Hello,
We are currently analysing binge drinking consumption in a specific population (communaute==1) according to sex (v1). The unweighed analysis produces results (see below), however, the weighed chi2 test does not produce 95%CI nor a p-value (see second table).
tab binge_drinking v1 if communaute==1, chi2 row
+----------------+
| Key |
|----------------|
| frequency |
| row percentage |
+----------------+
Bingedrinking |
| sexe
| Boys Girls | Total
----------------------+----------------------+----------
No binge drinking | 145 175 | 320
| 45.31 54.69 | 100.00
----------------------+----------------------+----------
Binge drinking | 30 11 | 41
| 73.17 26.83 | 100.00
----------------------+----------------------+---------
Total | 175 186 | 361
| 48.48 51.52 | 100.00
Pearson chi2(1) = 11.2927 Pr = 0.001
***Weighed anlysis
svy, subpop (if communaute==1) : tab binge_drinking v1 , percent ci row
(running tabulate on estimation sample)
Number of strata = 1 Number of obs = 361
Number of PSUs = 9 Population size = 2,270.4188
Subpop. no. obs = 361
Subpop. size = 2,270.4188
Design df = 8
-------------------------------------------------------
Binge drinking
|
| sexe
| Boys Girls Total
----------+--------------------------------------------
no binge | 43.73 56.27 100
| [43.73,43.73] [56.27,56.27]
|
Binge dr | 72.75 27.25 100
| [72.75,72.75] [27.25,27.25]
|
Total | 46.6 53.4 100
| [46.6,46.6] [53.4,53.4]
-------------------------------------------------------
Key: Row percentage
[95% confidence interval for row percentage]
Pearson:
Uncorrected chi2(1) = 10.8840
Design-based F(., .) = . P = .
We've looked at the weights but they seem fine. We've tried to do this analysis for other variables with similar population numbers and weights and it works. Any ideas why we're having this issue only for one variable and how to resolve it ?
Many thanks,
Emma Holmberg
We are currently analysing binge drinking consumption in a specific population (communaute==1) according to sex (v1). The unweighed analysis produces results (see below), however, the weighed chi2 test does not produce 95%CI nor a p-value (see second table).
tab binge_drinking v1 if communaute==1, chi2 row
+----------------+
| Key |
|----------------|
| frequency |
| row percentage |
+----------------+
Bingedrinking |
| sexe
| Boys Girls | Total
----------------------+----------------------+----------
No binge drinking | 145 175 | 320
| 45.31 54.69 | 100.00
----------------------+----------------------+----------
Binge drinking | 30 11 | 41
| 73.17 26.83 | 100.00
----------------------+----------------------+---------
Total | 175 186 | 361
| 48.48 51.52 | 100.00
Pearson chi2(1) = 11.2927 Pr = 0.001
***Weighed anlysis
svy, subpop (if communaute==1) : tab binge_drinking v1 , percent ci row
(running tabulate on estimation sample)
Number of strata = 1 Number of obs = 361
Number of PSUs = 9 Population size = 2,270.4188
Subpop. no. obs = 361
Subpop. size = 2,270.4188
Design df = 8
-------------------------------------------------------
Binge drinking
|
| sexe
| Boys Girls Total
----------+--------------------------------------------
no binge | 43.73 56.27 100
| [43.73,43.73] [56.27,56.27]
|
Binge dr | 72.75 27.25 100
| [72.75,72.75] [27.25,27.25]
|
Total | 46.6 53.4 100
| [46.6,46.6] [53.4,53.4]
-------------------------------------------------------
Key: Row percentage
[95% confidence interval for row percentage]
Pearson:
Uncorrected chi2(1) = 10.8840
Design-based F(., .) = . P = .
We've looked at the weights but they seem fine. We've tried to do this analysis for other variables with similar population numbers and weights and it works. Any ideas why we're having this issue only for one variable and how to resolve it ?
Many thanks,
Emma Holmberg
Comment