Describing a numeric variable. ANOVA | Other Commands | How to... |
means variable1 [/by=variable2][/Q or /Quiet]
Command abbreviated form:
mea
The MEANS command does a basic descriptive statistic for numeric variable1, stratified by variable2...
If the name of a second field is provided (variable2) ,means will also do
an analsysis of varianza , treating all other variables as factors
Both variables must be numeric (*)
(*) This will changed in future versions. Then only the first variable must be numeric
Output:
Age:
SEX1 Obs. Sum Mean Variance Std Dev ( 95% CI mean ) Std Err
1.000000 31 1142.00 36.84 581.94 24.12 27.99 45.69 4.33
6.000000 44 1619.00 36.80 385.93 19.65 30.82 42.77 2.96
SEX1 Minimum p5 p10 p25 Median p75 p90 p95 Max
1.000000 3.00 5.40 7.20 15.00 36.00 62.00 69.60 75.20 77.00
6.000000 8.00 9.50 11.50 16.25 35.50 53.75 62.50 68.00 72.00
Source SS df MS F p Value
Between 0.03 1 0.03 0.00 0.993
Within 34053.35 73 466.48
Total 34053.39 74 460.18
Bartlett's test for homogeneity of variance Chi2= 1.489 df(1) p= 0.222
If only variable1 is specified a descriptive summary of this variable is given plus a test to evaluate whether the means was 0.
If two variables name are specified a descriptive summary of variable1 for each value of variable2 is displayed.
Variability between groups, within groups and total are given.
The real output can be different depending on the options used (see below)
MEANS AGE (A descriptive summary of AGE for all records in the data set not afected for the select statement (if any))A cross tabulation of SEX by ILL) MEANS BP /by=SMOKE (A comparative analysis of variable BP stratified by SMOKE)Notice:
/Q or /Quiet | Do not show tablas at all |
$variable# $level# $Obs# $Sum# $mean# $var# $sd# $cfiL# $cfiH# $stdErr# $min# $p05# $p10# $p25# $p50# $p75# $p90# $p95# $max# |
Name of variable of stratification Value # of stratifiying variable (1 for the first value in alpha-numeric order) Number of observations (records) with value # in variable 2 Sum of variable1 for records with value # in variable 2 Average of variable1 for records with value # in variable 2 Variance of variable1 for records with value # in variable 2 Standar deviation of variable1 for records with value # in variable 2 Lower limit of the 95% IC of variable1 for records with value # in variable 2 Higher limit of the 95% IC of variable1 for records with value # in variable 2 Standar Error of variable1 for records with value # in variable 2 Minimun value of variable1 for records with value # in variable 2 Percentile 5 of variable1 for records with value # in variable 2 Percentile 10 of variable1 for records with value # in variable 2 Percentile 25 of variable1 for records with value # in variable 2 Percentile 50 (median) of variable1 for records with value # in variable 2 Percentile 75 of variable1 for records with value # in variable 2 Percentile 90 of variable1 for records with value # in variable 2 Percentile 95 of variable1 for records with value # in variable 2 Maximun value of variable1 for records with value # in variable 2 |
set table design stat =line[box][filled][system] set table design summary =line[box][filled][system] |
Main tables design will be acording with this command. The default value is line Summary table design will be acording with this command. The default value is line T-Student or Barlett's test will be included if option /T is added |
No specific references |