The influence of social media

In prior posts, I’ve discussed the significance of the title of my blog: “The medium is the message.” Our communication (message) is influenced by the method in which we communicate (medium). For example, when two people interact in person, then there is a two way conversation where both participate. However, most media tend to be a one-way conversation from a content producer to its consumer, such as books, newspapers, magazines, television, and movies. The Internet transforms this prior paradigm because it allows for one-way (the traditional media available online), two-way (email), and free for all discussions (chat rooms and discussion boards). This also is similar to Karlsen’s (2015) concept of opinion leaders influencing others within their networks. So, we’ve established that the medium has changed. How has the message changed?

One example is political messages through Twitter: the infamous example of President Donald Trump. His words become news all over the world from the traditional news media, social media, and blogsphere. Penney (2016) establishes the popularity of political messages on social media and examines their potential influence on the electorate. He argues that political campaigns have become similar to marketing campaigns in an attempt to persuade voters and establish a candidate’s brand.

What are the effects of the presidential tweets? An earlier article by Fuchs (2012) casts some doubt on the effects of social media. He states that the traditional media seems prone to exaggeration  that social media brings together “mobs” and this has created discussion around the technological effects of social media, which distracts from greater societal issues (why that “mob” formed in the first place due to inequity and inequality). However, Fuchs might be very relevant here. Trump’s tweets may distract from modern social problems and larger political issues. While the news media and politicians focused on “the wall,” what other political issues did not get attention? In addition, does quantity make up for quality if repeated tweets continue to gather attention?

I’ll admit that I feel a bit old because I remember the days when the  political commentators discussed the week’s events on television on Sundays. However, this sphere has expanded and anyone might gather enough followers to be an online opinion leader, although the most popular examples often have successful writing backgrounds too (such as Ariana Huffington or Ann Coulter).

The digital self

In Superconnected, Chayko (2017) discusses techno-socialization and the many variations of our self image. The digital environment and being face-to-face influence how we present ourselves. This made me think of several different examples:

  • Online surveys—the main problem with taking these personality quizzes for fun is whether we answer them based on what we might really do versus what we should do. Our network, via social media, tells us that we should complete these fun surveys and then tell all our friends about the results.
  • Massively multiplayer online role-playing games—no one is an blade-wielding sexy elf in real life. Everyone is pretending to be their fantasy character in a fantasy world. (Why do many men play female characters? Why do some games only feature male avatars?) Gamers form relationships, especially for more collaborative games, that require guilds or raids. Cosplay from these games also has become popular. Blizzard’s World of Warcraft might have had peak numbers of 10 million players worldwide.

While I could think about the non-digital predecessors for each of these examples, our world has exploded with new possibilities engendered by the digital age. Some make life more complicated and others easier, but technology helps us “lifelog” and document our events too.

 

The fallacies of “objective” biological data

Joel Best’s (2016) introduction to social problems outlines the problems with objectivism, the idea that a social problem exists when there is an objective measure of harm to society. However, this is a deceptive idea. What concept is truly objective? In the sixteenth century, “civilized” society would not have judged slavery to be a social problem. In the twenty-first century, we might consider slavery to be a fundamental human rights issue. What is a fundamental human right? The United Nations has a Declaration of Human Rights. Article 5 states: No one shall be subjected to torture or to cruel, inhuman or degrading treatment or punishment. Two controversial counterarguments immediately come to mind: 1) national security needs and 2) the death penalty. And down the rabbit hole we go… which is exactly why objectivism is problematic.

Browne’s (2009) article on Digital Epidermalization provides an excellent example of how biometric data has a patina of science, which is objective, right? However, biometric facial scans cannot distinguish darker skinned features as well as on lighter skin and Asian-descent women have fainter fingerprint ridges (along with elderly individuals and members of certain professions), which lead to a measurable failure of the technology in known instances. In addition to these racial implications, the science is not completely certain and highly subjective to user error.

In the above image, the American Civil Liberties Union (ACLU) used Amazon’s Rekognition product, which misidentified 28 members of Congress as matches to arrest photos. The ACLU cited a significant concern: “Nearly 40 percent of Rekognition’s false matches in our test were of people of color, even though they make up only 20 percent of Congress.”

As another example, direct-to-consumer genetic testing has become popular in the mainstream audience to “uncover your origin” or “commit to a healthier you, inspired by your genes.” I’ve discussed this topic a bit before because these companies essentially have “black box” proprietary methodologies for creating neat charts about your origins. This Science News article even compares the varying results from five different services. Alternatively, this Ancestry.com blog post nonchalantly describes how results can be different across four siblings:

Wait, what?? A Google search provided a few more articles that seem to confirm that ancestry results can differ just like appearance for siblings.

Anyway, these services offer results that seem definitive for an average person who doesn’t read the fine print explanation. Basically DNA samples have been taken from current populations and extrapolated to represent regions. One of the most convincing arguments about the problem with this approach also deals with race: South Africa. In the 17th and 18th centuries, the Dutch colonized the region and settlers became known as the Boers (or Afrikaners). This resulted in both White and Black South Africans. So, how does an ancestry test account for this historical context? If you had a White South African grandmother, would the result reveal a South African heritage or Dutch heritage? The companies make a decision to represent results based on subjective factors, so the science is not completely objective and certain as the charts make it seem.

What’s the lesson here? That there may be no objective truth? (Is teal actually a blue or green?–the answer might depend on the brain of the eye of the beholder.) This might be more of a philosophy question. Or that science is portrayed as objective and certain, but is often not really either? Science is an iterative process of hypotheses, experiments, and validation, where established ideas become theories, but still might be overturned by new evidence. We tend to forget that in science, the only certainty is uncertainty.

 

Social problems

Joel Best subscribes to the constructionist view of social problems. This means that people define social problems through a process of claiming an issue, using media coverage to persuade others to agree the issue exists, convince those in power to fix the problem, implement a solution, and have an outcome.

Let’s take a person I’ve talked about before: Neil DeGrasse Tyson, a famous astrophysicist. The Washington Post stated that he is “likely the world’s most beloved astrophysicist — a strong ambassador for the flagging agency.” Tyson is an expert in his field with advanced degrees and media credits that include academic publications, the popular press, television, and radio. He’s also testified in front of Congress about the future of the National Aeronautics and Space Administration (NASA) in 2012.

Tyson’s testimony on the “Past, Present, and Future of NASA” still rings true today, so I’d like to deconstruct it as a claim to a social problem. According to Best, a persuasive claim has three components: grounds, warrants, and conclusions.

1. Grounds – The speech begins with a dramatic scenario that illustrates the problem: “Currently, NASA’s Mars science exploration budget is being decimated, we are not going back to the Moon, and plans for astronauts to visit Mars are delayed until the 2030s—on funding not yet allocated, overseen by a congress and president to be named later.”

2. Warrants – The audience should care because of nationalism: “For a while there, the United States led the world in nearly every metric of economic strength that mattered… In fact, most of the world’s nations stood awestruck by our accomplishments.”

economic growth: “When you innovate, you lead the world, you keep your jobs, and concerns over tariffs and trade imbalances evaporate.”

and inspiration: “Yet audacious visions have the power to alter mind-states—to change assumptions of what is possible. When a nation permits itself to dream big, those dreams pervade its citizens’ ambitions.”

3. Conclusions – The problem can be fixed by allocating more money to NASA: “For twice that—a penny on a dollar—we can transform the country from a sullen, dispirited nation, weary of economic struggle, to one where it has reclaimed its 20th century birthright to dream of tomorrow.”

The testimony inspired a #Penny4NASA as a citizen movement.

Now, here’s a recent chart of the NASA budget:

The commentary alongside this chart noted that presidential directives for space largely have failed because of a lack of increased funding to support new initiatives.

Another recent commentary stated that the public favors increased funding for NASA.

I searched Google News for the headlines for “NASA funding” or “NASA budget”:

While the story about unlimited funding is amusing, this issue is not at the forefront of current news events. Funding for space exploration is a cyclical issue awaiting further opportunity. The current news events have been discussing the impact of the government shutdown on some space related programs and contractors, but not its lack of budget.

Tyson also has been relatively quiet, but has been dealing with several accusations of harassment in his past (an effect from the #MeToo social movement) since early December 2018.

 

 

Dual axis charts

I’ve never tried a dual axis chart. It’s not readily available as an easy option in Microsoft Excel. I don’t think I’ve even noticed examples of dual axis charts anywhere. As a result, I started by googling dual axis charts and discovered that their use is controversial. For example, Cole from the Storytelling with Data website preferred to focus on the most significant aspect of the story although this website presents an interesting alternative that seems to work. This is another example that discusses various ways to portray the same information with different messages.

So, what would I like to try?

One of my previous research projects has been to examine the connection between public opinion of science and federal funding of science. I could try the following combinations:

  1. Public opinion of science and public opinion about federal funding of science over time
  2. Public opinion of science and budget estimates for federal funding of science
  3. All of the above. However, I admit that this might have terrible results.

Initial Experimenting in Excel

Chart 1: Support for federal funding of science (agree scale) and interest in scientific discoveries (interest scale). This didn’t end up being a “real” dual axis chart because there wasn’t really a need for multiple axis. Both variables measure survey responses, just on different scales. The federal funding variable has an additional available year (2006). It’s kind of messy to have two variables on the same chart like this. Maybe the most obvious message from this visual is that 2010 has a dip in the number of survey responses for both variables? The next message might be that “agree” and “moderately interested” have more responses than a stronger sentiment responses (“strongly agree” or “very interested”).

 

Chart 2: Support for federal funding of science vs federal funding levels of science (in millions). A “real” dual axis chart. However, I’m not getting a clear message from the chart below. The support for federal funding of science seems to dip in 2010 while the federal funding amount peaks. However, the data actually shows a lower total number of responses that year, not necessarily more negative responses.

Chart 2a: Percentage of support for federal funding of science vs federal funding levels of science (in millions). A little data manipulation into percentages to even out the differences in total responses and another attempt. The left-hand axis going up to 120% bothers me, but not sure how to fix it because it seemed to affect the right-hand axis. This version of the chart makes it much clearer that most respondents tend to agree rather than any other response. There also seems to be not much of a relationship between support for federal funding and actual federal funding numbers, even considering a possible delay in effects. For example, higher support in 2008 might mean a higher funding level in 2010. However, what caused the dip in funding in 2014? Not levels of support…

Chart 3: Nope. Too terrible to contemplate.

Using Tableau

Coming soon…

Tableau continued

Tableau Public has not been the most user friendly and intuitive interface to learn.

  • Importing files became the first hurdle. The data came from the General Social Survey (GSS). I downloaded the file in SPSS. However, the GSS dataset is very large and became an unintelligible mess when imported into Tableau.
  • Next, I tried a single variable in a table format in Excel. This resulted in a small confusing jumble. I ended up watching several videos on importing and organizing data, which involved using the Data Interpreter and creating pivots.
  • Then, I started creating a chart while following along with another video. This helped figure out a little about the dimensions/measures and columns/rows. I tried out a few formats and liked the side by side circle charts. I’m not sure if my original vision of little figures is possible.
  • Finally, I though the final hurdle would be integrating the three charts into a single dashboard to place into this post. Another video showed how to make all three of my charts visible in a side-by-side format. Unfortunately, this revealed some data issues with one of the charts, so I had to verify the actual numbers and troubleshoot the original variable filters to remove extraneous data.
  • I looked briefly at filters in the dashboard, but it seems to want to filter by sheet. Ideally, I would set some sort of filter by year to highlight a certain year’s results across the three variables.

Here’s the current chart: https://public.tableau.com/profile/giny.cheong#!/vizhome/PublicOpinionofScienceGeneralSocialSurvey2006-2016/Dashboard1?publish=yes

Other thoughts:

  • Some of the filters options on the right-hand side of the dashboard should not be available because they don’t add anything. However, selecting by year does seem to work.
  • I see a small issue in the first chart where the don’t know category should be in the middle, but I can’t seem to change it and save  without an internal error.
  • The difference in chart scales also is a minor irritation. Maybe converting the original data to percentages would correct the issue between sample size for each variable and responses.
  • Overall, I found this exercise to be difficult because of a complete unfamiliarity with using the program although I had heard of it before.

Updated chart with fixes:

Dashboard 1

Timeline

For a timeline visualization, I would like to portray public opinion about science over time from survey data. The graphic would show something other than a line or bar graph, such as little figures indicating a certain number of people. For example, one particular survey question asked if the benefits of scientific research have outweighed the harmful results, or have the harmful results of scientific research been greater than its benefits (SCIBNFTS). The possible answers include: 1) benefits have been greater, 2) about equal, 3) harmful results have been greater, and 4) don’t know. In 2016, 1,318 total respondents answered this question. The majority chose that the benefits have been greater, approximately 73% or 1,007 respondents. If one little figure equals about 10 people, then the graphic impact of 1,007 versus the 112 who answered that the harmful results have been greater is obvious. The timeline portion would come in at the bottom of this graphic with a slider to move within time. In the General Social Survey, this question has been asked in 2006, 2008, 2010, 2012, 2014, and 2016. On the left hand of the graphic, I may want to include tabs for additional related variables, such as NEXTGEN: Because of science and technology, there will be more opportunities for the next generation – 1) strongly agree, 2) agree, 3) disagree, 4) strongly disagree, 5) don’t know. Here’s a badly sketched out idea of what I’m envisioning:

Here’s a traditional bar graph version of the variable described above (SCIBNFTS):

I googled “timeline graphic” and “public opinion” to see what might come up in a search, but found relatively static line graphs or bar charts from Roper Center or Pew Research Center. Kaiser featured an interactive graphic for public opinion on health reform law that had check boxes to select for opinion, party identification, income, age, gender, and race/ethnicity categories. I also have no idea what might have been used to create the graphic. In addition, FiveThirtyEight had an interesting approach to opinions on Donald Trump. Scrolling over with the mouse pointer moves a bar across the graphic to show time versus opinion rating. Underneath this visualization, there is a listing of individual polls and also mini visualizations to show the ratings of past presidents for comparison. At the bottom of the post, the authors provide their data as downloadable files, but I’m still unsure what tool they used to produce the visualizations. (Is there a trick for looking at the web code and figuring this out?)

I downloaded Tableau Public to see what it might help me do… Since I’ve never worked with this tool, I needed to convert my data to a usable format first. Using a General Social Survey file produced a giant mess of a data, when I wanted a few specific variables. Tableau doesn’t seem to accept SPSS files, so I took an output table and copied it into Microsoft Excel. The table mostly reproduced in the app, but then I had no clue what to do next, so I went looking for the how-to videos online.

To be continued…

The Dream Portfolio

From one of the suggested links, I stumbled upon the Information is Beautiful Awards 2017. However, I did not find all of the award winners to have accessible examples. For example, this site measures the individual scientific impact in different fields. The squiggly lines look quite artistic, but not intuitive for explaining the information within the visualization. I liked this site more for the visual representation of the journey for six asylum seekers. There is a balance between visual simplicity and explanatory text to understand the message. This Washington Post example is more typical of what strikes a balance between informative/educational and visually creative in support of the narrative. Call me academic, but I’m used to text as the main form of communication and visualizations as support.

As a result, my current vision would be modeled after the example of the Washington Post. A website wireframe would center around a narrative on public attitudes about science in the United States from the late 1970s through the mid-2000s. This might include responses to particular relevant questions from cross-sectional survey data over nearly thirty years, such as asking about interest or knowledge in scientific discoveries; support for federal funding of science; or whether science enriches our everyday lives. Survey data also have a demographic aspect that is usually interesting: do certain population groups have more positive or negative opinions? (For example, African-Americans have been cited as more negative about scientific research, which is attributed to historic abuses in human testing.) Elements of the narrative might include:

  • What is the overall trend?
  • What complicates this picture (i.e., certain issues or demographic groups)?
  • How might information be combined to present a visual narrative rather than a single aspect of data?
  • Why is this topic important? Is it possible to say anything about future implications?

 

 

 

Why this course?

Data visualization is ideally supposed to make complex data more accessible, understandable, and usable. I think this occurs for two main reasons: 1) the data is boring, but the author wants to highlight a particular message or 2) the data is very technical, but the author needs to present the information to a layperson.

Edward Tufte presents a great example of how engineers failed to clearly present technical data prior to the Space Shuttle Challenger accident. In summary, the engineers had thirteen technical charts and diagrams to present their case for why the launch should not occur. Their managers remained unconvinced about the chances of an issue because their presentation had technical data, but did not make the explicit visual connection between temperature and failure, such as the chart below that shows launches chronologically.

Example of a problematic chart that lacks a legend and clarity into the presentation of data.

Tufte criticized their charts and reconfigured a table to show clearer causality between temperature and failure (below). If the engineers had been able to give a clear and concise chart, then the message might have been that colder weather increases the chance for failure rather than that the engineers said a failure might occur without any compelling evidence to support their concerns.

Tufte’s suggested revision for the presentation of the same data.

The accident investigation found fundamental issues in the safety and mission assurance culture. In other words, these engineers had no way of reaching the highest levels of management directly to convey their message. They presented the information to their managers, who sent it up the management chain, which diluted the urgency of the message. Tufte also would add that an improved presentation might have been able to convince managers and allow the message to reach a decisionmaker.

This is one of my favorite examples for why careful thought and planning for visual presentation is important. However, I recognize most instances for data visualization do not have life or death consequences unless being “bored to death” counts. The data usually is exciting only to the researcher, so a visual presentation helps make the research more interesting to other people.

In addition, the various required research methods and statistics classes generally focus on learning methodology. The results get dumped into very basic charts or tables straight out of SPSS.

My personal goal for this class would be to use the opportunity to assess prior work to prepare for the practicums requirements. The general practicum will synthesize federally funded scientific research and public opinion. The in-depth practicum will be a case study of direct to consumer genetic testing and some public misconceptions around the science. I think the general practicum presents an opportunity to work with information that others may not find as interesting. The in-depth practicum has the potential for technical explanations to show why an ancestry test cannot reliably provide the neat percentages adding up to 100%.

 

Scale-free networks

A scale-free network shows a power law distribution where there is a predictable imbalance. The 80/20 Rule is an example of a power law distribution: 20% of the population holds 80% of the wealth (except it doesn’t seem to be these numbers anymore).

In a scale-free social network, a few nodes have a high number of connections and most have a low number of connections. The few nodes become “hubs” in the network (as seen by the degree distribution in the graph on the right). In a random social network,  most nodes have an average number of connections with a wider degree distribution (as seen in the graph on the left).

Internet evangelists, such as Clay Shirky, have heralded the potential for every consumer to also act as a producer of online content, resulting in a newfound democratic medium. However, power law distributions in online networks present a significant obstacle to this democratic ideal. Shirky would argue that power law affects large networks, such as the the most popular news sites in the public sphere, but may not affect small networks, such as family and friends reading your blog. If you’re hoping to follow Adriana Huffington’s example with your blog, the likelihood of becoming the next Huffington Post is slim because the existing top news blogs have a significant advantage of widespread recognition and reputation. (For example, when you read the previous, you’ve probably heard of the Huffington Post before.)

Next, think about how many blogs we might read on a regular basis. This number is finite, although the exact number might be higher or lower depending on our responsibilities.  Meanwhile, the number of blogs has grown dramatically. For example, Tumblr reported 357.7 million blogs in July 2017.

Prior posts have discussed the conscious and unconscious influence of our networks on our choices. Therefore, we’re more likely to read blogs that our friends read. Multiply this line of reasoning by millions of individuals and millions of existing blogs. The result is the power law distribution in blogs. A few blogs, such as the Huffington Post and others in the chart below, end up with more power than others because of their high readership.

So, the Internet reflects similar inequalities in social media networks as offline networks.  Melvin Kranzberg’s first law of technology applies here as a good reminder: “Technology is neither good nor bad; nor is it neutral.” The technology (power of networks in this case) does not have an inherent ethical value because it’s the users of the technology that determine the ethical tone of the application, which can certainly be labeled as good or bad.

Privacy Statement