Virata manual

Fitting models

Explaining one column with several others, whatever kind of column it is.

What we have

  • Explains a number using other columns, by fitting a straight line.
  • Explains a yes/no outcome.
  • Explains a yes/no outcome, with a slightly different curve.
  • Explains a count — how many times something happened.
  • Explains a count that is more spread out than Poisson allows.
  • Explains an outcome with ordered categories — poor, fair, good, excellent.
  • The same ordered-category model, with the bell-curve shape.
  • Explains a choice among categories with no order.
  • Draws the coefficients of the model you just fitted, with their error bars.

Coming later

  • NB1
  • exposure() and offset()
  • irr, to print incidence rate ratios
  • or, to print odds ratios instead of coefficients
  • predict and margins afterwards
  • predict, test, lincom and margins afterwards
  • rrr, to print relative risk ratios
  • vce(cluster ...), which is refused rather than quietly ignored
  • a margins plot, since margins itself is not built
  • a test of the proportional-odds assumption
  • frequency weights (fw), which are refused
  • plotting two models side by side