Let

1

• Lets continue to review some of the statistics
  youve learned in your first class
• Bivariate analyses (two variables measured at a
  time on each observation) are done both
  graphically and numerically. There are 3 ways to
  look at the bivariate analyses
• categorical vs categorical (use PROC FREQ and
  PROC GCHART)
• categorical vs quantitative (use PROC
  UNIVARIATE with the BY statement on the
  categorical variable - also PROC BOXPLOT and PROC
  TTEST NPAR1WAY)
• quantitative vs quantitative (use PROC GPLOT,
  PROC CORR, PROC REG)

2

• / for two categorical variables /
• PROC FREQ TABLES marshform run quit
• PROC GCHART VBAR marsh/ groupform run quit
  / what does the next one do? /
• PROC GCHART VBAR marsh/ sumvarplantht typemean
  groupform run quit
• / now one categorical, one quantitative /
• PROC SORT BY marsh
• PROC UNIVARIATE DATApadgett PLOT VAR plantht
  BY marsh run quit
• / this gives side-by-side BOXPLOTS - can also
  use PROC BOXPLOT after SORTing /
• PROC BOXPLOT DATApadgett
• PLOT planthtmarsh run quit

3

• Now what about two quantitative variables??
• / first plot the two /
• PROC GPLOT DATApadgett
• PLOT totmassplantht run quit
• / get the correlation coefficients /
• PROC CORR run quit
• /this gives descriptive statistics and
  correlations of all numeric statistics/
• PROC REG MODEL totmass plantht
• PLOT totmassplantht run quit / there are
  many options to REG /