For , decide whether
converges absolutely, converges conditionally, or diverges.
Do not test the alternating part alone. The terms look alternating, and Leibniz's test tempts you immediately. But is not an odd function: and do not cancel, and the leftover bias is what actually decides this problem. So expand instead of testing.
Expand the logarithm. With and ,
The second term has no sign change at all — that is the hidden bias.
Handle the alternating piece. converges for every by Leibniz (the terms decrease monotonically to zero), and converges absolutely exactly when . So this piece never causes divergence.
Handle the quadratic piece — this is the decider. is a -series with ; it converges iff , i.e. . When this piece diverges to while the alternating piece stays bounded, so the whole series diverges.
Assemble the three regimes. The remainder is dominated by the term, so it never changes the verdict:
State the result.
The moral: for with alternating , always expand to second order before applying an alternating-series test.
Need to solve a different problem like this? Open the solver →