|
|
|
METHOD
Throughout its three stages, PROFECI makes use of a diverse range of methods.
|
•
|
In its first stage, PROFECI relies on several qualitative methods, as well as some limited quantitative data, in order to understand the practices and interaction patterns involved in the construction and negotiation of public projections.
| |
•
|
In its second stage, PROFECI builds upon the results from the qualitative investigation to develop quantifiable categories and computational strategies suitable to automatically identify and classify public projections in different kinds of discourse.
| |
•
|
In its third stage, PROFECI applies these computational methods to trace the evolution, reception and impact of public projections over time.
|
At present, PROFECI is in its first, qualitative stage.
STAGE I: Analytical Framework
The first stage of PROFECI's research serves to obtain a thorough understanding of how both professional forecasters and lay publics project possible futures and receive and interact with one another's projections. For this purpose, PROFECI's first in-depth case study focuses on the three Israeli 2019-2020 general elections. Specifically, PROFECI collects five kinds of data:
At the same time, PROFECI is also collecting news coverage and selected actors' social media content on the US presidential elections and the Brexit referendum.
Public opinion survey
|
◼
|
Who do we collect data from?
| |
PROFECI is administering a multi-wave panel survey to a representative sample of the Israeli voting population.
| |
◼
|
How do we collect data?
| |
According to current plans, there will be nine waves over the duration of the three successive general elections. The survey is administered through the public opinion research company iPanel.
| |
◼
|
What data do we collect?
| |
The survey collects data on three broad fields of interest: First, we ask what people expect to happen in the upcoming elections – which party will be the largest, who will become new prime minister, etc. – how confident they are about this and how they feel about these projections. Second, we ask what news media and other information channels people rely on to keep up with current developments and inform their expectations. Third, we collect a range of data on political attitudes and orientations (political trust, left-right-placement, etc.). We also collect some basic demographic data (gender, age, education, region of residence, mother tongue).
| |
◼
|
How do we analyze the data?
| |
PROFECI's analysis of the survey data focuses on the identification of common patterns in participants' expectations over time, as well as systematic associations between their projections and their use of information, collective identities or more specific political attitudes. There is no analysis of individual participants' responses.
| |
◼
|
How do we protect the data?
| |
All data collected in PROFECI's survey waves is fully anonymized. PROFECI possesses neither contact data nor any other personally identifying information for any survey participant. We match the same participant's responses across survey waves using a random number identifier created at the first survey wave, which does not permit us or anyone to connect the responses to a particular person. All identifying information is held by iPanel, which does not have any access to the data, and PROFECI has no access to the identifying information.
|
Focus group interviews
|
◼
|
Who do we collect data from?
| |
PROFECI has convened five waves of focus group interviews with Israeli voters. Each group comprised 7-12 participants. Four groups were convened in Hebrew language and comprised Jewish voters, and one group was convened in Arabic language and comprised Arab voters.
| |
◼
|
How do we collect data?
| |
Each group was convened five times. The Hebrew language groups were convened at the premises of IPSOS, a public opinion research company in Ramat Gan. The Arabic language groups were convened at community centers in the Haifa area. All discussions were videorecorded and transcribed.
| |
◼
|
What data do we collect?
| |
In each group meeting, we asked participants to discuss their expectations for the upcoming elections: Why do they expect specific events to happen? How certain are they about these expectations, and how do they feel about these? What do these projections mean for them personally and for society and the state more generally?
| |
◼
|
How do we analyze the data?
| |
PROFECI's analysis of the focus group transcripts focuses on common patterns in how participants construct and justify their projections about future events and developments, and how they interact with one another in their discussion of shared or contested projections.
| |
◼
|
How do we protect the data?
| |
All transcripts of the discussions are fully anonymized, with no identifying information retained in the data. The videorecordings are stored in a safe, encrypted location and will be destroyed as soon as they are no longer required for creating the transcripts. Personal contact information was held only by IPSOS and the researcher convening the Arabic language group, who are not involved in the processing and analysis of the data.
|
In-depth Interviews with journalists and professional forecasters
|
◼
|
Who do we collect data from?
| |
PROFECI conducts interviews with selected journalists, experts and professional forecasters who regularly publish projections regarding the outcomes and implications of the Israeli general elections.
| |
◼
|
How do we collect data?
| |
All interviews are conducted in person in a location of the interviewee's choosing. Interviews are audiorecorded and subsequently transcribed.
| |
◼
|
What data do we collect?
| |
The interviews serve to better understand how different actors form and communicate their predictions about future events and developments. Interviews are semi-structured to accommodate the specific positions and practices of each interviewee, but generally cover both their use of information to predict future events, their communication activities, and their perception of the impact of their own activities on the public debate.
| |
◼
|
How do we analyze the data?
| |
PROFECI's analysis of the interview transcripts focuses on identifying professional forecasters', journalists' and experts' distinct strategies for formulating and validating future expectations, as well as their use of information, theories and models in this process. In addition, we aim to understand how these actors refer to one another, and rely on one another's projections to inform their own expectations.
| |
◼
|
How do we protect the data?
| |
All interview transcripts are carefully anonymized. Audiorecordings are kept in a safe location and will be destroyed when they are no longer needed to create the transcripts. Throughout the interview, interviewees have the opportunity to restrict our use of any parts of the interview and can decide what can be cited and how it can be attributed in any research outcomes created by PROFECI. PROFECI will not disclose the identity of any interviewee without the explicit consent of the interviewee.
|
Social media content published by selected political actors
|
◼
|
Who do we collect data from?
| |
PROFECI collects social media data from key political office holders (presidency, parliament speaker, heads of parties and parliamentary groups represented in parliament), the leading political commentators in major media, and a handful of other public figures who actively comment on political affairs using social media and have high visibility social media.
| |
◼
|
How do we collect data?
| |
All Twitter content is retrieved from the public Twitter API. All Facebook data is collected directly from the actors' publicly available Facebook pages.
| |
◼
|
What data do we collect?
| |
For each selected actor, we identify any active, public social media accounts. From each account, we collected the full text of actors' social media posts, as well as some statistical information about shares, likes, and the number of comments/replies. No contents from comments or replies by third users are collected.
| |
◼
|
How do we analyze the data?
| |
PROFECI's analysis of social media data serves two main purposes. On the one hand, we study how different kinds of actors present their projections on social media. On the other hand, we are interested in tracing specific projections from their original presentation throughout their multiplication and discussion by other actors and on the mass media.
| |
◼
|
How do we protect the data?
| |
All collected social media content is published by major public actors and already widely accessible in public, if not covered in the news. Given the high visibility of selected actors, it is inevitable that some actors could be identified based on their posts. In order to mitigate that risk, PROFECI will store recorded data in a pseudonymized fashion and will, beyond legal rules for citations from public communication, not quote verbatim from this data without the author's consent. To protect these actors' right to control their own publicly available data, the raw content collected from social media will be kept strictly confidential. PROFECI exclusively collects social media data that is fully public and does not collect or store any contents posted by third actors.
|
Mass media coverage
|
◼
|
Who do we collect data from?
| |
PROFECI collects the textual and, to some extent, audiovisual coverage of leading television channels, newspapers, and digital news outlets, as well as selected additional news media that are influential within specific publics beyond the mainstream media.
| |
◼
|
How do we collect data?
| |
All news content is retrieved from commercial news archives or directly from the news outlets' publicly available web pages.
| |
◼
|
What data do we collect?
| |
Our data collection includes all news coverage directly related to the elections, the election campaigns and election outcomes. Beside the content of news coverage, we store some metadata (date, author, classification).
| |
◼
|
How do we analyze the data?
| |
PROFECI's analysis of news media content serves two main purposes. On the one hand, we study how projections, both by journalists and other, cited actors, are presented and discussed in the news. This analysis serves to better understand how journalists transform available projections to become part of the news. On the other hand, we trace what kinds of projections are presented by what outlets and how they evolve over the course of the election campaign.
| |
◼
|
How do we protect the data?
| |
As all collected content is public news coverage, the need for data protection is diminished. To protect the data owners' rights to control their own content, no collected content will be rendered available to third parties.
|
General Statement regarding Data protection and Privacy
PROFECI's protection of all personal data are in full compliance with the requirements laid down by ▸The Hebrew University and the ▸European Research Council (ERC).
The University's ▸Privacy Notice provides information on how The Hebrew University collects and uses personal information when a human subject takes part in one of our research projects. The up-to-date Privacy Notice can be downloaded ▸here.
For any question regarding research participants' rights, the ways to exercise such rights and the Privacy Notice please contact Roy Azrad, the Hebrew University's Data Protection Officer (DPO), by e-mail at ▸privacy@huji.ac.il.
|
|
|
|
|
