Time
series analysis is a type of longitudinal data analysis that deals with time-bound events such as repeated measures in time.
Such longitudinal data can be summarized graphically by using a time series plot of y against time.

Three
types of patterns can be described: time trend, seasonal pattern, and random or irregular patterns, A pattern may be a mixture
of more than one of those described above.

_{ }A scatter-plot of data against
its immediate predecessor is another graphical way of identifying trends in time.

Moving
averages may be used instead of raw scores to make the time series plot more stable.

Time
series plots show trend and can be used for forecasting.

** **

**2.0 FORECASTING**

Forecasts can be qualitative or quantitative. Quantitative forecasts use the time series graph for
extrapolation.

** **

**3.0 CORRELATION AND REGRESSION MODELS**

LONGITUDINAL REGRESSION MODELS

In a longitudinal regression model, time becomes the independent or ‘x’
variable. The regression equation is like the usual linear regression equation.

AUTOREREGRESSION

Auto-regression is when a regression model is used to relate a variable to
its immediate predecessor. If they are related we know that there is a time trend.

AUTO-CORRELATION

Autocorrelation is correlation between a variable and its immediate predecessor. If there is correlation
we know that there is a time trend.