Since this thread hasn't been answered for a day, I'll give it a go. I will presume the figures are
not deseasonalised. As a result, this is actually a pretty big question so I'll give you a run down on how to actually do the problem. Any questions, just ask.
1. You need to find the quarterly averages for 1998. As an example, I'll do it:
Quarterly Average =

Quarterly Average =

2. Next, you need to calculate the seasonal index. The formula is:
SI =


SI
quarter 1 1998 =

3. Repeat step 2 for quarter 2, 3 & 4 for the rest of 1998.
4. When you add up all the SI's, they should equal to 4 since you have 4 "seasons".
5. Repeat steps 1-4 for 1999 and 2000.
6. Once you have all the relevant seasonal indices, you are now ready to calculate the deseasonalised figures. The formula is:
Deseasonalised figure =

For example, the deseasonalised figure for quarter 1 of 1998 would be:
Deseasonalised figure =

7. Repeat step 6 for all the quarters in every year.
8. Now that you have the deseasonalised figures, you are ready to plot the graph. As you know on the y-axis you have unemployment and on the x-axis you will have the number of quarters. As I have a TI-89, I will be using Stats/List. Enter the values in, graph it and using the least squares regression line, I can find the equation in terms of unemployment and quarters.
9. Use the equation to predict what will happen in quarter 1 of 2001. This means that you will have to sub in 13 in the equation because quarter 1 of 2001 is quarter 13, so to speak.