The stability of Talentoday's mYti© tool is essential to obtaining reliable, consistent measurements. We have therefore taken measures to confirm the questionnaire's validity, accuracy and sensitivity.
The questionnaire results were calibrated using a large sample, which allowed us to standardize and ensure the psychometric validity of these results. Thanks to the quality of our database, the most recent calibration was conducted with a sample of 2,350,430 individuals.
We also calibrate to take into account differences due to cultural biases and due to the language in which the questionnaire was taken. We analyzed the trends by taking into account the respective cultures of each country in order to standardize and adapt their respective translations.
Internal consistency evaluates the accuracy and reliability of a test, which is measured by Cronbach's alpha (1951) See: Cronbach's Alpha. This assessment tool is a calculation method for the covariance of the items and questions for each criterion, allowing us to measure their homogeneity. The higher the covariance, the more the obtained score corresponds to the measured data.
Cronbach's alpha is calculated using the following:
is the number of questions (items)
is the variance of the question's variance of a personality or motivation trait
is the variance of the question (item), i
The American Psychological Association recommends alpha greater or equal than 0.7. All the items in Talentoday's mYti© questionnaire have a Cronbach's alpha greater than 0.7.
The average alpha for each item in the questionnaire is 0.74.
The questionnaire's stability is measured by the consistency of the variance between an individual's results the first time t1 and the second time t2 taking the test. This test-retest method makes it possible to calculate a correlation coefficient r between the two assessments, t1 and t2. According to the American Psychological Association's recommendations, if this coefficient is high enough (r greater than 0.6), that means that the results were stable over the course of the subsequent six-month period, as defined by the personality and motivation measurements.
Note that after six months, certain criteria may change.
Test-retest analysis methodology:
- Administer a test to a sample of users.
- Re-administer the same test to the same group, six months later.
- Calculate the correlation coefficients of the collected results.
We confirmed that the 28 mYti© criteria all presented a correlation greater than 0.6 in the retest at six months.
A test is valid if it measures what it is supposed to measure. This validity defines the meaning of the scores: they reflect what they are supposed to measure, no more no less. Thanks to this validity, we can draw specific conclusions about an individual based on their score - in this case their personality and motivations.
A predictive validity study compares the results from the psychometric assessment tool and an external assessment index. We therefore calculated the correlation between the mYti© score and a second score, which evaluates the same measurements using a different method.
Predictive validity analysis methodology:
- Administer the mYti© questionnaire to a sample of subjects.
- Next, ask the managers of these subjects to assess them using a second questionnaire that evaluates the same measurements.
- Finally, ask each subject to conduct a self-assessment using the second questionnaire.
After analyzing and comparing the results, the relationship between the mYti© scores, the managers' assessment scores and the self-assessment scores indicates a high degree of predictive validity.
Representativity of the dataset
Talentoday has a pool of over three million assessed candidates. Users are of all ages and categories: adolescents, students, employees, job seekers, and retirees. 160 nationalities are represented in this data.
People have taken the Talentoday questionnaire from from all over the world: America, Europe, Asia, Oceania, and Africa.
The sample meets representative criteria: 142 companies with over 200 employees (Apple, IBM, Walmart, etc.) and 402 universities with over 1,000 students (University of California, Berkeley, National University of Singapore, etc.). This diversity allows us to apply individual analysis to an extensive pool of candidates - the starting point for large-scale psychometrics.
Comparison with other assessment tools
Myers-Briggs Type Indicator (MBTI)
Launched in 1962, the Myers-Briggs Type Indicator (MBTI) is an assessment tool based on work of Carl Jung that measures psychological preferences in how people perceive the world and make decisions.The test consists of a series of questions that measure preferences on the dimensions of Introversion - Extroversion, Sensing - Intuition, Feeling - Thinking, and Judging - Perceiving. How you score on those answers fits you in to one of 16 different personality types. Despite its popularity, this test is not suited for hiring assessment and the The National Academy of Science review committee concluded that there are not sufficient research to justify its use in career counseling.
Launched in 1928, the DISC model is similar but simpler than the Myers Briggs. It measures only two dimensions of personality.
Big Five and Gallup's StrengthFinder
The IPIP Big-Five Factor Markers come from the statistical study of responses to personality items using a factor analysis. It identifies the core five personality traits from unlimited personality variables: extraversion, neuroticism, agreeableness, conscientiousness and openness to experience. Academic psychology commonly uses this model of personality assessment. Tests based on the Big Five such as the Gallup StrenghthFinder have in general a better scientific validation and are in compliance with International Test Commission standards.