|
<< Click to Display Table of Contents >> Navigation: Reference Manual > Tables > Table statistics |
This document explains the table based statistics available.
See also Significance testing and distinction analysis in Data drill down window
If respondent weighting has been applied to a table, Companion maintains weighted and unweighted counts for all cells and is able to use the unweighted figure where appropriate when calculating statistics.
If you are confident your weighting procedures are relevant and correct then you can still produce meaningful statistics.
If quantity weighting has been applied or table arithmetic (other than addition/overlay) has been done, then a some of the statistics will not be reliable because the real sample sizes will not be known.
See Statistics formats for the relevant format options.
There are two sorts of questions that will produce various statistics when used as the rows of a table:
These are quantities: integer or float entries. For example salary, height or volume.
Any respondents with undefined values (U) are excluded from the statistics calculations.
If you only want to see the statistics you can use a "Stats only" table but this will not work for formats ILE, ILH, ILL, MED, and MOD. For these formats you can also use format NDIS.
The average value in each column will be shown by default, see format AVG.
These are usually single-coded rating scales to which score values have been assigned. For example:
Like a lot (2)
Like a little (1)
Indifferent (0)
Dislike a little (-1)
Dislike a lot (-2)
Don't know
These can be five or ten point scales when only the first and last scores have a response label, for example:
Agree totally (5)
(4)
(3)
(2)
Disagree totally (1)
Don't know
Only responses with a score value attached will be used in the calculations, other responses will be ignored.
A multi-coded entry can be used but respondents should only go into one of the responses with a score value, for example:
Like a lot (2)
Like a little (1)
Indifferent (0)
Dislike a little (-1)
Dislike a lot (-2)
Top 2 boxes (any like)
Bottom 2 boxes (any dislike)
Don't know
NOTE: response numbers are never used to calculate statistics, only the score values attached to the responses are used. To use response numbers copy to an integer variable.
The mean score in each column will be shown by default, see format AVG.
If you only want the statistics to be shown on the table you can use format NDIS.
The values to be used for statistics are usually shown in parentheses as shown above, see format PSV.
See Statistics formats for the relevant format options.
The following statistics may be produced:
Base for statistics, format BST
Sum of values, format SUM
Sum of squares of values, format SSQ
Average or mean score, format AVG
Standard deviation, format SDV
Standard error, format SER
Error variance, format EVR
Mean score divided by standard error, format MSE
The decimal places are controlled by formats DPA and DPS.
See Statistics formats for the relevant format options.
These formats are not available on a "Stats only" table.
Medians, format MED
Quartiles to Percentiles, format ILE
Maximum value, format ILH
Minimum value, format ILL
The calculations above assume that the rows are in ascending order.
IMPORTANT: You must always use format RNA when tabulating quantity questions with list all rows if you are using any of these formats.