"interpolating f in x = −2,0,2"
To me that sounds like you should use f(x) at x = -2, 0, and 2, rather than x = -2:2, which is x = -2, -1, 0, 1, and 2.
format long
a = 1./2.80;
f = @(x) (8.*a.^3)./(x.^2+4.*a.^2);
p = polyfit([-2 0 2],f([-2 0 2]),2)
And don't forget about "What is the value of interpolation polynomial in x = 1?"
polyval(p,1)
% visual aid:
% rational function f:
plot(-2:0.01:2,f(-2:0.01:2))
hold on
% polynomial fit from x = [-2 0 2]:
plot(-2:0.01:2,polyval(p,-2:0.01:2),'--r')
% also plot the polynomial fit from x = [-2 -1 0 1 2], for reference:
p_old = polyfit(-2:2,f(-2:2),2);
plot(-2:0.01:2,polyval(p_old,-2:0.01:2),':m')
