Lying, donations, and preferences for NGOs: Evidence from Spain

6000 are surveyed online in Spain on political attitudes and rural-urban perceptions (representative sample). The survey is originally in Spanish.

Halfway through the survey, subjects are asked to think about a number between 0 and 5. Then they are asked to reveal the number and at the same time told that the number will be converted into € for a given NGO chosen at at random among 4 NGOs. The goal is to assess willingness to donate depending on individual characteristics and NGOs location of headquarters and main goals.

There are 4 NGOs chose. 2 have the headquarters in Catalonia (Barcelona). One promotes the Catalan culture; the other one seeks to help out political refugees and migrants. The other two NGOs are located in the rest of Spain (Madrid). One seeks to help out political refugees and migrants. The other one promotes LGBTQIA+ rights.

There are two slight variations of the experiment, detailed in the experimental design part.

2023-05-12

2023-05-28

Regional preferences.
Political preferences.
Socioeconomic preferences.
Lying behavior.

Regional preferences. For all subjects, we have their location of residence. We examine whether people from certain regions are more likely to donate to certain NGOs depending on the location of the headquarters.

Political preferences: we have information on past vote, intention to vote, Left-Right self assessment, assessment of main problems in the country, etc. We examine whether these can predict the level of donations.

Socioeconomic preferences. We have information on subjects' background (locality, income, gender, age,...). We examine whether these can predict the level of donations.

Lying behavior: Given subjects are asked to think a number between 0 and 5, we would expect the mean to be 2.5. We examine if the mean for the different groups and subgroups deviates from this. It could be argued that the mean is not necessarily 2.5. We will also examine whether there are significant deviations of the mean reported number across different groups and subgroups.

Confidence in others. Major problems in society. Attitudes towards climate change. Knowledge of people around subject. Prioritization of rural areas.

Confidence in others: in general, most people can be trusted (strongly agree – strongly disagree).
Major problems in society: from a list, select the major problems Spain faces today.
Attitudes towards climate change: should economic growth be prioritized vis a vis climate change mitigation policies? (question framed as options being mutually exclusive).
Knowledge of people around subject: in your place of residence, what % of the population do you personally know?
Prioritization of rural areas: to what extent should rural areas be the focus of policies by the central government?

Setup A (3,000 participants):

First screen. "Before proceeding to the next question, please think of a number from 0 to56. When you have it in mind, press continue".

[once button is pressed]

Second screen: "We have randomly selected [NGO here], based in [location here], which is dedicated to [goal here], to make a donation. The amount of the donation for [NGO here] is equivalent in euros to the number you have thought of. What is the number?

[Of every ten participants, we will take one at random to make the relevant donation. We will pass all the evidence of the transfer to [company carrying out survey] so that they can validate that we have indeed done all of them as promised]"

Setup B (3,000 participants):

First screen. "Within this project, we will randomly select one in ten participants to make a donation to a randomly chosen NGO (in your case: [NGO here] – based in [location here], which is dedicated to [goal here]). Before proceeding to the next question, please think of a number from 0 to 5. When you have it in mind, click continue."

[once button is pressed]

Second screen. "The amount of the donation is equivalent in euros to the number you have thought of. What number had you thought of? If you are selected, we will make the donation for this amount and we will pass all the evidence of the transfer to [company carrying out survey] so that they can validate that the We have done everything as we have promised"

In all cases the name of the NGO contains a hyperlink for participants to check this is a real NGO with a real webpage where donations can be made. This is also provided so that subjects can collect more information on the NGO before proceeding.

Randomization carried out by an algorithm designed by the company carrying out the survey.

Individual. Below the breakdown by location size and NGO.

CATALUNYA
Participant lives in Total OPEN ARMS ACCEM FELGTB ÒMNIUM
Small town 480 120 120 120 120
Medium sized town 320 80 80 80 80
Small city 320 80 80 80 80
Large city 480 120 120 120 120
1600
Rest of Spain
Participant lives in TOTAL OPEN ARMS ACCEM FELGTB ÒMNIUM
Small town 1320 330 330 330 330
Medium sized town 880 220 220 220 220
Small city 880 220 220 220 220
Large city 1320 330 330 330 330
4400

Yes

4 types of locations are identified. Small town (0-5K), 30%. Medium sized town (5-25K), 20%. Small city (25-100K), 20%. Large city (100K+), 30%. The % refers to the expected number of observations from each type of town in the sample.

6000. 1,600 from Catalonia, the other 4,400 from the rest of Spain.

3,000 in each treatment.

Using stata [power twomeans 2.5, diff(-0.25) n1(220) n2(220)] we can see that this yields a power level of 0.33. (Alpha=0.05. Standard deviations assumed to be equal to 1.7). For any of the smallest possible comparison groups, we can detect a difference in means of 25 cents with power 0.33 and significance 0.05. The effect size should be 45 cents for us to reach a power level of (approx) 80% (0.79)