An image question fails in a specific way. You look, something jumps out, you commit to it, and you never examine the rest of the picture. The finding that jumped out is usually the obvious one, and in a well-set question the obvious finding is frequently the distractor.
The fix is a routine you run every time, regardless of what you noticed first. It costs about ten seconds and it is the difference between reading an image and reacting to it.
A routine that transfers across image types
- Name what you are looking at before interpreting anything — modality, projection, stain, magnification, region. A surprising number of errors are answering correctly about the wrong kind of picture.
- Say the obvious finding out loud and then deliberately set it aside. You are not ignoring it; you are stopping it from monopolising the rest of your look.
- Sweep systematically. Whatever order you choose, use the same one every time. Consistency is what stops you skipping a region.
- Read the stem again with the findings in mind. The stem usually contains the one clinical detail that decides between two plausible readings.
- Commit, then check the options — not the other way round. Reading the options first invites you to find what they suggest.
The first thing you notice is what the examiner expected you to notice. It is rarely the whole answer.
The second kind of image question
There are two quite different uses of a figure, and they need different handling. In one, the picture illustrates the explanation — it helps you remember after the fact. In the other, the picture is the question: identify the arrowed structure, name the finding. For the second kind you cannot answer without reading the image properly, and a passing glance guarantees a guess.
Our Spot Diagnosis app deals only in the second kind, so you can practise the routine deliberately rather than meeting those questions at random.
Where a tutor helps, and where it does not
A tutor grounded in the question's own explanation can talk you through what the figure shows and why a particular feature is decisive. That is useful after you have committed to an answer.
What it should not do is look at an arbitrary image you upload and diagnose it. That is exactly the ungrounded use these guides keep warning about, and in radiology and pathology the failure is both easy to produce and hard for a candidate to catch.
What to practise on radiographs specifically
Most candidates look at a film and hunt for the abnormality. The more reliable habit is to establish the normal first: is the film adequate, is the patient rotated, is the exposure reasonable. Half the trick questions in image sets turn on a technical factor rather than a pathological one, and you cannot see that if you skipped straight to hunting.
Histopathology has a different failure mode
On a slide, the error is usually magnification. People commit at high power to something they would have read correctly at low power, because architecture is visible at low power and cellular detail is seductive at high. Establish the architecture first, then go closer to confirm — the reverse order produces confident, specific, wrong answers.
Clinical photographs and the anchoring problem
A photograph of a rash or a face triggers pattern recognition almost instantly, and once triggered it is very hard to dislodge. This is the strongest form of the anchoring problem, which is why the discipline of naming the obvious finding and then deliberately setting it aside matters most here.
Why figures make an explanation stick
A diagram gives a fact a second route into memory: you can retrieve the picture when the words will not come, and the spatial relationship often carries the reasoning rather than merely illustrating it. This is why almost every question in our bank carries a figure drawn for it, rather than a stock image attached afterwards — 12,607 of the current 12,607.
Practise on figures you have already seen
Reviewing an image you have met before is not wasted. The second look is when you notice the features you missed while you were busy being right for the wrong reason, and it is far quicker than a fresh question.