Everything from the 2 lessons of this topic on a few pages: the key idea of each lesson, the diagram to remember, the points to learn and the traps examiners set. Read it, cover it, test yourself.
Write the equation as \(f(x) = 0\). If \(f\) is continuous on \([a, b]\) and \(f(a)\) and \(f(b)\) have opposite signs, then there is at least one root between \(a\) and \(b\). To show a root is \(\alpha\) correct to 2 d.p., show the sign changes between \(\alpha - 0.005\) and \(\alpha + 0.005\).
In a sketch-and-count question the marks are for two correct graph shapes in the right places AND a stated conclusion ("one point of intersection, so one root"). A sketch with no statement does not earn the final mark.
Compare with zero, not with each other: you must rearrange to \(f(x) = 0\) first. Checking that \(\ln 2 < 3 - 2\) and \(\ln 3 > 3 - 3\) is fine, but only if you say clearly that the sign of \(\ln x - (3 - x)\) has changed.
No sign change does NOT mean no root. \(f(x) = (x-1)^2\) has a root at \(x = 1\) but is positive on both sides; and two roots between \(a\) and \(b\) cancel out the sign change. A sketch protects you from both traps.
If the sequence given by \(x_{n+1} = F(x_n)\) converges to \(\alpha\), then \(\alpha = F(\alpha)\): the limit is a root of the equation \(x = F(x)\). Iterate until successive values agree to the accuracy required, then give the root to that accuracy.
Keep \(x_{n+1}\) and \(x_n\) distinct while iterating, but set them BOTH equal to \(\alpha\) (or \(x\)) when finding the equation. Writing \(\alpha = \dfrac{2x}{3} + \ldots\) with a mixture of letters goes nowhere.
Do not round each iterate before feeding it back in: use the full calculator value (ANS). Rounding early can stall the sequence or change the final digit.
A start close to the root does not guarantee convergence: the rearrangement decides. In an exam, if your values run away, recheck the arithmetic and calculator mode (radians!) before concluding that the iteration fails.