These four fictional examples illustrate different ways of gathering information. They do not show who can run an experiment, who should take a job or how to identify either type.
Start with two colleagues in the same lab, working from the same set of inconsistent measurements.
INTP and ISTP: a replication fails, and which end to work back from
(Fictional scene) A colleague's published result will not reproduce, twenty per cent off. With the same method and the same rig in front of them, one approach re-derives the colleague's model to see whether a premise was assumed away without being stated; the other walks the written steps again and checks whether the actual state at each step matches the description. Under the same conditions, INTP and ISTP can both take either approach.
Re-deriving the model can uncover an unstated assumption, but the paper might omit a detail needed to test it. Checking the procedure can surface a difference in the setup, provided the written method is clear enough to compare. The colleagues could time-box both checks and compare what they found before choosing an explanation.
Record what was known, who was responsible for each part, what the procedure required and which evidence changed the conclusion. That helps explain the decision without assigning a type.
INTP and ISTP: teaching a colleague to rerun your code
(Fictional example) A colleague is taking over the analysis scripts. Both people have the same code, dataset and environment. One way to help is to explain how the modules fit together so the colleague can adapt the script. Another is to demonstrate a complete run so they can see the inputs, outputs and settings. Either type can teach both ways.
A walkthrough can get the colleague to a working result, while a discussion of the design can help them adapt the code later. Choose based on what they need to do on their own, how much time they have and whether the environment is working. Teaching style cannot distinguish INTP from ISTP.
INTP and ISTP: whether the outlier goes into the report
(Fictional example) One reading remains unusual after repeat measurement. One question is whether the reading fits a plausible mechanism. Another is whether the records document a change in the measurement procedure. Either type can ask both. The study’s prespecified rules govern how to handle and report the outlier.
A proposed explanation belongs in the report as a hypothesis until evidence supports it. An additional control may help if the outlier could change the result and the study permits the check. If the team cannot verify the cause, report that limitation and how it might affect the conclusion.
INTP and ISTP: the shared flat where the network keeps dropping
(Fictional example) At 8 p.m., a video call drops and the router light turns orange. A pattern-based approach asks whether several devices or neighbours lose service at the same times. An observation-based approach compares which devices are affected and what the router indicates when the connection drops. Either type can do both. The light alone cannot distinguish a provider issue from a local problem.
Reports from neighbours can suggest a wider problem. Comparing affected devices can narrow down a local issue. Both findings are clues that can guide the next check. Neither shows what type the person troubleshooting is.
Log the time, duration and device affected. That gives the provider concrete information to work with when you report the problem.