much like how statistical process control works in quality monitoring in manufacturing, ongoing monitoring relative to normal baselines may help supervisors to know when something is an abberations (a bad day for example) vs. a statistically significant trend that needs attention. Supervisors could be alerted if an employee is trending up or down from their normal baseline attitude- automatically by the software.
I know many people think this is too “big brother”, it seems likely to me that in 1-2 generations people will no longer view constant monitoring with the negativity and fear that they do today. We are just not used to it yet. When the kids who have less need for privacy of today become the bosses of tomorrow, they are not going to view this type of monitoring in the same way that management today sees it.
Carrie

[h Schneider]

Just knowing that your communication was being monitored for emotion might cause you to 1. Fake happiness (most people who value their jobs would do this) 2. Avoid all emotionally revealing language or 3. Intentionally load your language with negativity (the disgruntled rebels).
Have you heard of "smile and move " initiatives? I think squelching negativity has positive effects even if it's not sincerely embraced by all.If the disgruntled few are encouraged to communicate professionally and to focus on solutions then all benefit. I'm interested in the impact of the "toxic "coworker on morale versus the "pollyanna ".

[Josh Denton]

Part of the problem will be limited to those who use the technology and the segment of that population that express pleasure or displeasure with some organizational issue(s). I wonder if too much reliance on a sample of the organization would sway management to implement strategies when there is not a problem to begin with. This also assumes that management finds improving well-being to provide positive ROI.

What is your opinion of organizational use to electronically monitor employee well-being?

Thanks for the post.

Josh