Why we did this study

This report is part of CNTI’s broader 2024 “Defining News Initiative” which examines questions surrounding this theme in policy, technological developments and the views of journalists, in addition to public perceptions. The survey data in this report measure the public’s perceptions of news, journalism and technology in four countries: Australia, Brazil, South Africa and the United States. These data were collected in parallel with a global survey of journalists. 

CNTI was motivated by several overarching questions:

1. How does the public navigate new ways of being informed?
2. Where do they see journalism fitting in?
3. How can journalism do a better job of communicating its unique value?

Answers to these questions are central for better understanding the evolving information ecosystem. As CNTI learned in a series of focus groups last year in the same four countries, people are putting a lot of work into getting themselves up to speed on news. At the same time, many are actively tuning news out, expressing a sense that it is overwhelming. This is the first in a series of CNTI reports that examines these ideas. 

As with all CNTI research, this report was prepared by the research and professional staff of CNTI. This project was financially supported by CNTI’s funders. 

How we did this

CNTI’s survey questionnaire was developed internally by our research team and advisors in consultation with Langer Research Associates. Focus groups were initially run in each of the four counties which informed the development of the questionnaire. In addition to references made throughout this report, themes from these focus groups may be found in a series of essays available on CNTI’s website.

In partnership with Langer Research Associates, the data were collected through different vendors in each country:

• Australian data are from an Infield International RDD CATI/cell phone sample conducted from September 4 to October 1, 2024. The total sample size was 1,000 respondents. The design effect was 1.39 and a margin of error of 3.7 points. The survey was only available in English. All interviews were conducted by telephone, and the median interview length was 18 minutes and five seconds. There were 41 interviewers used by Infield International all of whom were trained. The sample collection age categories were: 18-24, 25-34, 45-54, 55-64 and 65+.

• Brazilian data are from an Inteligência em Pesquisa e Consultoria (IPEC) RDD CATI/cell phone sample conducted from September 13 to October 3, 2024. The total sample size was 1,000 respondents. The design effect was 1.67 and a margin of error of 4.0 points. The survey was only available in Brazilian Portuguese. All interviews were conducted by telephone, and the median interview length was 16 minutes and 49 seconds. There were 76 interviewers all of whom were trained. The sample collection age categories were: 18-29, 30-44, 45-59 and 60+.

• South African data are from an Infinite Insight RDD CATI/cell phone sample conducted from September 23 to October 16, 2024. The total sample size was 1,012 respondents. The design effect was 1.48 and a margin of error of 3.7 points. The survey was available in English (n = 811), Zulu (n = 138), Sesotho (n = 24), Sepedi (n = 17), Setswana (n = 12) and Xhosa (n = 10). All interviews were conducted by telephone, and the median interview length was 19 minutes and 17 seconds. There were 47 interviewers all of whom were trained. The sample collection age categories were: 18-24, 25-34, 35-49, 50-64 and 65+. 

• United States data are from Ipsos’s probability-based online KnowledgePanel® and was conducted from September 12-21, 2024. A total of 1,670 panelists were initially selected and 1,053 completed the survey. A total of 28 respondents were removed during the quality control process, yielding a final sample size of 1,025 respondents. The design effect was 1.13 and a margin of error of 3.3 points. The survey was available in English (n = 983) and Spanish (n = 42). All surveys were self-administered online, and the median interview length was 10 minutes and 23 seconds. The sample collection age categories were: 18-29, 30-44, 45-59 and 60+. 

Technical reports from the survey vendors are available upon request from info@innovating.news.

How we weighted survey data

Each country’s data were weighted using demographic variables (i.e., age, sex, education and macroregion). While each individual country used a different age breakdown for sample collection, we opted to recode the age breakdown into the following categories: 18-29, 30-44, 45-54 and 55+. These categories were used to ensure there were at least 100 weighted respondents in each group — South Africa has a younger population, especially compared with the United States’ older population. 

For specific questions about the sample frames, weighting procedures and/or additional survey details, please send an email to the research team at info@innovating.news. 

How we protected our data

Data collection was done by the country-specific vendors listed above. The survey included individual-level information such as age, gender, race, political ideology and macro-region. Survey data supplied to CNTI from the vendors did not include names or specific locations of respondents. Each respondent received an unique identifier.

It would be very difficult, if possible at all, to identify survey respondents because CNTI did not collect personal contact information or contact respondents directly. The survey data for this project are securely stored in an encrypted folder that is only authorized to the core research team at CNTI.

How we analyzed the data

Data were analyzed using the R statistical computing language. In addition to base R functions, several packages (libraries) were used to clean and analyze the data. These included: googledrive, pewmethods, pollster, survey and tidyverse. We recoded missing, refused and don’t know responses into a catch-all category to keep the sample sizes for each country consistent across each question that did not explicitly have survey logic (i.e., questions that were asked to every respondent).

Data were analyzed using the demographic sample weights supplied by vendors and Langer Research Associates. These weights were applied to every analysis and topline in this report. Non-statistical analyses consisted of toplines and crosstabs. Statistical analyses found in this report consisted of Chi-square proportion tests (see the “How we tested for statistical significance” section below). 

Exploratory analyses were run to learn about the data and responses within each country. These methods consisted of both linear (ordinary least squares) and non-linear regression (logistic regression, ordered logistic regression). Upcoming reports will implement these approaches to present more information about country-specific attitudes and behaviors.

Preliminary coding of open ends

Each country’s open-ended responses were analyzed separately. We were conservative in large part because of the multilingual nature of our data set. All responses from across the four countries were translated into English.

These translations may not fully reflect the nuance found in the original languages. Terms do not necessarily have exact equivalents between languages, thus we used the translations to stand for larger themes or concepts and avoided attempts to analyze for nuances of meaning against similar terms. 

Responses to both the PSDEFJSM and PSJSMPRO questions were analyzed by examining the most frequent words used within each country’s data to gain a general understanding of prevalent and shared ways that respondents might conceive of the (1) differences between news and journalism and (2) traits of those who produce journalism. 

We will examine the full set of open-ended responses for both the (1) defining news versus journalism and (2) traits of journalism producers items in an upcoming report.

How we addressed data quality