Party points out defects and asks TRE-MT to challenge MT Data research

The party points out a series of supposed objective technical deficiencies in the research, such as, for example, capture by self-selection; fundraising through paid and targeted boosting on Meta’s social networks; structural incompatibility between the declared PPT method and the collection instrument, which does not capture neighborhoods or census sectors; Regional section without source and materially incorrect that distorts the sample weight and numerical inconsistencies are not reconciled.

According to the party, the defects highlighted are structurally identical to those that have already led TRE to challenge research from another institute.

Research by boosting According to Agir, the virtual modality of the research used paid boosting, reinforcing that such a tool allows the institute to target the public that wants the research.

“The content served by paid promotion (Meta Ads) allows the represented party, or whoever has contracted the promotion, to direct the display of the ad and the form to audiences segmented by the platform's own criteria, such as location, age, interests, browsing behavior and previous engagement

with pages or content of a political nature, among other targeting parameters available in Meta's ad manager. This direction is the opposite of the “routine navigation” and randomness declared in the registration: it replaces the user’s spontaneous and uncontrolled exposure with a prior selection, made by the advertiser himself, of which audience will have access to the questionnaire”, argued the party, in the representation.

The party also questions the Probability Proportional to Size (PPT) methods, based on municipalities and census sectors obtained from IBGE, for the purpose of distributing interviews in the “7 regions” of the State.

"It turns out that the questionnaire actually applied in the field, now attached to this representation, does not contain any question intended to capture the neighborhood, street or census sector of the interviewee. Worse still: the questionnaire treats the answers as OPTIONAL. The only filter of a geographic nature contained in the instrument is the question in item 05 of the instructions to the researcher (“Do you vote in this municipality? [if you do not vote, thank and end the interview]”), which only confirms the municipality of voting, without any consequences for inferior territorial unit”, concluded Agir.

Agir requests TRE to immediately suspend the research and notification of the Goal to present the boosting criteria adopted by the institute. On the merits, he requests the merits of the representation and the imposition of a fine on the institute.