For the selection of taxpayers in each treatment, a strategy of stratified randomization or "by blocks" was followed, which uses each taxpayer’s geographic location to define the strata or blocks. In the case of taxpayers with a correspondence address outside the municipality, an additional artificial block was created since these people may show highly idiosyncratic non-observed characteristics.

Within each block, for each taxpayer randomly assigned to a treatment, one more was assigned to the control. The groups looked like this:

Control: since there are three treatments, 60% of the taxpayers were assigned to the control group (approximately 15,000)

The treatments:
The rest were distributed equally in the treatments within each stratum (about 2,500 per group). To ensure proper balance between the groups, the random selection was run 1,000 times and the random draw was selected which showed the best balance for all pre-experimental covariances, controlling the Dummies by stratum.

The wording of the messages and the selection of the images were prepared by a communications team in coordination with the municipal authorities and tested in a small focus group. In each case and following the standard recommendation in the literature, the objective was to provide the combination of words and images that would have maximum impact when transmitting the intended message.

Treatment 1, deterrence: the first treatment included a message in the tax bill which gave taxpayers a simplified example of the costs of non-payment, by calculating the cost of a hypothetical debt of AR$1,000 after a given year of 2% cumulative interest. In particular, the message stated that for a hypothetical debt of AR$1,000, the taxpayer would have to pay AR$269 for late payment at the end of the year. The message included the image of a judge's hammer and added that the municipality would take administrative and legal measures in case of non-payment.

Treatment 2, legitimacy: the second treatment presented a message on the bill with information on the real use of income by the municipality which explained the investment in works. It highlighted the number of public lighting and sewer connections installed by the Junín local government in the previous six months. The message was also accompanied by a picture of an easily recognizable "men working" road sign. This message was designed to change taxpayers’ perception of the justification of the tax in order to influence their assumptions about the quality and quantity of public services provided by the municipality.

Treatment 3, equity: the third treatment introduced a message on the tax bill about CVP compliance levels. The message asked if taxpayers were aware that only three out of 10 taxpayers did not pay their taxes. It also added a phrase questioning whether the taxpayer was actually paying his/her taxes ("And you?"), which aimed to capture the essence of the arguments of moral persuasion.

Taxpayers of real estate tax

The literature has found that people tend to comply more with their taxes if they believe that others pay, and if they believe that the government uses the money better. Also, a better understanding of the fines and possible legal consequences that result from non-payment can also increase payment.