HW DUE 05.12
What makes a graph 'bad'? What aspects do you find the most frustrating about graphs?
Find a couple of examples of what you think are 'bad' graphs, for whatever reasons you choose. Be sure to have reasons and be able to talk about what makes them a bad graph.
---
Medical Helicopters, redux
What's an "easy" situation?
Working Phoneathon
What's an "easy" call?
Baseball Players
Median Average Salary in US
Civil Rights Voting Act
Paper on Acceptance Rates, the "classic" paper in all of this.
paper will be emailed to you.
Naked Stats, regression:
Example was weight, attached to height
each category changes the value:
weight 4.4 (95% 4.0 to 4.8)
Age 0.08 (.01 to .2)
and so on. we build a model, looking at the interaction between all of these.
some were CATAGORICAL -- if yes, add/subtract some amount
this builds a model to help us predict weight based upon other factors
I think this is based in an econ frame of mind, as it is one way that we talk about groups, with ANOVA being the other (see below)
paper referenced about MBAs: http://scholar.harvard.edu/files/goldin/files/dynamics_of_the_gender_gap_for_young_professionals_in_the_financial_and_corporate_sectors.pdf
with NY Times Context: http://www.nytimes.com/2009/04/29/business/global/29iht-riedgenper.html
How is this different than ANOVA:
short answer, it isn't, but the null hypothesis is a little different:
http://www.allanalytics.com/author.asp?section_id=1413&doc_id=252823&piddl_msgorder=thrd#msgs
"Use regression when you aren't sure whether the independent categorical variables have any effect at all. Use ANOVA when you want to see whether particular categories have different effects."
Bad Graphs: