Prediction · Exercise 16

contains real data for the first two decades of flu cases reporting.
\begin{array}{|c|c|c|}\hline \text{Year} & \text{# flu cases diagnosed} & \text{# flu deaths} \\ \hline \text{Pre-1981} & \text{91} & \text{29} \\ \hline \text{1981} & \text{319} & \text{121} \\ \hline \text{1982} & \text{1,170} & \text{453} \\ \hline \text{1983} & \text{3,076} & \text{1,482} \\ \hline \text{1984} & \text{6,240} & \text{3,466} \\ \hline \text{1985} & \text{11,776} & \text{6,878} \\ \hline \text{1986} & \text{19,032} & \text{11,987} \\ \hline \text{1987} & \text{28,564} & \text{16,162} \\ \hline \text{1988} & \text{35,447} & \text{20,868} \\ \hline \text{1989} & \text{42,674} & \text{27,591} \\ \hline \text{1990} & \text{48,634} & \text{31,335} \\ \hline \text{1991} & \text{59,660} & \text{36,560} \\ \hline \text{1992} & \text{78,530} & \text{41,055} \\ \hline \text{1993} & \text{78,834} & \text{44,730} \\ \hline \text{1994} & \text{71,874} & \text{49,095} \\ \hline \text{1995} & \text{68,505} & \text{49,456} \\ \hline \text{1996} & \text{59,347} & \text{38,510} \\ \hline \text{1997} & \text{47,149} & \text{20,736} \\ \hline \text{1998} & \text{38,393} & \text{19,005} \\ \hline \text{1999} & \text{25,174} & \text{18,454} \\ \hline \text{2000} & \text{25,522} & \text{17,347} \\ \hline \text{2001} & \text{25,643} & \text{17,402} \\ \hline \text{2002} & \text{26,464} & \text{16,371} \\ \hline \text{Total} & \text{802,118} & \text{489,093} \\ \hline\end{array}

Do you think a linear fit is best? Why or why not?

Its source prints no answer for this problem, so there's nothing to check against.

OpenStax, Introductory Statistics 2e · §12.5 Prediction · Exercise 16