Show:
Establishing Causation
SAMPLING
Ice Cream Study
Harvard Registrar List
200 students
Question Wording

ALIEN
Reading
Grab Bag
100
I want to argue that having an alien encounter causes people to be more intelligent. What’s my independent variable? What’s my dependent variable?
IV: having an alien encounter; DV: intelligence
100
What is the target population here? What is the sample?
Target population: undergraduates at Harvard; Sample: the 200 students selected for the study.
100
Who were you with during your alien encounter?” Choices: a) family member, b) friend, c) sibling
Categories are not exhaustive or mutually exclusive.
100
What was the main finding in Pager’s article, and how did she methodologically design her study?
She found that having a criminal record negatively affected employment opportunities, and there was an interaction between race and criminal record. While having a criminal record was bad for all job applicants, it was worse for African-Americans with criminal records (in fact, an African-American without a criminal record had lower prospects than a White applicant with a criminal record, though the difference wasn’t statistically significant).
100
Name some potential values of a variable called “religious affiliation.”
Catholic, Protestant, Jewish, Other
200
What’s an intervening variable? What’s an antecedent variable?
An intervening variable goes between the IV of interest and the DV: IV → intervening variable → DV An antecedent variable affects both the IV and the DV, and comes before both: Antecedent variable DV IV (or, if it’s easier to visualize this way): DV antecedent variable IV
200
What is your sampling frame?
Registrar’s list of students
200
“Did the aliens talk to or poke you when you were in their ship?” Choices: a) Yes, b) No
Double-barreled question.
200
Summarize the main take-home messages we got from the “defining poverty” readings.
There are many ways to define poverty. One distinction in the types of definitions lies whether an absolute or relative measure is best. Figuring out how to define poverty is important because there are political ramifications of this decision.
200
Define reliability. Define validity
Reliability - consistency across contexts and researchers. Validity - the extent certain inferences can be made from a particular measurement. Am I measuring what I think I am measuring?
300
A researcher claims that kids who watch TV are more aggressive. Name the IV, DV, and two possible control variables she could include (think: what else might explain this relationship?)
IV: TV watching; DV: aggression; possible controls: age of child, family income, family structure (single parenthood?), gender of child, daycare use.
300
What kind of variable (nominal, ordinal, interval, ratio) is the variable “average number of times you eat ice cream in a week”?
This is a ratio variable because it has a meaningful “zero” value (0 number of times you eat ice cream has an actual meaning (unlike interval variables like temperature, where “0” degrees doesn’t mean you have “no” temperature): you ate no ice cream this week).
300
Craft two versions of the same question for an alien about his/her planet—one in forced-choice format and one in open-ended format.
Forced-choice: “On a scale of 1-10 with 10 being strong affinity, what are your feelings towards earthlings?” Choices (circle one): 1 2 3 4 5 6 7 8 9 10 Open-ended: “What are your feelings towards earthlings?”
300
List some of the variables that predicted tolerance in 1950s America (Stouffer)
Stouffer: gender, education, religion, etc.
300
Give a conceptual, then operational, definition of “healthiness.”
Conceptual: Not Having a health problem affecting daily life. Operational: Number of visits to Health Services per month.
400
What does it mean to control for a variable, and why is it important? Think about replication and explanation
You control for a variable in order to confirm that the relationship of interest (IV → DV) is not spurious. Controlling for a variable means accounting for that variable in your analysis. If the original relationship still remains after controlling for the variable, then you have replicated the original relationship. If the original relationship goes away after controlling for the variable, then you have explained away the original relationship. Interpretation is just how we describe a case (in words) after we’ve accounted for controls (think of “interpreting” the results)
400
Give an absolute and a relative measure of “whether someone likes ice cream
Absolute: Ask someone, “Do you like ice cream?” and take their answer at face value. Relative: Ask people, “On a scale of 1-10 with 10 being LOVE, how much do you like ice cream?” Average the results. You find that the mean is 8.5. Any value lower than 8.5 indicates that you don’t like ice cream, and any value higher than that indicates that you do like ice cream.
400
Do you not like aliens because of their large craniums, green complexion, disgusting smell, or strange voices?” Choices: a) none of the above, b) large craniums, c) green complexion, d) disgusting smell, e) strange voices, f) all of the above
Using negatives (“not”), which is confusing; loaded words and leading question; choices are not exhaustive; horrible wording
400
What do Pearson et al mean by aversive racism?
Aversive racism is a form of prejudice which characterizes the biases of those who are politically liberal and openly endorse non-prejudiced views, but whose unconscious negative feelings and beliefs get expressed in subtle, indirect and often rationalizable ways
400
What is the difference between stratified sampling and cluster sampling?
stratified: Every group is represented. (useful if between group variation is high) cluster: Only some of the sub-group is represented. (useful if between group variation is low)
500
What is the unit of analysis in the following example: Using detailed 33 category occupational classification and inspecting the proportion of workers in each occupation who use computers and their mean weekly hours of use, we were able to identify six discrete clusters for comparison. Three fell at the high end of the computer use scale, while the other three were at the low end.
The unit of analysis is occupation. Each occupation has a proportion of workers who use computers, and a mean weekly hours of use per worker. The occupations were then classified into clusters and compared, but it is the occupation that takes on the value of “proportion who use computers” (the variable).
500
You want a sample of 100 students for your study. Explain how you would draw a simple random sample for this study. Explain how you would draw a stratified random sample (stratified by house). Explain how would draw a clustered sample (clustered by house)
Simple Random Sample: Put all the registrar names in a hat and draw 100 names randomly. Stratified Random Sample (by house): Let’s say that Harvard’s campus is made up of 5 houses. Twenty percent of students live in Pineapple House, 25% in Apple House, 10% in Grape House, 10% in Banana House, 10% in Cherry House, and 25% in Orange House. So we’re going to randomly select 20 people from Pineapple House (that’s 20% of 100), 25 people from Apple House (that’s 25% of 100), 10 people from Grape House, 10 people in Banana House, 10 people in Cherry House, and 25 people from Orange House to get our total sample of 100 people. Clustered Sample (by house): We’d randomly select, say, 2 clusters (houses). Then from those two houses only, we randomly select 50 students from each.
500
“Do you agree with the President’s recent decision to send spaceships into space to research the social stratification mechanisms underlying societies in Mars?” Choices: a) Agree, b) Neither agree nor disagree, c) Disagree
Using “the President” will steer the respondent into agreeing because he’s a popular/authoritative figure; inappropriate wording (“social stratification mechanisms” may not be a familiar term to everyone)
500
According to Furstenberg, what are 2 independent variables that predict the transition to adulthood?
Education, Social class/income
500
What were the main points in Everything is Obvious?
Everything is not as obvious as they seem. Watts discusses how we often think we know more about the world than we actually do, and illustrates many examples of when “common sense” fails us. Social science often highlights this idea that what’s seemingly obvious isn’t really obvious. See examples he discusses, as well as the details of his arguments.