Is the OBA a bullshit analysis?

Honestly, we don't believe in online assessments

Most assessments try to pin you down by asking your opinion on multiple subjects.

This method is the main cause of their cruel lack of reliability, for 3 reasons:

1. Your opinion on a topic can change over time and be influenced by outside factors:
   "After a day of work, do you prefer a night out with friends or to stay at home to read a book?"
   At this very moment I am tired so I would rather go home than go party > the answer varies depending on exogenous factors
2. ‍It's quite easy to guess what is analyzed behind each question, which creates biases in the analysis results
   ‍"You often lose your things"
   They are trying to determine whether or not I am organized > it's a job interview, I better say ***
3. Your opinion on a subject does not reflect the way you operate. Two people who function in opposite ways can share the same opinion about some things, but not through the same thought process.
   ‍"Would you rather have a child who's nice but stupid, or a child who's mean but intelligent" > this question is trying to determine if you are using rational or emotional thinking
   ‍
   Person A: I’d rather have a nice child because the most important thing is to be nice to others - everyone should
   Person B: I’d rather a nice child because the most important thing is to succeed in life, and in order to do that building relationships is more important than pure intelligence
   
   Person A and Person B have picked the same answer: having a nice child. Yet, their thought processes leading to this decision were different. Most assessments aren't able to spot those differences, which leads to biased results.

For these reasons, the OBA is not based on direct questions asking for your opinion, but on stimuli words that test your brain's responsiveness.

Have you ever seen questions as weird as the ones in the OBA?

The brain is made up of 18 main areas, each one responsible for both our strengths and limitations at the same time.

Each area is sensitive to certain words and concepts that the OBA uses in its questions. Your answers are measured to determine how strongly you are reacting to each of these. This enables us to identify how your brain has been evolving since your birth, as well as what areas you will develop next.

The OBA is so powerful that for 2 people reading the same sentence, one will be clueless about the meaning, while the other will perfectly understand what the question means.

This method allows the OBA to be much more precise despite a very quick analysis ≃ 15 mins.

We have tested the evolution of results over time

The most reliable method for verifying the robustness of an assessment is the "test retest" method.

This involves administering the same assessment to the same person under the same conditions in order to observe the difference in results between two administrations.

The closer the results are in a short period of time, the more reliable the model is.

The challenge of this method is to determine the time between the two administrations: a time that is too short would logically lead to similar results, but a time that is too long could skew the results as the person could have evolved between the two administrations.

For the OBA, we have determined the following method:

48 people completed the OBA at 4 different times: the first time, 1 month later, 6 months later, and 12 months later to ensure that the results were the same.

In these tests, 82% of respondents obtained the exact same result (same profile letter) between the first administration and the one at 12 months (compared to 6.25% by chance alone).

For those who did not obtain the exact same profile, 96% of them obtained at least one of the letters they had during the first administration (compared to 25% by chance alone).

The best assessments on the market obtain a score of around 75-85%.

Correlation of questions analyzing the same traits (factor analysis)

In the OBA, each question analyzes one or more traits.

Based on this premise, questions analyzing the same traits should have similar results during administration.

To analyze this point, we used the Pearson coefficient (r) which measures the interdependence of 2 factors.

In our case, we measure the score difference between question 1 and question 2 which both analyze the same item A.

We performed this calculation on a sample of 500 people for all 48 questions of the analysis.

We set a goal of r=0.20 for each question. In the first iteration, 91% of our questions achieved r=0.40, validating the initial correlation objective. The remaining questions were more complex (as they evaluate several elements at once). We adjusted them on 8 consecutive versions.

Finally, we were able to achieve our goal for all questions, ensuring a high level of result reliability over time.

Bias and areas for improvement

Every assessment is imperfect and must constantly evolve. Currently, we identify 3 areas for improvement:

1. The analysis of the introversion and extraversion dimensions is less reliable than the other items. Currently, we do not take these dimensions into account in our recommendations.
2. The validity of data over time (test retest) needs to be renewed and on a larger sample.
3. The complexity of questions and the principle of stimulus words can make administration more complex for a person with a less important academic background and/or bias certain analyzed items.