Strategic Ignorance and Privacy - Would entertaining content trigger information avoidance?

2022-05-11

2022-08-17

We will measure the likelihood of avoiding privacy information measured by an indicator variable for information preferences on whether the participant chooses to gather information.

The actual time that participants spend on making the information decision, measured by the number of seconds before clicking the button.

The basic design of the experiment will consist of participants, recruited through the online platform Prolific, watching two videos. Before watching the second video, they will have to decide whether to read some privacy terms or to proceed directly to the video. We will randomize the sample into four different treatments, varying along two dimensions: the videos’ content, either entertaining (Entertain) or normal (Control), and the minimum time indicated that it would take to the read the privacy notice (at least 30 or 10 seconds).We will measure the actual time participants spend on the different elements throughout the task.

Introducing the task, consent form and privacy saliency
The participants will go through a consent form and an information page before watching the first video. We will inform the participants that they will be asked to watch videos and rate them. In the consent form we will give them accurate information about privacy, and how we handle data. In our case the data will be handled in compliance with European Privacy Regulations—GDPR (General Data Protection Regulation). In addition to the information about the handling of personal data, participants are informed that they can leave at any point of the experiment. After the participants have accepted the general terms of the task, they will be asked to answer two questions on the information they have just received, as attention checks. If their answers are wrong, they will be redirected to the consent and information forms to read again, and will be given the opportunity to answer the attention check questions a second time. One of the questions in the attention check will be about the privacy terms of the study, and it will also contribute to make the participants aware of privacy. If they give wrong answers for a second time, they will not be allowed to continue the study. We will also ask if the participants know the company Telenor before starting the study, to control for the effects of potential familiarity of the brand.

Videos - content treatment
For the study we will use four different videos, conducted and produced for the multinational telecommunication company Telenor. Two of them are entertaining commercials, both related to how the mobile technology could be useful. The two other videos are informative and made for brand building purposes; one is focused on online safety and the other explains how mobility data from the mobile network could be used to understand the spread of diseases. To select the videos to be used in the main study, we ran a pretest of the videos, also using the Prolific platform. In the pretest we included six videos, three that could be considered more entertaining and three that have a more informative and educational content. We then measured the level of entertainment for each of the six videos. The pretest confirmed our expectations regarding the level of entertainment, and in the experiment we will use those videos that had the highest and lowest levels of entertainment. The participants in the Control treatment will watch two videos with low levels of entertainment, while those in the Entertain treatment will watch two videos with high levels of entertainment. All four videos are available online:

1. Normal video 1: https://www.youtube.com/watch?v=1xMZjt84CNg
2. Normal video 2: https://www.youtube.com/watch?v=zoMLbKoU28w
3. Entertaining video 1: https://vimeo.com/135555757
4. Entertaining video 2: https://www.youtube.com/watch?v=GDLU5Y8fmzk

The participants in the Control treatment will first watch Normal video 1, which was rated with the lowest level of entertainment in the pretest, followed by Normal video 2; and participants in the Entertain treatment will watch Entertaining video 1, which was rated with the highest level of entertainment, followed by Entertaining video 2.

Privacy notice and time treatment
At the beginning of the first video there will be a quick reminder about the privacy terms and the only choice given to participants will be to continue to the video. The pop-up shows the following text: “Privacy Reminder. For this video, your personal data will be handled as described in the consent form.”. At the beginning of the second video, there will be another message about privacy. In this case, the pop-up will warn about a potential change in the privacy policy. The message will be the following (either 10 or 30 seconds): “The privacy policy might be different for this video. You can read about the privacy policy for this video by clicking on the "Read the policy" button. Reading the policy may take at least 10 (30) seconds. You can also go directly to the video by clicking on the "Agree to the policy" button.”. The participants will have to choose whether they want to read the privacy policy or continue directly to watch the video (and therefore, agreeing to the policy without reading). Since the participants are informed in the consent page that they can leave the study at any point, if they do not want to choose any of the two options, they will be allowed to leave the study. Half of the participants will be given an indication that reading the notice will take at least 10 seconds, while the other half will be given an indication of at least 30 seconds. The actual content in the privacy notice will be the same as the one given in the consent page, so there will be no actual changes to the privacy policy.

Post-experiment survey
After watching the videos, the participants will be asked a set of questions. To check the participants’ attention during the task, they will be first asked questions about the content of the videos. If their answers are wrong for two times in a row, they will not be allowed to continue participating in the study. They will then be asked how much they enjoyed the two videos they have watched, and their secondary beliefs about the level of enjoyment the first video gave to the other participants in the study (this question will be incentivized with a bonus payment). We will also collect basic demographics, and some other background characteristics of the participants. Namely, age, gender, level of privacy concern, risk and time preferences and trust in the telecommunication company Telenor.We will also ask about their views on how they believed reading the privacy notice could have affected their experience of watching the second video. In addition to the information that we collect, we will receive demographic information on country of residence and birth, with status of employment from the online crowd-sourcing platform Prolific.

We plan to run each of the four treatments on separate days, all of them on weekdays from Monday to Thursday at the same time. Subjects are added to a block list when they have participated to one treatment, to make sure that they cannot participate in more than one.

The randomization is done at the the individual level. We have a between-subject design, where one individual is assigned to only one treatment.

No

Same as the number of observations, which we plan to be 1600 (400 per treatment) participants, recruited on the online platform Prolific.

1600 participants, recruited on the online platform Prolific.

400 participants for the entertain-30 seconds treatment, 400 for the entertain-10 seconds treatment, 400 for the normal-30 seconds treatment and 400 for the normal-10 seconds treatment.

0.9 power, with 0.17 effect size, standard deviation 0.025