Visual Binning
This facility lets you interactively create groups (bins, categories) from a continous variable and visually control the process.
The example below "bins" urb (% of people living in cities in various
countries, i.e. creates a categorical variable from it.
- On the first dialogue specify the variable(s) you want to recode ("bin")
- Then the following dialog comes up:
- Specify a name for the Binned variable; SPSS
already supplies a default Label, but you can
of course change it.
- You now could enter (Grid) cutpoints, i.e.
the (upper) limits that separate two intervals (bins) and the
corresponding label. (HIGH means
the highest value, in our case a value of 100.0).
- If you want to split the variable regularly,
use Make Cutpoints:
- Equal Width Intervals defines
evenly spaced intervals. Fill in either the width or the
number of intervals.
- Equal percentiles on scanned cases
defines groups/classes with roughly the same number of
observations. Fill in either the number of cutpoints or the
%width by entering the percentage of cases in each group.
For example if you indicate 3 cutpoints or
25% of the cases, you will create four groups containing each one fourth
of the observations.
- Cutpoints at mean/standard deviations: Define a
cutpoint at the mean and at ± 1,2 or 3 standard deviations.
- Then Apply
The Visual Binning panel now appears as shown: the bins (interval borders) have been filled in,
according to the rules you just specified. Note that HIGH simply means
up to the highest value of the variable.
It is strongly recommended to fill-in value labels for the intervals (the first
label has been added as an example.
If you wish you can also move bin borders interactively, by just grabbing and
moving the limit:
Related documents