Table 4 Characteristics of the basic methodological considerations (autocorrelation, non-stationarity, seasonality) (Only for ITS with aggregated unit)
Characteristics | n | % |
|---|---|---|
Considered all of three methodological issues (autocorrelation, non-stationarity and seasonality) (N = 149) | ||
 Yes | 14 | 9.4 |
Considered at least one of three methodological issues (N = 149) | ||
 Yes | 117 | 78.5 |
Autocorrelation | ||
Autocorrelation acknowledged (N = 149) | ||
  Yes | 108 | 72.5 |
Autocorrelation acknowledged (ITS study used ARIMA model) (N = 24) | ||
 Yes | 20 | 83.3 |
Autocorrelation acknowledged (ITS study used Non-ARIMA model) (N = 125) | ||
 Yes | 88 | 70.4 |
Autocorrelation identified methods (N = 108) | ||
 Durbin Watson test | 40 | 37.0 |
 ACF | 9 | 8.3 |
 Cumby-Huizinga test | 5 | 4.6 |
 Ljung-Box2 test | 3 | 2.8 |
 Others a | 4 | 3.7 |
 Not reported | 47 | 43.5 |
Adjusted for autocorrelation (N = 108) | ||
 Yes | 65 | 60.2 |
 No adjustment for autocorrelation (after statistical test) | 16 | 14.8 |
 Unclear | 27 | 25.0 |
If yes, which method was used? (N = 65) | ||
 ARIMA | 31 | 47.7 |
 GLS | 16 | 24.6 |
 OLS with Newey-West standard errors | 12 | 18.5 |
 Add lag terms | 5 | 7.7 |
 Generalized Estimating Equation | 1 | 1.5 |
Non-stationarity | ||
Non-stationarity acknowledged (N = 149) | ||
  Yes | 20 | 13.4 |
Non-stationarity acknowledged (ITS study used ARIMA model) (N = 24) | ||
 Yes | 11 | 45.9 |
Non-stationarity acknowledged (ITS study used Non-ARIMA model) (N = 125) | ||
 Yes | 9 | 7.2 |
Non-stationarity identified methods (N = 20) | ||
 Augmented Dickey-Fuller test | 8 | 40.0 |
 Plot the raw data | 1 | 5.0 |
 Not reported | 11 | 55.0 |
Adjusted for non-stationarity (N = 20) | ||
 Yes | 15 | 75.0 |
 No adjustment for non-stationarity (after statistical test) | 4 | 20.0 |
 Unclear | 1 | 5.0 |
If yes, which method was used? (N = 15) | ||
 ARIMA | 11 | 73.3 |
 Others b | 4 | 26.7 |
Seasonality | ||
Seasonality acknowledged (N = 149) | ||
  Yes | 60 | 40.3 |
Seasonality acknowledged (ITS study used ARIMA model) (N = 24) | ||
 Yes | 15 | 62.5 |
Seasonality acknowledged (ITS study used Non-ARIMA model) (N = 125) | ||
 Yes | 45 | 36.0 |
Seasonality identified methods (N = 60) | ||
 Augmented Dickey-Fuller test | 4 | 6.7 |
 Plot the raw data | 3 | 5.0 |
 Others c | 6 | 10.0 |
 Not reported | 47 | 78.3 |
Adjusted for seasonality (N = 60) | ||
 Yes | 41 | 68.3 |
 No adjustment for seasonality (after statistical test) | 9 | 15.0 |
 Unclear | 10 | 16.7 |
If yes, which method was used? (N = 41) | ||
 Add seasonality terms | 18 | 43.9 |
 ARIMA | 13 | 31.7 |
 Fourier function | 8 | 19.5 |
 Othersd | 2 | 4.9 |