Variables, correlation and causality
The two goals of this lab are first, to give you more practice with variables and Excel
charts, and second, to teach you how to be careful in drawing conclusions about what
causes something. Correlation doesn't by itself prove causation!
EXERCISE 1: Generating independent variables
A variable is something that varies. Imagine you are trying to improve Lasell’s
graduation rate by helping students with their obstacles to success, and you want to
know what student situations and qualities predict dropping out or graduating (your
dependent variable). Brainstorm a list of 6 or more variables by which students vary,
and star the ones you think might be relevant independent variables in student success.
(E.g. *Commuting vs residential.)
______________________ _______________________
______________________ _______________________
______________________ _______________________
______________________ _______________________
EXERCISE 2 - See if there’s a correlation between two variables, and display it in
an Excel chart.
Below are the professor’s imaginary results from surveying full-time faculty and part-
time (adjunct) instructors about whether they experience a lot of stress.
THE SURVEY DATA
Stressed (1) or not
First name of instructor PT or FT very stressed (2)
Aldo PT 1
Betty PT 2
Clyde PT 1
Dean PT 1
Emory PT 2
Franny PT 2
Gerardo PT 1
HoWen PT 1
Iggy PT 2
Jessalyn PT 1
KaShawn PT 1
Lefty PT 1
Martha PT 1
Nestor FT 2
Olive FT 2
Paolo FT 1
Queesha FT 1
Rikki FT 2
Selma FT 2
Tiny FT 1
UThan FT 2
Victor FT 2
Wendy FT 1
Xavier FT 1
Yolanda FT 2
Zulaika FT 2
a) Tally the faculty members’ responses and write the totals in each cell of this
bivariate (2-variable) table.
Not very stressed
Stressed (1) (2)
PT
FT
b) In Excel, open a new spreadsheet and type in your bivariate table above. Insert a
column chart, format it and add a title, and paste it here:
c) Speculate on whether there’s a causal relationship or a spurious relationship. Write a
couple sentences about faculty members’ lives and two or more stress factors that
might explain what you see, using the terms “antecedent variable” and “intervening
variable” if relevant.
(Example: “Maybe full-time faculty tend to be older, and part-time adjunct faculty tend to be younger, and
maybe young adulthood is more stressful than middle age in general, because of concerns about finding
affordable housing and having babies. If so, age could be an antecedent variable which causes both
variables, PT/FT and stress.”
___________________________________________________________
EXERCISE 3 - Analyzing correlations, evaluating causality
Below are 6 examples, ranging from preposterous to meaningful, where someone looks
at two variables that correlate. For as many as you can before the lab time is up (saving
10 minutes for reflection in Exercise 4), fill out the “Elements of a Study” table and hand-
draw the U-shaped diagram connecting the variables.
Example:
0. Famous old example that appears in many methods textbooks:
The towns in Holland with the most storks also have the highest birth rates. Therefore storks bring
babies. (FYI: That’s what many European children in the 19th century were told when they asked where
babies come from.)
Elements of a study
Number or Title 0. Famous old example
Hypothesis Storks bring babies.
Unit of analysis Towns in Holland
Independent variable Number of storks
(IV)
Dependent variable Birth rates
(DV)
Evaluate the This is a spurious relationship. That’s obvious by common sense, because we
relationship between know where babies actually come from, and it’s not storks!
IV and DV:
Causal or spurious? In fact, rural areas have more storks and also have higher birth rates than
Why do you think so? urban areas. So the correlation has an antecedent variable, rural/urban, that
Use terms “post hoc,” explains both variables.
“intervening variable”
and “antecedent
variable” when
relevant.
Independent variable= # storksDependent variable = #
births
A=More storks Y =More births
B=Fewer storksX=Fewer births
Elements of a study
Blank form to copy under each example if you don’t want to fill in by hand
Number / Title
Hypothesis
Unit of analysis
Independent
variable (IV)
Dependent
variable (DV)
Evaluate the
relationship
between IV and
DV:
Causal or spurious?
Why do you think so?
Use terms “post hoc,”
“intervening variable”
and “antecedent
variable” when
relevant.
EXAMPLES TO ANALYZE
1. Floyd was poor country boy, but he wanted to be a rich man. He spent much of his
time trying to figure out what made people rich. He drove through rich
neighborhoods, visited rich relatives, read books by rich authors, and watched "Who
Wants to be a Millionaire?,” seeking for a clue to what makes people [Link], one
day he put it all together. Eureka! Every rich man he knew wore a silk tie. Everybody
in the rich neighborhoods wore a silk tie when they went off to work, his rich uncle
wore a silk tie to all those fancy parties, all the backflaps of those rich authors' books
sported a picture of them with a silk tie, even Regis Philbin wore a silk tie. "This must
be it," he said as he went off to Macy's to make his purchase.
2. In 1974 Massachusetts passed a law requiring a special license to carry a handgun
outside the home or business. The law is supported by a mandatory prison
sentence. Studies by Northeastern University professors documented that after the
law was passed handgun homicides in Massachusetts fell 50% and the number of
armed robberies dropped 35%.
3. Many prestigious colleges, such as Ivy League schools, conduct studies that show
that graduates of their college make more money after graduating than graduates of
other colleges. They say this justifies their high tuition prices.
4. Re-read the abortion/crime theory in today’s reading (in Box 2.1, also in text on pp. 20, 21,
22 and 24). This was a controversial study, so put at least two competing
evaluations in the last line.
5. Calvin & Hobbes cartoon
6. Autism and organic food chart
EXERCISE 4 - REFLECTION
You’ve now listed variables, analyzed and diagramed a correlation, and evaluated
causal claims – congratulations!
Write a paragraph reflecting on how this kind of analysis of correlations and causation
could be useful in daily life and in the political realm.
If you need suggestions, here are some prompts to spark your thinking:
· What health fads have you heard promoted by dubious causal claims?
(For example, “Kale makes you thin!” Actually affluence affects both weight and fresh vegetable
consumption. But some consumers disregard overall nutrition advice to follow fads. Understanding
spurious relationships can lead to better food choices.)
· Have you ever heard a real candidate or elected official make a claim that you
thought was a false interpretation of a real correlation?
(For example, US presidents (including the current one) sometimes claim credit for job growth, but blame
other factors for job loss. Actually there are many causes of the number of jobs in the US economy, some
affected and some not by federal policy. But overstating presidential influence can sway voters to cast
votes against incumbents during recessions and for incumbents during economic booms. Being careful
about causal claims can lead to more informed voting.)
· What social factors are often overlooked in bogus misinterpretations of
correlations?
· How could you use the word “stereotypes” in describing the perils of logical
fallacies about causality?
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
