Wednesday, December 17, 2014

Cheese Consumption and Number of People Who Died Tangled in BedSheets, And Other Examples of Spurious Correlations

Looking for examples of spurious correlations to illustrate for your students the dangers of confusing correlation and causation?  Tyler Vigen offers over 40 examples of spurious correlations he derived from real data, including this one between cheese consumption and people who have died by getting tangled in their bed sheets (r = 0.95!) :


Or Tyler Vigen made it possible for you (or your students) to discover your own spurious correlations: click on "Discover a New Correlation."

Students may also find Vox's "Eight Ways to Be A More Savvy Science Reader" a helpful read on how to evaluate scientific evidence.

Read more:
http://tylervigen.com/
http://www.vox.com/cards/savvy-science-reader/scientific-evidence-intro-card#E7053536

TeachingwithData.org resources:
Dancing Statistics: Explaining the Statistical Concept of Correlation Through Dance (http://www.teachingwithdata.org/resource/3908)
Statistics Can Be Misleading (http://www.teachingwithdata.org/resource/3005)

Monday, December 15, 2014

Racial and Ethnic Wealth Gaps Highest In 30 Years

A recent Pew Research Center analysis of data from the Federal Reserve’s Survey of Consumer Finances shows that the Great Recession has widened wealth inequality along racial and ethnic lines.  The wealth gap between Whites and Blacks is now at its highest since 1989, and the gap between Whites and Hispanics is at its highest since 2001.

"The wealth of white households was 13 times the median wealth of black households in 2013, compared with eight times the wealth in 2010 [...]. Likewise, the wealth of white households is now more than 10 times the wealth of Hispanic households, compared with nine times the wealth in 2010."



The Great Recession has affected all households, but not all have benefited from the recovery: "From 2010 to 2013, the median wealth of non-Hispanic white households increased [...] by 2.4%.  Meanwhile, the median wealth of non-Hispanic black households fell 33.7%, [...]. Among Hispanics, median wealth decreased by 14.3%."

Factors that may explain the widening of the wealth gaps include:

  • Differences in the extent to which White and minority median incomes were hit by the Recession.
  • White households are more likely to own financial assets, such as stocks, which have recovered in value more quickly than housing.
  • The decrease in asset ownership has tended to be proportionally greater among minority households.


Read more:
http://www.pewresearch.org/fact-tank/2014/12/12/racial-wealth-gaps-great-recession/#more-265631
http://www.federalreserve.gov/econresdata/scf/scfindex.htm

TeachingwithData.org resources:
Wealth Inequality in America (http://www.teachingwithdata.org/resource/3922)
Attitudes about Racial Discrimination and Racial Inequality in the US: A Data-Driven Learning Guide (http://www.teachingwithdata.org/resource/3431)
Social Class and Attitudes about Inequality: A Data-Driven Learning Guide (http://www.teachingwithdata.org/resource/3459)
Recession Trends (http://www.teachingwithdata.org/resource/3067
Income Inequality in the US (http://www.teachingwithdata.org/resource/3182
Race and Ethnic Inequality (http://www.teachingwithdata.org/resource/3101

Thursday, December 11, 2014

Parable of the Polygons: Interactive Exercise To Teach/Learn About Bias and Racial Segregation

Vi Hart and Nicky Case have created an interactive based on Nobel-Prize winning game theorist Thomas Schelling’s 1971 model of racial segregation.  The game is about how "harmless choices can make a harmful world,"
"a society of triangles and squares who are all only very slightly “shapist”: they actually prefer living in diverse neighborhoods, but they also demand that a certain proportion of their neighborhood are like them. The game starts with one simple rule: Shapes want to move if less than 1/3 of their neighbors are the same shape as them. The object of the game is to drag and drop unhappy polygons until they are all in a position that makes them happy."



The game makes three important points:

  1. Small individual bias can lead to large collective bias.
  2. The past haunts the present: "Your bedroom floor doesn't stop being dirty just coz you stopped dropping food all over the carpet. Creating equality is like staying clean: it takes work. And it's always a work in progress."
  3. Intervention is necessary to maintain diversity.
Go to http://ncase.me/polygons to play the game!

Read more:
http://www.washingtonpost.com/blogs/wonkblog/wp/2014/12/09/how-tiny-individual-biases-have-huge-cumulative-effects-on-racial-segregation/
http://ncase.me/polygons/
http://www.stat.berkeley.edu/~aldous/157/Papers/Schelling_Seg_Models.pdf

Wednesday, December 10, 2014

Americans Have Grown Increasingly Supportive of Torture

A number of polls have shown that an increasing share of the American public views torture as justifiable.  According to the Pew Research Center,
"when the Pew Research Center first surveyed on the subject in July 2004, a narrow majority (53%) said the use of torture to gain important information from suspected terrorists could be only rarely or never justified. Opinion has shifted since then, with more Americans finding torture acceptable. In August 2011, a narrow majority (53%) of Americans said the use of torture could be often or sometimes justified, while 42% said it could only rarely be justified or not be justified at all."
An Associated Press poll conducted in 2013 confirmed the Pew findings.


In spite of the recent release of a damning Senate report on the CIA's use of harsh interrogation techniques during the George W. Bush administration, experts expect the public to remain supportive of torture, perhaps due to the polarization of the electorate: Pew found that a "large majority of Republicans (71 percent) said torture could be at least sometimes justified, compared with 51 percent of independents and 45 percent of Democrats."




Read more:
http://fivethirtyeight.com/datalab/senate-torture-report-public-opinion/
http://www.huffingtonpost.com/2014/12/10/torture-poll_n_6301036.html?utm_hp_ref=politics
http://www.pewresearch.org/fact-tank/2014/12/09/americans-views-on-use-of-torture-in-fighting-terrorism-have-been-mixed/

TeachingwithData.org resources:
The War on Terror: A Data-Driven Learning Guide (http://www.teachingwithdata.org/resource/3465)
Respondants reporting whether they think it will be necessary to give up some civil liberties to curb terrorism (http://www.teachingwithdata.org/resource/2884)

Monday, December 8, 2014

Interactive Mapping Tool Shows How Young Adults Today Compare With Previous Generations

Census Explorer, the Census Bureau's interactive mapping tool, has released “Young Adults: Then and Now.”  The tool uses data from the 1980, 1990 and 2000 Censuses and the 2009-2013 American Community Survey (a part of the decennial census), and allows users to explore characteristics of the young adult population (age 18-34) across the decades.  The American Community Survey especially provides detailed data on a wide range of economic and social indicators aggregated at several geographical levels (from block groups to state and country).
According to Jonathan Vespa, Census Bureau demographer, "Many of the differences between generations examined within these latest data reflect long-term demographic and societal changes. Three decades of decennial census statistics combined with the latest American Community Survey statistics give us a unique view of how — and where — our nation is changing. In this case, we can look at the changing characteristics of young adults over the last few decades."

Here are some selected findings from the Census Bureau report:
In 1980, 30 percent of the population was age 18 to 34, compared with 23 percent today.

The percentage of young adults today who are foreign born has more than doubled since 1980 (15 percent versus 6 percent).

  • All states have higher proportions of foreign-born young adults than 30 years ago.  The increase was larger in the West and Northeast, where 21 percent and 18 percent, respectively, are now foreign born, compared with 12 percent and 8 percent 30 years ago.  
  • Only 9 percent of young adults in the Midwest and 14 percent in the South are foreign born, up from 3 and 4 percent, respectively, in 1980.  
  • One in four young adults, or 17.9 million, speaks a language other than English at home. That proportion is higher still in New York, New Jersey, Texas, New Mexico and Nevada (where it is about one in three) but is highest in California (where it is about one in two).

More millennials are living in poverty today, and they have lower rates of employment, compared with their counterparts in 1980:

  • One in five young adults lives in poverty (13.5 million people), up from one in seven (8.4 million people) in 1980.
  • Today, 65 percent of young adults are employed, down from 69 percent in 1980.
Prior generations of young adults were more likely to have ever served in the armed services: 9 percent were veterans in 1980, compared with 2 percent today.

Millennials are more educated than young adults in 1980:  22 percent have a college degree, up from 16 percent in 1980. States with the largest share of young college graduates are in the Northeast, including Massachusetts, New York and New Jersey.

Unlike in prior generations, the majority of millennials have never been married, reflecting continued delays in getting married: Only about three in 10 young adults have ever been married, down from six in 10 in 1980.  The state with the highest share of married young adults is Utah (51 percent); the lowest is Rhode Island (25 percent).

For more data, charts, and maps, go to: http://census.socialexplorer.com/young-adults/#/report/full/nation/US

Read more:
http://www.theatlantic.com/business/archive/2014/12/where-to-move-in-america-if-you-want-a-job-more-money-or-lots-of-single-people/383434/?utm_source=FB1205_3
http://census.socialexplorer.com/young-adults/#/report/full/nation/US
http://www.census.gov/newsroom/press-releases/2014/cb14-219.html
http://census.socialexplorer.com/pop-flash/

TeachingwithData.org resources:
Social Change: A Data-Driven Learning Guide (http://www.teachingwithdata.org/resource/3458)
Economy Track: Employment to Population Ratio (http://www.teachingwithdata.org/resource/2936)
Marital Trends (http://www.teachingwithdata.org/resource/3116)
Family Change 19550 to 1990 (http://www.teachingwithdata.org/resource/3112)
Interacting with America (http://www.teachingwithdata.org/resource/3041)
Education in America (http://www.teachingwithdata.org/resource/3124)

Wednesday, December 3, 2014

Country Deeply Divided Over Ferguson Grand Jury Decision

A new ABC News/Washington Post poll shows that views of the grand jury decision in the Ferguson, MO, case are strongly divided along racial, generational, and ideological lines.

As a whole, respondents are split evenly: 48 percent said that they approve and 45 percent said that they disapprove of the grand jury’s decision not to bring criminal charges against police officer Darren Wilson in the fatal shooting of black teenager Michael Brown.  Respondents are similarly split, 48-47 percent, on whether the federal government should bring civil rights charges against Wilson.

But White respondents are much more likely than Black or Hispanic respondents to approve of the decision not to indict Wilson: 58 percent of White people polled approve of the grand jury decision, compared to only 9 percent of Black people and 32 percent of Hispanic people.


The poll also indicates that views are sharply divided along partisan lines: 76 percent of Republicans, but half of independents and only 27 percent of Democrats, approve of the grand jury's decision. As for bringing civil rights charges against Wilson, about three-quarters of Democrats, fewer than half of independents and, just 21 percent of Republicans say they'd approve.

"There's a similar divide by ideology, with approval for the grand jury action ranging from 74 percent among strong conservatives to 47 percent of moderates and 29 percent of liberals. At the same time 62 percent of liberals say they'd approve of the federal government bringing civil rights charges; 51 percent of moderates agree, dropping to 29 percent of strong conservatives.

The generational differences are equally sharp, with 62 percent of seniors approving of the grand jury decision, compared with 30 percent of those under age 30. And while two-thirds of millennials approve of efforts to pursue a civil case, just a third of seniors agree."


These results are consistent with the results of a 2013 Pew Research Center survey highlighting that "Blacks and Whites have very different views about many aspects of race — from confidence in the police to progress on racial equality."

Read more:
http://www.langerresearch.com/uploads/1144a47Ferguson.pdf
http://abcnews.go.com/blogs/politics/2014/12/racial-generational-political-divisions-mark-americans-attitudes-on-ferguson/
http://www.vox.com/xpress/2014/12/2/7319597/ferguson-grand-jury-poll
http://www.pewsocialtrends.org/2013/08/22/kings-dream-remains-an-elusive-goal-many-americans-see-racial-disparities/

TeachingwithData.org resources:
Race and Ethnic Inequality (http://www.teachingwithdata.org/resource/3101)
Gun Violence in America (http://www.teachingwithdata.org/resource/3864)
Fear of Crime (http://www.teachingwithdata.org/resource/3155)
Crime Victimization in the US: A Data-Driven Learning Guide (http://www.teachingwithdata.org/resource/3437)
Crime and Victims Statistics (http://www.teachingwithdata.org/resource/3261)

Tuesday, December 2, 2014

New Study Finds No Support For The "Mean Girls" Argument

The belief that relational (indirect) aggression is a predominantly female behavior has been popularized by some research studies and the media.  Girls are frequently depicted as harming others by manipulating or damaging relationships, whereas boys are thought to be more overtly aggressive.  But a new study conducted by researchers at the University of Georgia (Pamela Orpinas and Caroline McNicholas) and Southeastern Louisiana University (Lusine Nahapetyan) found that while there are gender patterns in relational aggression, these patterns are not the ones we think.

Orpinas and colleagues used a sample of 620 randomly selected sixth graders from nine middle schools in Northeast Georgia.  Participants were followed and surveyed every year from grade 6 to grade 12, in order to identify developmental trajectories of relational aggression perpetration and victimization, investigate possible gender patterns, and find out whether there is overlap between perpetration and victimization.

Scores of relational aggression perpetration were calculated by measuring how often during the 30 days prior to completing the survey participants: (1) did not let another student be in the group; (2) told students you would not like them unless they did what you want; (3) tried to keep others from liking another student by saying mean things about him/her; (4) spread a false rumor about someone; (5) left a student out of an activity on purpose, (6) said things about another student to make other students laugh.  Relational aggression victimization scores measured how often survey participants were victims of those same acts.


The researchers found that:

  • Participants tended to follow three distinct trajectories of relational aggression perpetration (low, moderate, and high declining) and victimization (also low, moderate, and high declining).
  • Relational aggression perpetration and victimization scores decreased with age.
  • At each grade level, mean scores of perpetration of relational aggression were higher for boys than girls, and mean scores of victimization were higher for girls than boys.
  • Boys were overly represented in the higher trajectories of perpetration, while girls were overly represented in the higher trajectories of victimization.
  • There was substantial overlap between trajectories of perpetration and victimization: two thirds of participants were classified in concordant groups.

The authors concluded that "more research is needed to understand the multiple factors that influence relational victimization, such as same-sex and other-sex victimization, sensitivity to rejection, social status, value of the relationship, and other personal and interactional characteristics."

Read more:
http://news.uga.edu/releases/article/uga-study-mean-boys-not-mean-girls-rule-at-school-1214/
http://onlinelibrary.wiley.com/doi/10.1002/AB.21563/pdf

TeachingwithData.org resources:
Indicators of School Crime and Safety (http://www.teachingwithdata.org/resource/3878)
Body Image, Gender, and School Experience in Adolescence (http://www.teachingwithdata.org/resource/3836)
Religion among Teens: A Data-Driven Learning Guide (http://www.teachingwithdata.org/resource/3239)