FAQ

In the open-ended text question section, we utilize Machine Learning and Natural Language Processing (NLP). This technology is a subfield of Artificial Intelligence (AI) and deals with written texts from humans.

Closed-ended questions consist of scientifically standardized items and scales (6-point Likert scale) used to measure job-relevant personality traits. These closed-ended questions represent self-assessment. Additionally, Zortify uses Natural Language Processing (NLP) to obtain an objective external evaluation. The algorithm has been trained to reliably draw conclusions about the personality profile of the participants through the analysis of the open-ended text (6 questions at the end of the questionnaire).

This is done intentionally and is a common practice in all social sciences to identify biased response tendencies (participants always tick on the far left or far right of the scale). This also ensures that the candidate answers the test attentively and does not always choose the same answers.

The benchmarks show the profile of the participant and their average proficiency compared to the respective benchmark. With Zortify Select, you can choose from four benchmarks: Young Professionals, Talents, Executives, Top Executives. The selected benchmark is displayed as a line in the graph (spider chart).
The areas in the spider chart and percentiles, indicate the comparison to the overall normative sample.

There is no right or wrong interpretation for scores. The decisive factor is always the fit between the personality profile of the participants and the requirement profile of the job.
The scores represent tendencies in the participant’s behavior. It is important to note that the Big 5 personality traits are bipolar scales, and a lower percentile does not indicate something negative or a bad value, but simply shows a tendency towards the opposite pole of the scale.
The values shown are percentiles, i.e. percentile ranks within the cohort, and not raw values on a scale.

In general, moderate levels of self-centerdness , in combination with moderate to high levels of agreeableness (from the Big 5), are not a cause for concern. As we display percentiles rather than raw scores, moderate subclinical narcissism scores may appear alarming, but a percentile rank of 70 for example currently indicates moderate values.
Caution is advised when high scores in Strategic Manipulation are measured. People scoring high on this scale are typically very focused on their own advancement, power retention, and expansion. They often make ruthless decisions in their own favor and are not concerned with the impact on the company and its employees. High levels of impulsivity should also be seen as a warning sign.
Clients who use our products receive comprehensive training from our consultants and are given access to the Zortify network, where experiences are shared. In addition, consultants are available to answer questions about specific profiles and to advise our clients accordingly.
Personality tests are a source of information in the selection process. If the personality profile reveals potential risks (“red flags”), we recommend checking these in more detail (e.g. through specific questions in the behavioral interview, reference checks).

If the scores are above 70, it should be taken as an opportunity to thoroughly examine and question the candidate again, as this may be a sign of potential risk to the team and the company as a whole. The scores indicate possible tendencies. For example, a reference check would be useful, as well as targeted situational questions or case studies and presentations

The formulations are derived from the combination of the different traits (Big 5) & states (entrepreneurial capital), i.e. the respective personality profile. They are written very generic and indicate tendencies in behavior.
The number (ratio of opportunities vs. potential risks) depends on the respective personality profile and can vary.

The answers are analyzed by the AI to create the external assessment in the personality analysis. In addition, we display the answers at the end of the report as additional information for the recruiter. They can also be useful for the participant, in case they want to review their answer later. Often, the answers provide interesting starting points for the interview process as well.

We recommend being upfront and 100% transparent with candidates about the assessment process. This creates trust. Be sure to emphasize that the analysis is only one part of the recruiting process and that the final decision is always based on multiple factors.

The process can be set up individually according to the customer’s wishes. The report can be sent only to the recruiter, but also simultaneously to the participant(s). Our recommendation: First, send the report to the recruiter, then share the report with the candidate before the next interview. This way, both parties have the same starting point to discuss the results in the next interview.

Zortify’s USP is the combination of the closed-ended items (self-assessment) and the open-ended text questions (NLP). Large differences between self-assessment and NLP can indicate that someone is not very self-reflective (under- or overestimation) or has tried to influence the result in the self-assessment part. If this is the case, this can be explored further by asking specific questions to the candidate to better understand their approach to the assessment (e.g. behavioral questions, checking references).

We are currently checking whether it is possible to influence the outcome of the analysis using ChatGPT responses. However, we have over 1000 parameters running in the background. Therefore, it is very difficult to manipulate the result with ChatGPT as it does not know the parameters that underlie our personality analysis.

In addition to the participant’s own ranking, the NLP suggests a ranking of PPCs based on the participant’s personality analysis (external perception). The PPCs refer to the individual preferences of a person – regardless of their soft skills. The assessment by the AI can help to uncover hidden potentials.

Personal assessments may significantly differ from the NLP-based assessment as it is merely a matter of preferences. In general, we use this information to support teaming processes such as in project groups.

Numerous parameters and variables are incorporated into the evaluation through the nodes of the neural network (AI with over 1,000 nodes).
The focus is not on individual words or content, but always on the overall picture of content, syntax, and word choice.

Our algorithms are language-independent. Personal data (such as origin, gender, age) is not taken into account during the training of the AI in order to minimize bias. This ensures that no population group is overrepresented in comparison and avoids possible bias/discrimination.

Yes, Zortify Select helps decision-makers to make a more unbiased selection. It’s about the fit between personality and job profile. For example, a third person could make a pre-selection based on anonymized analysis results.

It is important that the questions are answered intuitively (see also video with instructions at the beginning of the analysis). The candidate should not think about it for too long. There is always an interaction between the person and the situation. Sometimes we behave a little differently in our professional and private lives or even in different professional situations. Nonetheless, personality is relatively stable in the long run and when people identify themselves. Therefore, it is rather unlikely that the individual is a completely different person at home than at work. If participants cannot answer a question intuitively, it is advisable to answer the question in a professional context.

The data used for self-assessment and open text questions that we have used to train our algorithms is based on a diverse group of participants that differ in profession, employment status, language, and gender. Moreover, the EU will soon publish an AI regulation that specifies that AI should be trained with bias-free and neutral datasets. Our goal is not only to meet this new standard but also to continuously develop our technology through our own research.

The participants in our analysis so far are mainly individuals living in Western Europe.

We currently do not differentiate between corporations, SMEs, or startups. The use of Zortify Select is independent of the size of the company as well as the position (junior, senior, management).

We want to emphasize that Zortify’s personality analysis combines traditional personality assessments with AI-based personality assessments. Our analysis is based on proven personality models, the validity and reliability of which have been demonstrated through decades of research. Zortify also conducts its own validation studies to ensure the scientific foundation of the assessment. Additionally, the assessment was developed by a highly qualified team consisting of data scientists with PhDs in the field of NLP and psychology, among others. We work to the highest ethical standards and have very high expectations of our work. Moreover, we are an accredited research institution and are authorized to conduct independent research and produce expert reports.