[Goose]

Even if it's just a pictorial or graphical presentation of valid results for publication, if there is time or other pressure such that someone would feel a need to substitute another image from some other part of the research that shows the same thing, there has to be a culture such that bringing the problem to light so the team or team leader can address it feels better to the pressured minion than doing the wrong thing.

I basically agree that the culture is important. Rigor and honesty are values that need to be emphasized. That said, not all "pressure" is induced by the PI. Grad students/postdocs want to finish/get good recommendations. They have agendas too. First much scientific research/paper publication is "interesting" but not spectacularly important. It is very, very probable that nobody will ever repeat your experiment or rely on your result. Probably nobody will even scrutinize your publication very closely. You also aren't building a product based on this result so you won't get caught that way either. You will get a publication, the PI gets a publication, nobody cares about the result, you don't get caught. Even in the very, very improbable event somebody does try to repeat your results and fails, you will be long gone. Nobody can "prove" anything. In most cases, nobody will even try. You can just claim the group who tried to duplicate your result just has "bad hands" :). In short, you will get away with it.

Second, much research is done by individuals. It isn't a "team". Nobody else sees the results as a matter of course. Even when no fraud is intended, anomalous results are often neither recorded nor reported. You only have to make it "work" once. On top of that, lab notebooks in "academic research" settings are notorious for being incomplete, inaccurate,  and not living up to "standards". This is an area where PIs often talk a good game but then don't follow up. They leave it to their "senior minions" to do so. Shouldn't be the case but almost always is. You probably won't get caught by somebody reading your notebook and finding time gaps and results missing that should be there. Again, you will get away with it.

Some people when confronted with a situation where 1) they won't get caught, 2) no obvious material harm will ensue to others, and 3) it will make said people's lives MUCH easier, will "cheat". This may be the ultimate example of the statement "Morality is what you do when nobody is looking". Not sure this is easy to change.

[Goose]

Even if it's just a pictorial or graphical presentation of valid results for publication, if there is time or other pressure such that someone would feel a need to substitute another image from some other part of the research that shows the same thing, there has to be a culture such that bringing the problem to light so the team or team leader can address it feels better to the pressured minion than doing the wrong thing.

I basically agree that the culture is important. Rigor and honesty are values that need to be emphasized. That said, not all "pressure" is induced by the PI. Grad students/postdocs want to finish/get good recommendations. They have agendas too. First much scientific research/paper publication is "interesting" but not spectacularly important. It is very, very probable that nobody will ever repeat your experiment or rely on your result. Probably nobody will even scrutinize your publication very closely. You also aren't building a product based on this result so you won't get caught that way either. You will get a publication, the PI gets a publication, nobody cares about the result, you don't get caught. Even in the very, very improbable event somebody does try to repeat your results and fails, you will be long gone. Nobody can "prove" anything. In most cases, nobody will even try. You can just claim the group who tried to duplicate your result just has "bad hands" :). In short, you will get away with it.

Second, much research is done by individuals. It isn't a "team". Nobody else sees the results as a matter of course. Even when no fraud is intended, anomalous results are often neither recorded nor reported. You only have to make it "work" once. On top of that, lab notebooks in "academic research" settings are notorious for being incomplete, inaccurate,  and not living up to "standards". This is an area where PIs often talk a good game but then don't follow up. They leave it to their "senior minions" to do so. Shouldn't be the case but almost always is. You probably won't get caught by somebody reading your notebook and finding time gaps and results missing that should be there. Again, you will get away with it.

Some people when confronted with a situation where 1) they won't get caught, 2) no obvious material harm will ensue to others, and 3) it will make said people's lives MUCH easier, will "cheat". This may be the ultimate example of the statement "Morality is what you do when nobody is looking". Not sure this is easy to change.