--Many other kinds of data benefit from log transformations:
Where 0 is the "bottom" and larger values can be thought of naturally
as multiples of smaller ones.
Where the histogram distribution is J-shaped, many observations at
small values and fewer and fewer at larger and larger values. E.g.
earthquake severity, national populations.
--We usually use log base 10, for ease in interpretation. Then
raw value log The leading
log digit tells what place
1-10 0-1
the leading raw digit takes.
10-100 1-2
100-1000 2-3
Categorical data: relations
It's very important when giving/reading "The percent is..." to know
exactly what group is on the bottom of the fraction. (What group
is on the top is usually clear, but the bottom is not always.) Example.
Jargon: Two-way table = contingency table = crosstab(ulation)
Marginal distributions: each variable separately, summing over
the other varable to get totals. (Usually expressed as proportions.)
E.g. Gender. Labor
Force Participation.
Conditional distributions: values for one variable, confined
to or "given" a particular value for the other.
E.g. For Women only, dist.
of Labor Force Particip. For Men only, dist. of Labor
Force Particip.
For people in labor force, dist of Gender. For people not in labor
force, dist. of Gender.
- - - - - - - - - - - - - - - - - - - - - - - - - -
Reading: Sections 2.5, exponential growth &
2.6, categorical variables. I will discuss Simpson's Paradox
in class.
| Hand in:
Sec. 2.5 exponential p.190 2.73 stocks Not SPSS--just calculator, or perhaps Excel For the following you will need to Transform your y-data to a new variable in SPSS. Use Transform>compute: Use the function LG10( ) for the log base 10. Graphing: you can graph the transformed variable, or in the Chart editor, Chart>Axis>y-axis>Scale:log will graph the original variable on the log scale.) 2.74 gypsy moths (do b. with calculator) 2.77 U.S. population Sec. 2.6 Categorical : Study "Monroe Comm. College" handout. BY HAND: (pp. 201-3) 2.81, 2.82, 2.83 2.87 Work through the SPSS manual, pp. 66-72, to see how to analyse predigested tables. The SPSS Data>Select cases (filtering) and Data>Weight cases choices are also useful generally. |
Read, discuss
2.89 cocaine relapse |
Optional |
| Sievers home | Math251-Fall01/DayP12.htm | 9pm | 9/25/01 |