Unit Weights
Unit-weighted regression proceeds in three steps. First, predictors for the outcome of interest are selected; ideally, there should be good empirical or theoretical reasons for the selection. Second, continuous predictor variables are changed to Z scores. Third, the predictors are added together; the sum is called the variate. This variate is used as the predictor of the outcome, also expressed in z scores. The relationship of this variate to the outcome is assessed with the Pearson R correlation.
One small variation on unit-weighted regression is to make the weights not one, but one divided by the number of predictors. Thus, with three predictors, the weight of each variable is 1/3; with four predictors, the weight is 1/4; and so on. The value of this variation is that the variate is already in z score form.
A second variation occurs when predictors are binary. In this case, the predictors are scored as one (present) or zero (absent).
Read more about this topic: Unit-weighted Regression
Famous quotes containing the words unit and/or weights:
“During the Suffragette revolt of 1913 I ... [urged] that what was needed was not the vote, but a constitutional amendment enacting that all representative bodies shall consist of women and men in equal numbers, whether elected or nominated or coopted or registered or picked up in the street like a coroner’s jury. In the case of elected bodies the only way of effecting this is by the Coupled Vote. The representative unit must not be a man or a woman but a man and a woman.”
—George Bernard Shaw (1856–1950)
“What do you believe in?—In this, that the weights of all things must be determined anew.”
—Friedrich Nietzsche (1844–1900)