AI-supported analyses following an employee survey: Making motivation, emotions, and skills visible

Using an AI-supported process, Cubia can measure specific characteristics relating to motivation, emotions, and skills from text responses to open questions, in addition to classic analyses such as clusters or summaries. Based on renowned research results, the new process enables a much deeper insight into the psychological structures of an organization or parts thereof. The process was developed in collaboration with Prof. David Scheffer (NORDAKADEMIE) and is hosted entirely by Cubia AG. This means that no data is transferred to third parties or cloud solutions.

What was previously hidden: AI-based text analysis provides new insights

For almost 100 years, scientists have been working on measuring people’s motives, emotions, and competencies with the help of content analyses of written or spoken texts. The research work and its transfer into practice by Harvard professor David McClelland was groundbreaking in this field. In the past, these texts were evaluated by trained experts according to specific criteria. Scientific publications and meta-analyses prove the astonishing validity and prognostic reliability of this method. For example, David McClelland and his colleagues were able to show that leadership success could be predicted with statistical significance 16 years in advance. Today, thanks to AI-supported methods, evaluations can be carried out automatically and thus scaled enormously.

The written or spoken texts by employees of an organization are stimulated by guiding questions.

Cubia AG and Prof. David Scheffer have been working with this technology for more than 20 years. They developed an early form of automated evaluation. Since then, David Scheffer at NORDAKADEMIE, University of Applied Sciences, has continued to refine this technology with the help of machine learning and neural networks and expanded it to cover more and more skills. This method has been validated in an assessment center at NORDAKADEMIE for 20 years. There, competencies such as focus on results, creativity and innovation, as well as communication and cooperation skills are observed by trained HR managers from partner companies and serve as validation criteria for the automatic text analysis method.

The method has already been tested in practice in numerous customer projects with companies such as AIRBUS, BCG, EDEKA, and Hamburger Hochbahn. The method has also been used in NORDAKADEMIE’s selection test for many years. The CAPTA Institute (Computer Aided Psychometric Text Analysis) ensures the scientific validity of the method with several doctoral students.

Theoretically, all open-ended questions in surveys can be evaluated automatically using this AI-based method. However, since the neural network has been trained on the four questions mentioned above, these should be included if possible.

What characteristics can be measured by motivation analysis?

The “big three” implicit motives can be measured: the need for commitment/security, the need for quality standards/excellence, and the need for power/autonomy.

Employees who give commitment-motivated answers in a survey strive for harmony and cohesion. They avoid competitive situations and criticism. This aspiration is very valuable in private life and also in some professions. In most professional roles, however, especially those in tough, competitive environments, a high level of attachment motivation is a disadvantage and should be changed if possible. This is especially true at the management level.

However, the average value for the attachment motive in an organization should not fall too low. Good, trusting relationships with colleagues and superiors are important! If the average value falls below 10%, this is alarming.

People who report high values for performance motivation in their responses strive for excellence. They set themselves challenging but achievable performance goals and are guided by quality standards. They are ambitious and want to learn, but sometimes this drive comes at the expense of social skills. Extreme average values in an organization or team indicate a culture of “lone wolves.”

If many employees give highly power-motivated answers to the open questions, then the organization or team has a particularly strong desire to shape things and a lot of “power.”

They want to make a difference, change circumstances, influence people, and develop. To this end, they are willing to work very hard and achieve a lot. They want to be independent and rise to the top. However, caution is advised with extreme average values: situations are then interpreted too one-sidedly as competition in which only the superior wins. Such a team then has similarities with a shark tank.

Which emotions can be measured?

Intrinsic motivation and enthusiasm are a particularly favorable implementation style for the three motives. People who answer the open questions in this style rely on their optimism and positive energy. This implementation style gives all three motives a charming, sometimes even charismatic touch.

Extrinsic motivation and impact orientation, on the other hand, are more emotionally restrained. People who demonstrate this style when answering the open questions focus on correctness and a certain degree of role conformity.

Switching between positive and negative emotions is an important prerequisite for persistence. Such answers to the open questions are helpful in very difficult situations and transformation processes. Higher average scores indicate that employees and managers are always looking for new solutions.

Emotional pressure and task orientation in the answers to the open-ended questions should be viewed with caution. They indicate a culture in which mistakes must be avoided at all costs. Such employees want to work efficiently even under great pressure, but they tend to conceal mistakes and mistrust each other.

Negative emotions and caution in the responses clearly indicate bad experiences in the work situation. If fear, feelings of powerlessness, anger, and annoyance dominate the responses to the open questions, then there is definitely a need for action.

Which skills can be measured?

Skills cannot be measured at the individual level, but only at the group level. At least six employees must be aggregated in order to draw conclusions about skills within the team.

In psychology, skills that indicate resilience are considered particularly important. To do this, the switch between different motivational and emotional systems must be demonstrated at the team level.

In this way, skills such as self-regulation, focus, agility, willingness to learn, and openness to new experiences can also be measured.

What are the implications of the motivation analysis?

David McClelland has shown that such open-ended questions and so-called operant responses are often more meaningful than closed questions. A detailed analysis of the responses with the support of AI provides very deep insights into the inner workings of an organization. Broken down into departments and teams, detailed recommendations for action can be developed for managers. If, for example, a transformation process elicits predominantly power-related and emotionally negative responses that indicate low resilience, then the change phases in which leadership coalitions are formed and “quick wins” are achieved for employees (see Kotter’s change model) must be critically reviewed again to ensure they are being fulfilled. Predominantly emotionally positive, commitment-related responses, on the other hand, may indicate a culture of comfortable inertia that must be broken down for change to be successful, for example, by communicating the urgency according to Kotter.