Time-series autoregressive integrated moving average (ARIMA) modelling results of the impact of austerity-related and prosperity-related events on four monthly suicide time series studied in Greece from 1983 to 2012
All suicides | Male suicides | Female suicides | All suicides + potentially misclassified suicides | |||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Abrupt permanent test | Gradual permanent test | Abrupt temporary test | Abrupt permanent test | Gradual permanent test | Abrupt temporary test | Abrupt permanent test | Gradual permanent test | Abrupt temporary test | Abrupt permanent test | Gradual permanent test | Abrupt temporary test | |||||||||
Size | Size | Growth rate | Size | Decay rate | Size | Size | Growth rate | Size | Decay rate | Size | Size | Growth rate | Size | Decay rate | Size | Size | Growth rate | Size | Decay rate | |
September 1997 | −0.42 | 0.26 | 0.42 | 2.21 | 0.77 | −0.16 | −0.13 | 0.44 | 0.17 | 0.99 | 0.15 | 0.13 | 0.13 | 0.07 | 0.95 | 0.58 | 0.51 | 0.37 | 0.08 | 0.62 |
June 2000 | −1.25 | 0.01 | 1.07 | 2.68 | 0.78 | 0.05 | 0.05 | 0.64 | 0.18 | 0.77 | 0.87 | −1.79 | 0.94 | 0.71 | 0.45 | 0.92 | 1.38 | 0.53 | 0.14 | 0.88 |
January 2002 | −0.44 | −0.51 | −0.18 | −0.38 | 0.83 | −0.37 | −0.77 | −0.46 | −6.62* | 0.84*** | −0.28 | −0.21 | 0.53 | 2.09 | 0.06 | −1.12 | −3.41 | 0.91 | 1.24 | 0.45 |
August 2004 | 1.01 | 0.51 | 0.52 | −2.67 | 0.42 | 0.71 | 0.27 | 0.67 | 1.33 | 0.82 | 0.30 | 0.23 | 0.59 | 0.24 | 0.74 | 0.13 | −0.16 | 0.64 | 0.33 | 1.01 |
October 2008 | 1.95 | 3.57 | 0.78 | 4.43 | 0.36 | 3.24*** | 5.30 | 0.65 | 2.01 | 0.03 | 1.23 | −1.63 | 0.32 | 0.43 | 0.68 | 0.14 | 0.07 | 0.82 | 0.04 | 1.12 |
March 2010 | 12.31 | 0.23 | 1.07 | 0.28 | 0.62 | −0.80 | −0.69 | 0.44 | 10.70 | 0.07 | −1.30 | −0.79 | 0.48 | 0.11 | 0.50 | −3.42 | −3.02 | 0.32 | 7.78 | 0.69 |
May 2010 | 0.04 | 0.10 | 0.59 | 1.50 | 0.75 | −2.08 | −1.51 | 0.35 | 0.15 | 0.69 | −0.82 | −0.23 | 0.78 | 0.61 | 0.53 | −4.62 | −1.87 | 0.61 | 0.63 | 7.28 |
May 2011 | 0.48 | 0.44 | 0.60 | 0.39 | 0.95 | −6.02 | −6.26 | 0.24 | 0.01 | 0.83 | 2.37* | 4.31*† | 0.80*† | 3.43 | 0.66 | −3.81 | −4.27 | 0.40 | 0.07 | 0.94 |
June 2011 | 11.20*** | 5.87 | 0.50 | −7.68 | 0.20 | 5.16*** | 10.42*† | 0.79*† | 11.86*† | 0.89*† | −3.19 | −5.99 | 0.67 | −5.68 | 0.22 | 10.20*** | 6.54 | 0.37 | −6.30 | 0.14 |
September 2011 | 1.24 | 1.30 | 0.21 | 7.33 | 0.59 | −5.35 | −3.06 | 0.75 | 8.83 | 0.12 | 1.31 | 1.83 | 0.55 | 0.40 | 0.89 | 1.46 | 1.65 | 0.45 | 8.98 | 0.47 |
February 2012 | 1.69 | 1.39 | 0.52 | −9.68 | 0.46 | 8.65 | −3.06 | 0.75 | −9.03 | 0.12 | 0.98 | 0.56 | 0.49 | 0.14 | 1.42 | 2.33 | 0.12 | 1.27 | −10.34 | 0.61 |
April 2012 | 4.64 | 2.39 | 0.42 | 0.16 | 1.70* | −9.03 | −9.18 | 0.04 | 9.81* | 0.78*** | 0.44 | 0.37 | 0.43 | 0.10 | −1.71 | 6.55 | 6.67 | 0.01 | −0.22 | 0.65 |
All suicide models: ARIMA(0,0,0)(0,1,1)12. Q(24 lags)=13.6. Male suicide model: ARIMA(0,0,0)(0,1,1)12. Q(24 lags)=21.6.
Female suicide model: ARIMA(1,0,1)(0,1,1)12. Q(24 lags=19.8. All suicides + potentially misclassified suicides model: ARIMA(1,0,1)(0,1,1)12. Q(24 lags)=16.2.
Size: numerator coefficient. Growth rate and decay rate: denominator coefficient. Abrupt permanent test: zero order transfer function applied to step.
Gradual permanent test: first order transfer function applied to step. Abrupt temporary test: first order transfer function applied to pulse.
*p<0.05; **p<0.01; ***p<0.001.
†The coefficient was statistically significant but produced a poorly fitting estimate of the time series and thus was rejected.