Question 159471
A regression line is the best fit of a straight line using the data points.
It doesn't necessarily go through all or any of the points but fits the data best (least error at each of the data points).
A regression analysis uses the data to find the slope, m, of the line, and the y-intercept, b.
{{{y=mx+b}}}
.
.
.
Let's see if we can guess the solution without doing a full blown regression analysis. 
I'm not sure if that was the intent of the problem.
First look at the data,
{{{drawing( 300, 300, -2, 18, -5, 45,grid( 1 ),
circle( 12, 29, .2 ),
circle( 14, 35, .2 ),
circle( 10, 28, .2 ),
circle( 9, 23, .2 ),
circle( 15, 36, .2 ),
circle( 8, 18, .2 )
)}}}
You can guess the approximate slope by using the first and last points,
{{{m[apx]=(36-18)/(15-8)=2.57}}}
So the answer must be 2.4, closer to 2.4 than 1.2.
You know that {{{b=y[ave]-mx[ave]}}}
{{{x[ave]=(12+14+10+9+15+9)/6=11.333}}}
{{{x[ave]=(29+35+28+23+36+18)/6=28.167}}}
{{{b=y[ave]-mx[ave]}}}
{{{b=28.167-(2.4)(11.333)}}}
{{{b=28.167-27.2}}}
{{{b=0.97}}}
The closest answer is 1.2.
{{{y=1.2+2.4x}}}
.
.
.

The value for b was off because of roundoff error in m. 
The actual calculated value for m=2.3814, with b=1.178.