📸 ARIMA / Time Series Forecasting menu in SPSS
What Is ARIMA?
ARIMA (AutoRegressive Integrated Moving Average) models and forecasts univariate time series data. ARIMA(p,d,q) notation: p=autoregressive terms, d=differencing order to achieve stationarity, q=moving average terms.
Applications: economic forecasting, hospital admissions trends, environmental monitoring, sales projections.
Running ARIMA in SPSS
Step 1: Define time variable: Data → Define Dates (select Year, Month, etc.).
Step 2: Analyze → Forecasting → Create Models.
Step 3: Add the time series variable to Dependent Variables. Method: Expert Modeler (SPSS automatically selects best ARIMA).
Step 4: Statistics: Goodness of fit, Ljung-Box Q. Save: Predicted values → OK.
📸 ARIMA model statistics — goodness of fit summary
APA Reporting
An ARIMA(1,1,1) model was identified via SPSS Expert Modeler (BIC=-184.2). The Ljung-Box Q test confirmed that residuals were white noise (Q=12.3, p=.412), indicating adequate model fit. The model was used to generate 6-month forecasts with 95% confidence limits.