Distributions of two

variables
From one variable to two
• Having examined the distributions of each variable separately
(univariate distributions), we can now look at the relationship
between two variables (bivariate distributions).
• Take the example of the `happiness' scale. Suppose we were to
include the variable `gender' (or sex), in addition to the happiness
scale. We could present this first as two separate univariate
distributions: one for male and the other for female. These two
distributions are called conditional distributions because they define
the distribution of happiness conditional on gender.
From one variable to two
• Total percentage

• Column percentage

• Row percentage
INDEPENDENT AND DEPENDENT
VARIABLES
Researchers often measure independent and dependent variables in
studies to test association or cause-and-effect relationships.
• The independent variable is the cause. Its value is independent of other
variables in your study. Independent variables are also called
Explanatory variables (explain an event or outcome), or Predictor
variables (used to predict the value of a dependent variable).
• The dependent variable is the effect. Its value depends on changes in
the independent variable. Dependent variables are also called Response
variables (they respond to a change in another variable), or Outcome
variables (they represent the outcome you want to measure)
Identifying independent vs. dependent variables
Distinguishing between independent and dependent variables can be tricky
when designing a complex study or reading an academic research paper.
• Recognizing independent variables
• Is the variable manipulated, controlled, or used as a subject grouping method by
the researcher?
• Does this variable come before the other variable in time?
• Is the researcher trying to understand whether or how this variable affects
another variable?
• Recognizing dependent variables
• Is this variable measured as an outcome of the study?
• Is this variable dependent on another variable in the study?
• Does this variable get measured only after other variables are altered?

Does this variable come before the other variable in time?

DEPENDENT AND INDEPENDENT VARIABLES
• For most purposes, we are interested in a possible causative effect
within the relationship: if happiness is effected by gender, then the
corresponding variables should show a relationship or association
between them. The causing variable (or independent variable) and
the effected variable (dependent variable).
• The normal convention is to use the rows to denote the independent
variable (gender) and the columns to denote the dependent variable
(happiness), but sometimes this convention is reversed.
Cross Tabulation (CROSSTABS)
• The CROSSTABS procedure enables us to combine the two conditional
distributions into one table. Each row/column of the table represents
one conditional distribution.
• The ROWS correspond to the categories of the first variable (sex), the
COLUMNS to the categories of the second (happiness). Each cell
represents the simultaneous occurence of one category from each
variable: e.g. female who are not too happy define a cell with 112
cases and 12.9% of all female.
Interpreting association in crosstabs
• Association: look at the configuration of cells with the highest
percentage in every row. When the configuration is not vertical-
straight or random, then we can assume that there is an association
or difference between variables.
• Direction of the association. For ratio, interval, and ordinal variables
we can assume the direction of the association (positive or negative)
when the configuration of the cells with the highest percentage in
table form a diagonal line. This interpretation is not apply for nominal
variable.
Exercise Respondent's Sex * General Happiness Crosstabulation

General Happiness
Not Too
Very Happy Prett y Happy Happy Tot al
Respondent 's Male Count 206 374 53 633
Sex % within
32. 5% 59. 1% 8.4% 100. 0%
Respondent 's Sex
Female Count 261 498 112 871
% within
30. 0% 57. 2% 12. 9% 100. 0%
Respondent 's Sex
Tot al Count 467 872 165 1504
% within
31. 1% 58. 0% 11. 0% 100. 0%
Respondent 's Sex

Respondent's Sex * I s Life Exci ting or Dull Crosstabulation

Is Lif e Exciting or Dull

Exciting Rout ine Dull Tot al
Respondent 's Male Count 213 200 12 425
Sex % within
50. 1% 47. 1% 2.8% 100. 0%
Respondent 's Sex
Female Count 221 305 29 555
% within
39. 8% 55. 0% 5.2% 100. 0%
Respondent 's Sex
Tot al Count 434 505 41 980
% within
44. 3% 51. 5% 4.2% 100. 0%
Respondent 's Sex