It can be seen that the relative proportion of representation of the participants in the North and West teams before and after did not change considerably, while the proportions in the South team increased, while that for the East team dropped.
However, as the following table of distributions show, the relative proportions of participants before and after “cleaning“ of data did not differ considerably:
Data Cleaning | Intervention | Control | Total |
---|---|---|---|
Original N (%) | 129 (24%) | 408 (76%) | 537 (100%) |
After | 83 (24.3%) | 258 (75.7%) | 341 (100%) |
As can be seen from the above table, the “data cleaning“ had little impact on the relative distribution of the intervention and control group. Hence, the following analyses will be presented on the “filtered“ data set.
The following table shows the number of HONOS assessments completed in different waves. Fewer HONOS completions were registered in Wave 1 and Wave 4.
Wave | N | Percent |
Wave 1 | 175 | 6.7 |
Wave 2 | 188 | 7.2 |
Wave 3 | 246 | 9.4 |
Wave 4 | 166 | 6.3 |
Wave 5 | 183 | 7 |
Wave 6 | 220 | 8.4 |
Wave 7 | 236 | 9 |
Wave 8 | 237 | 9.1 |
Wave 9 | 252 | 9.6 |
Wave 10 | 250 | 9.6 |
Wave 11 | 248 | 9.5 |
Wave 12 | 214 | 8.2 |
Total | 2615 | 100 |
My Question:
Is this pattern important? What does this variation of pattern of responses in HONOS completion tell us? Are there differences across intervention levels or regions? Should we be interested in this? Is this important?
In the next section, we review the results of the Generalising Estimating Equations (GEE) for this data, where the effects of the trend are examined along with exploratory data analysis with visual inspection of the trends.