Hi there!
I am stuck on something that seems so simple, yet I cannot find a straightforward solution.
I am conducting analysis on some survey data. For each question, I have the number of respondents who selected each category. For example:
Q1.please indicate your primary area of COVID-19 response work during the past 6 months
Options = 1) communications 2) Case management 3)COVID line 4)Surveillance and so on.
I had coded these options on stata from 1-7 (as there were 7 options for question 1).
I simply want to compare the proportion of respondents in each category because I want to know if they are statistically different.
right now I have various format of this variable. I have its original format as q1code (coded 1-7). I also created dummy variable for each option so I have 7 more categorical variables coded as 0/1, where 1 is the actual count I am interested in. I tried running a chi square on each of these categorical variables =
tab2 q1code2 q1code1 q1code3 q1code4 q1code5 q1code6 q1code7, row col chi.
This will give me 2x2 tables for each combination of those variables. For example=
| q1code1
q1code==2 | 0 1 | Total
-----------+----------------------+----------
0 | 34 3 | 37
| 91.89 8.11 | 100.00
| 69.39 100.00 | 71.15
-----------+----------------------+----------
1 | 15 0 | 15
| 100.00 0.00 | 100.00
| 30.61 0.00 | 28.85
-----------+----------------------+----------
Total | 49 3 | 52
| 94.23 5.77 | 100.00
| 100.00 100.00 | 100.00
Pearson chi2(1) = 1.2907 Pr = 0.256
However, I am not interested in the "0" group. I want to know if 3/52 (5.77) is significantly different than 15/52 (28.85). I can do a prtest for each of these variables against each other but that seems very inefficient. I am hoping there is a simple solution to this.
Thank you!
Meera
I am stuck on something that seems so simple, yet I cannot find a straightforward solution.
I am conducting analysis on some survey data. For each question, I have the number of respondents who selected each category. For example:
Q1.please indicate your primary area of COVID-19 response work during the past 6 months
Options = 1) communications 2) Case management 3)COVID line 4)Surveillance and so on.
I had coded these options on stata from 1-7 (as there were 7 options for question 1).
I simply want to compare the proportion of respondents in each category because I want to know if they are statistically different.
right now I have various format of this variable. I have its original format as q1code (coded 1-7). I also created dummy variable for each option so I have 7 more categorical variables coded as 0/1, where 1 is the actual count I am interested in. I tried running a chi square on each of these categorical variables =
tab2 q1code2 q1code1 q1code3 q1code4 q1code5 q1code6 q1code7, row col chi.
This will give me 2x2 tables for each combination of those variables. For example=
| q1code1
q1code==2 | 0 1 | Total
-----------+----------------------+----------
0 | 34 3 | 37
| 91.89 8.11 | 100.00
| 69.39 100.00 | 71.15
-----------+----------------------+----------
1 | 15 0 | 15
| 100.00 0.00 | 100.00
| 30.61 0.00 | 28.85
-----------+----------------------+----------
Total | 49 3 | 52
| 94.23 5.77 | 100.00
| 100.00 100.00 | 100.00
Pearson chi2(1) = 1.2907 Pr = 0.256
However, I am not interested in the "0" group. I want to know if 3/52 (5.77) is significantly different than 15/52 (28.85). I can do a prtest for each of these variables against each other but that seems very inefficient. I am hoping there is a simple solution to this.
Thank you!
Meera
Comment