PLUS: How is polling done around the world?

Is this email not displaying correctly? View it in your browser.

 
 

Thursday, September 5, 2019

 

Methods

 
 

 
TEXT ABOUT IMG GOES HERE
 

Using machine vision to understand gender in images

 

In recent years, computer-driven image recognition systems that automatically recognize and classify human subjects have become increasingly widespread. A form of machine intelligence called deep learning is the basis of these image recognition systems, as well as many other artificial intelligence efforts. In a new data essay, Pew Research Center’s Data Labs team describes how we designed a deep learning system capable of identifying men and women in digital images to conduct our studies focused on gender representation in Google image search results and images from news posts on Facebook. Learn about the facial features machine learning models rely on to assess gender with our interactive tool and read our new report about Americans’ views of facial recognition technology.

 
 

 
How is polling done around the world?
 

How is polling done around the world?

 

The methodology behind Pew Research Center’s international surveys can vary from country to country, and what works in one place might not work in another. In the latest installment of our Methods 101 video series, we look at some of the challenges of international polling, including government restrictions on survey work, political or social instability that can make it unsafe for interviewers to do their jobs, and a lack of qualified partners who can help administer surveys locally.

 
 

 
A week in the life of popular YouTube channels
 

A week in the life of popular YouTube channels 

 

As part of a continuing effort to understand the digital ecosystem and the flow of information online, Pew Research Center published a new report based on a comprehensive analysis of YouTube channels with at least 250,000 subscribers and the videos they posted over the first week of 2019. The report builds on a 2018 study of YouTube’s recommendation engine and is the first Center analysis to systematically assess the content of YouTube videos, using both machine learning (to identify English-language videos) and online crowdsourcing (to understand the topic and themes of each video).

 
 

 
 

How Pew Research Center counts unauthorized immigrants in the U.S.

 

The way Pew Research Center experts calculate the estimated number of unauthorized immigrants in the United States is the product of decades of work by Senior Demographer Jeffrey S. Passel, along with former colleagues at the U.S. Census Bureau and the Urban Institute. On the Center's Fact Tank blog, Passel addresses some common questions about the research techniques used to derive the unauthorized immigrant population estimate.
 

 
 

 
 

Why survey estimates of the number of Americans online don’t always agree

 

How many U.S. adults use the internet? While there is a lot of information available from large, high response rate federal surveys (as well as from surveys conducted by Pew Research Center and other organizations), these different sources of information measure internet use in ways that can be tricky to reconcile. In a recent Fact Tank post, we explore why internet use estimates differ between surveys and explain how our research experts approach the question.

 
 

 
 

How we check numbers and facts at Pew Research Center

 

All of the material we release at Pew Research Center goes through a rigorous verification process we call a “number check.” It's called a number check, but we check much more than the numbers. Although there are slight differences in the process depending on the nature of the data used in a given report, the process is rigorous no matter what we're checking. In a recent Decoded post, we provide a step-by-step walkthrough of how one of our researchers number checks a piece of domestically focused survey research.

 
 

 
 

Happy birthday, Decoded! 🎂

 

Pew Research Center’s Decoded blog focuses on the “how” behind our numbers. The blog features content ranging from survey methods to data science to data visualization, and allows researchers to build on and engage with our work.

We celebrated the first birthday of Decoded in June with a thread on Twitter highlighting some of the most popular posts from the past year. These included the first installation in our series on analyzing Center survey data in R, a post on how we use application programming interfaces (APIs) to collect website data, and a write-up from the Center’s data visualization design team on using small multiple charts.

Explore some of our latest posts:
 

 
 

Featured Datasets

 

Fall 2017 Media and Politics in Western Europe Survey

Survey of eight western European countries conducted Oct. 30–Dec. 20, 2017.

 
 

2018 National Survey of Latinos

Survey conducted July 26–Sept. 9, 2018 among a nationally representative sample of 1,501 U.S. Latino respondents ages 18 and older.

 
 

American Trends Panel Wave 33 (space exploration, environmental protection)

Survey conducted March 27–April 9, 2018 among 2,541 respondents.

 
 

Job Openings

 
 
 

Support Pew Research Center

 

Please support Pew Research Center with a contribution on the Center’s behalf to our parent organization, The Pew Charitable Trusts.

 
Donate
 
 
 
 

Pew Research Center is a nonpartisan fact tank. As a neutral source of data and analysis, Pew Research Center does not take policy positions.