Types of surveys
To get people to respond to the questions we offered two ways of answering, using Survey Monkey as the technology platform. The first was to answer the questions directly, on one’s own, by going to the online survey provided at www.abigproject.org/myanswers. The second way was to answer the four questions in groups. One survey from each group would be then downloaded online at www.abigproject.org/answers. Surveys were translated into 16 different languages. All surveys will continue to remain open until December 2012.
Types of outreach
Outreach to engage people in taking the survey was done through a series of strategies including: parties, social media, personal outreach to friends and family, and street-based canvassing. The lion’s share of communication for this initiative has been Internet-based. Note: this has meant that there is a disproportionate number of people with high levels of access to the Internet in our sample. (Click here to see demographics.) The project hopes to partner with groups which provide access to communities without high levels of access in the coming months. (Click here to partner.)
On September 20, the surveys were analyzed to assess the results of the first month’s responses to the four global questions. The process included
- Downloading all surveys from online platform
- Translating surveys from other languages into English using machine translation*
- Working with partners to extract, sort, and analyze data based on two levels: Popularity - words that were cited most often in the responses; and distinctiveness - the level of distinctiveness words had to each region (as well as level of agreement across regions)
- Creation of images that best depicted data (i.e. histograms, wordles, etc.)
- Translation of data back into all languages on website (in progress)
*It is important to note that machine translation is not ideal for many translation efforts – specifically because machines often use literal translations and cannot always determine the context of a word in a sentence. In all other elements of this project volunteer certified translators have been used to support translation. However, in the case of survey data translation it is believed that machine translation is most effective because a) the analysis focuses on a key word analysis (not concept or sentence analysis) so a literal translation of each word is most appropriate and b) because the size of the dataset, there would be no feasible way of translating all responses with the current capacity of volunteer translators.
We recognize that this is NOT a perfect process. We would like to acknowledge that some concepts may get lost in translation, and encourage others in similar efforts to find ways to make this process even more accurate in the future.
**Level of distinctiveness and agreement were determined based on the following formulas:
The report uses graphics to explain much of the results. This is done because we believe it will make understanding about the information easier for people from different backgrounds and because we value graphic imagery as a method to express ideas in ways that sometimes numbers and words alone cannot.
There are three types of graphics provided: 1) ‘Worldles’ (www.wordle.net ) which present words in order of those most recurring based on size; 2) Venn diagram – a Microsoft Word chart which helps represent areas of distinction and areas of agreement; and 3) histograms from BigML, Inc. database (www.bigml.com) . Histograms are interactive online and will be made available to the public soon.
Limitations of Study
Because there are always competing ideas for which countries are included in various regions, as well as which countries are currently recognized as countries, some of our data may be specific to our organization of world areas. See Appendix B in the full report for our breakdown of countries by region. Further, the types of people who filled out the survey may not represent the ‘typical’ viewpoints of people from that country or region. In some cases we only have one or a few people filling out the survey from a specific country. Because of our small sample size in some areas, readers should not seek to generalize the results from any particular region with too much vigor. More confidence in the data will accrue as the number of people taking the survey continue to build.
Further, because this is a primarily Internet-based initiative (using low or no-cost social media platforms and other technologies to support global organizing efforts) there is a strong bias of respondents who have access to the Internet. Approximately 89% of respondents had good or very good access to the Internet, which compares to statistics which suggest only about a third of the world’s population have good access. It is important to recognize that many people and communities without good access to the Internet often also have low access to other key economic resources such as food and safe housing. These findings are therefore not a representative sample of the global population in terms of economic or technology-based factors. Project leaders are making efforts to increase representation of these populations in the coming months.
Additionally, this research report is based on an initiative seeking to gain clarity on how people feel the world could be better, and then to share this data in ways that opens people’s hearts and minds – through art and music. It also seeks to build buy-in and engagement around the results by inviting as many people to participate as possible. There is a tension which exists between communicating the results with the utmost scientific rigor while also garnering interest in the project and making the information accessible to artists, musicians (and aspiring artists and musicians) who want to create projects based on the data. The authors have done as best as they could in meeting all of these goals while maintaining the integrity of the data by using scientific methods wherever feasible, and being as transparent about the process as possible.
The authors acknowledge that this report – as well as this project – is imperfect process. It is hoped that the data support the process of communication and collaboration, rather than the outcome of it. And it is hoped that future endeavors which seek to engage the globe in conversations will build on our work, and find improved ways to do this work.