Information-Score - BTS - Evaluation of information level (not part of the ReDem®-Score)

Estimated reading: 2 minutes

In this article, you will learn everything you need to know about the Information-Score (IS), which can measure the collective information level of responses.

How the Information-Score works and how to interpret it

The information score measures the collective information level per answer option. Answer options that received more votes than collectively predicted are assigned a high score.
If the respondents' information level is high and their information differences are low, all response options will have a positive value. The information score thus reveals information or information differences that are not visible. In general, the more response options have a positive IS, the lower the information differences. A low IS (below 50) indicates that there are larger information differences in the sample. The following diagram shows three answers: The first answer has a lower Information-Score compared to the other two, while the third answer has the highest relative Information-Score.
In order to use the Information-Score, projective question sets must be added to the questionnaire.

Collective responses are used

A key feature of the Information-Score is that it does not evaluate the individual answers of each respondent, but rather the collective response for each answer option.
Since the IS, like the SDS, evaluates collective response behavior and not the individual response quality of respondents, the IS is not part of the ReDem®-Score.

Scientific background

Together with the Prediction-Score (PS) the Information-Score is part of the so-called Bayesian Truth Serum (BTS). This concept originates from more than 15 years of research at the Massachusetts Institute of Technology (MIT), under the direction of Prof. Drazen Prelec. The results have also been published in various scientific publications.

Prelec D. A Bayesian Truth Serum for Subjective Data, Science, 2004

Radas S, Prelec D. Whose data can we trust: How meta-predictions can be used to uncover credible respondents in survey data, PLOSOne, 2019

To apply the BTS, the Information-Score must be combined with the Information score within the weighting function .

Do not use IS and SDS for the same question

It is important to note that the Information-Score and the Social-Desirability-Score cannot be used in conjunction with the same question. The Information-Score is not suitable for capturing socially desirable responses.

Applications and fields of use

Questions in which information discrepancies can generally occur.
Purchase intentions
Employee surveys
Customer surveys

If you have any further questions about the Information Score, please contact us at business@redem.io