Hi all,
I am working with a dataset of individuals surveyed twice following an intervention in three villages (one received no intervention, one received one intervention, the third received two interventions). Individuals are identified by caseID_num and time of survey (2016 or 2018) is indicated by the TimeSeries variable. The dependent measure, nQ19, is a count of the number of responses selected for Q19, "What ways do you protect yourself?", by each individual. Provided below is an example of the data.
The hypothesis is that the mean of nQ19 will be higher for the Villages that received interventions. To test this a one-way repeated measures ANOVA was run and here is the output from STATA
I am wondering why the margins and alternative epsilon's are not estimable (except box's conservative) and how to interpret the coefficients and p values for Village and caseid_num. Is two-way repeated measures ANOVA the incorrect analysis for this data?
Thank you in advance for taking the time to assist.
I am working with a dataset of individuals surveyed twice following an intervention in three villages (one received no intervention, one received one intervention, the third received two interventions). Individuals are identified by caseID_num and time of survey (2016 or 2018) is indicated by the TimeSeries variable. The dependent measure, nQ19, is a count of the number of responses selected for Q19, "What ways do you protect yourself?", by each individual. Provided below is an example of the data.
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input float nQ19 long Village float(caseid_num TimeSeries) 3 1 1 0 2 1 1 1 2 1 10 0 2 1 10 1 2 2 102 0 2 2 102 1 2 2 103 1 2 2 103 0 3 2 104 0 3 2 104 1 end label values Village village label def village 1 "Bugembe", modify label def village 2 "Kijinjomi", modify
Code:
. anova nQ19 Village caseid_num, repeated(Village)
Number of obs = 380 R-squared = 0.5531
Root MSE = 1.01177 Adj R-squared = 0.1085
Source | Partial SS df MS F Prob>F
-----------+----------------------------------------------------
Model | 240.70789 189 1.2735867 1.24 0.0668
|
Village | 2.3333333 2 1.1666667 1.14 0.3221
caseid_num | 235.04521 187 1.2569263 1.23 0.0798
|
Residual | 194.5 190 1.0236842
-----------+----------------------------------------------------
Total | 435.20789 379 1.1483058
Between-subjects error term: caseid_num
Levels: 190 (187 df)
Lowest b.s.e. variable: caseid_num
Repeated variable: Village
Huynh-Feldt epsilon = .
Greenhouse-Geisser epsilon = .
Box's conservative epsilon = 0.5000
------------ Prob > F ------------
Source | df F Regular H-F G-G Box
-----------+----------------------------------------------------
Village | 2 1.14 0.3221 . . 0.2884
Residual | 190
----------------------------------------------------------------
.
end of do-file
. margins Village
Predictive margins Number of obs = 380
Expression : Linear prediction, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Village |
Bugembe | . (not estimable)
Kijinjomi | . (not estimable)
Kyakabuzi | . (not estimable)
------------------------------------------------------------------------------
Thank you in advance for taking the time to assist.
Comment