CROSSTABULATIONS

1
CROSSTABULATIONS

• Problem Set 10

2
Crosstabulation

• Remember the standard convention is that the
  INDEPENDENT Variable is the COLUMN Variable and
  the DEPENDENT Variable is the ROW Variable.
• So in this table, PARTY ID is the column variable
  and Presidential Approval is the Row variable.
• In your worksheet, you will place tallies/IDs in
  cells but in your finished table, the cell
  entries are case counts/absolute frequencies
  (probably the converted in Column Percentages.
• Independent is a legitimate value of the
  ordinal variable PARTY ID and likewise moderate
  (IDEOLOGY), mixed (HEALTH INSURANCE), neutral
  (HAVE SAY), etc. Dont drop these values from a
  crosstabulation (or scattergram) they help
  assess the evidence for the hypothesis.
• Whether No Opinion/Dont Know is a legitimate
  value of PRES APPROVAL is more open to question.
• If you recode a variable (e.g., PRES APPROVAL,
  ABORTION OPINION), you must tell your reader how
  you did this. (Measurement must be public.)
• To assess an association between variables, you
  must look at the whole crosstabulation, not just
  particular cells (frequencies).

3
Association is Between Variables,Not Values
4
(No Transcript)
5
Problem Set 10Column Percentages