Effect of different interventions on attention to financial campaign

We conducted a field experiment that investigated the effectiveness of different communication methods on the awareness and actions of an underprivileged population in Israel: Ultra-Orthodox Jewish women with low objective and subjective financial literacy. The women were recruited from a class at a college for Ultra-Orthodox Jewish women. Ultra-Orthodox Jewish women tend to marry young and undertake paid work to support their husbands, who commit to full time religious study. For our experiment, we used the launch of the “Money Mountain 2” campaign in 2017. Like the original Money Mountain campaign, Money Mountain 2 was launched to help the population find inactive accounts and had the same website as the original Money Mountain, but it extended this service from the retirement savings accounts to other types of bank accounts.
In the field experiment, the interventions we tested included personal and non-personal digital and non-digital interventions (Laudenbach et al. 2018): (1) no intervention (for the control group), (2) an e-mail explanation of the financial campaign, (3) an e-mail explanation together with a video explanation featuring a professional actor, (4) a face-to-face explanation of the financial regulation given by an employee of the Bank of Israel (the organization in charge of banking regulation), and (5) an e-mail explanation given to part of the control group after they had filled out a baseline survey. This last intervention enabled us to isolate the effect of an e-mail on a group that had had an earlier encounter with a Bank of Israel employee (who handed out a baseline survey). We ran a survey on each group following the interventions to investigate their effectiveness.

2017-08-02

2017-09-12

We find that the interventions that include a more personal interaction (including an e-mail with an accompanying video presentation), increased the awareness of the campaign by more than 100% relative to the control group. These interventions were also more successful in raising the percentage of subjects visiting the website (from 14% in the control group, to between 16% and 28% in the treatment groups). Our conjecture is that a more personal interaction helps lower observation costs or perceived transaction costs.

All participants filled out at least one survey that was used to collect data on individuals’ characteristics and attention to the issue.
The interventions we examined in the experiment are as follows:
1. No intervention: Control group only filled out a baseline survey handed out by an employee of the Bank of Israel during class.
2. E-mail intervention: Received detailed information on the financial regulation in an e-mail and later filled out a survey in person during class.
3. E-mail–video intervention: Received detailed information on the financial regulation in an e-mail, along with a video presentation by a professional actor,90 and later filled out a survey in person during class.
4. Face-to-face intervention: Received a face-to-face explanation of the financial regulation from an employee of the Bank of Israel (the organization in charge of banking regulations) and later filled out a survey in person during class.
5. Double survey–e-mail intervention: Filled out a survey in person handed out by an employee of the bank of Israel during class, received detailed information on the financial regulation in an e-mail, and then filled out the same survey in person for the second time during class.
The e-mail provided (1) detailed instructions on how to access and use the website, (2) an explanation of the steps to take to close accounts if they find inactive funds (e.g., contacting the relevant bank), and (3) a direct link to the website. The video had an actor explain all this information using screenshots from the website. The face-to face explanation provided the same information.
The various interventions described above enable us to test the effectiveness of the mode of communication, in particular digital versus face-to-face. In addition, we examine the dimensional effect of adding a video presentation to an e-mail text, or of having the class meet and interact with a person before receiving more detailed information. The video presentation was 70 seconds (1:10 minutes) long.

The fact that we could enter and re-enter classes to make the interventions and collect the survey data enabled us to relay information in a supervised framework with the possibility of following up on the impact of the intervention on the individuals. Classes were randomly selected to be the control and intervention groups. Each class was part of a different academic program or curriculum such that students in one class were not taking courses with students in the other classes. To further minimize the risk of spillovers between treatment areas, the face-to-face intervention took place on a different campus in another city where there is another branch of the college. This branch is located in an area as central as the other campus but with lower socioeconomic attributes. Each intervention group was made up of two classes: one in education and one in a health profession, except for the face-to-face intervention group, which was made up of a single education class.

Randomization was done by class affiliation

No

5 classes

300 students

Control group - 78 students, Email intervention 33 students, Email-video intervention 42 students, Face-to-face intervention 29 students, Survey-email intervention 43 students