Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Statscraft 0.7.0, release v0.7.0

PARTIAL CORR

Partial correlations of variables controlling for others, with significance.

Syntax

PARTIAL CORR [VARIABLES=]varlist [WITH varlist] BY controls [(order ...)] [/...]
             [/SIGNIFICANCE=TWOTAIL | ONETAIL] [/STATISTICS=NONE CORR DESCRIPTIVES BADCORR ALL]
             [/FORMAT=MATRIX | CONDENSED | SERIAL] [/MISSING=LISTWISE | ANALYSIS  EXCLUDE | INCLUDE].
A list alone gives a square matrix, WITH a rectangular one. Without orders the partials control
for every control variable at once; an order k gives one block per set of k controls (up to five
orders, 25 lists). Coefficients come from the zero-order Pearson correlations step by step, and are
tested with t on M - k - 2 df, M the smallest weighted count of the correlations used. A coefficient
that cannot be computed is empty. CORR shows the zero-order correlations first, BADCORR only when one
of them is empty; DESCRIPTIVES the mean, standard deviation and N. CONDENSED shows the coefficients
alone, marked * at 0.05 and ** at 0.01; SERIAL shows each pair once. LISTWISE (default) uses the cases
valid on every variable of the list, controls included; ANALYSIS each pair's valid cases. Weighted
when a weight variable is set. Not supported yet: MATRIX.

Example

PARTIAL CORR /VARIABLES=satisfaction recommend BY age /STATISTICS=CORR /SIGNIFICANCE=ONETAIL.

statscraft help PARTIAL CORR prints the same text in the terminal. All commands