Quantcast

problems generated weighted analysis

classic Classic list List threaded Threaded
1 message Options
Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

problems generated weighted analysis

Debbie Hahs-Vaughn
I am stumped and hope that someone can offer some assistance.

I am trying to generate a weighted ANCOVA.  The outcome is a continuous
variable (minimum value 13.039 and maximum value 88.994), the independent
variable is dichotomous (0, 1), and the covariate has five levels.  A cross
tab of the independent variable by covariate shows that the smallest cell
size is 11.

The weighted sample size is 1460. There are some weights that are very
close to zero (e.g., less than .04), however it's less than 1% of the
cases.  When I generate the ANCOVA, the results produce a sample size of
only TWO for one of the groups (group A) of the independent variable (there
should be 187) and only 208 for the second group (group B) (there should be
1272).

In trying to problem solve, I ran a simple t test.  The t test produced a
sample size of 188 for group A and 1272 for group B.  I also ran a one-way
ANOVA, and it produced a sample size of 188 for group A and 1272 for group
B (see syntax below).  All the results for the t test and one-way ANOVA
were the same.

ONEWAY
  outcome BY group
  /STATISTICS DESCRIPTIVES
  /MISSING ANALYSIS .


I then ran ANOVA using the general linear model and it produced a sample
size of 2 for group A and 208 for group B (see syntax below).

UNIANOVA
  outcome  BY group
  /METHOD = SSTYPE(3)
  /INTERCEPT = INCLUDE
  /PRINT = DESCRIPTIVE
  /CRITERIA = ALPHA(.05)
  /DESIGN = group .


Any thoughts on what's happening and suggestions on what to do? Again, my
goal is to generate an ANCOVA.   Thanks in advance.
Loading...