"But I really can't remember doing it"

You may accept the experiments and still say that they don't describe you, because you genuinely can't remember. It's not that the picture is a bit faded. It's that when you reach for it, there's nothing there. Surely that's a real gap, and a real gap calls for a look.

Let me describe what van den Hout and Kindt think is going on, because it explains the gap. When you do something for the first time, you take in its particulars. The feel of the key, the sound of the click, the light on the door. That's perceptual processing, in their terms, and it's what makes a memory vivid. When you do the same thing again, it's familiar. The mind doesn't need the particulars any more, because it knows what a locked door is, and so it processes the act as a concept, "locked the door", rather than as an event with details. Familiarity pushes out perception. And a memory made of concept rather than perception feels thin. When you reach for it, you find "I locked the door" and no picture, and the absence of a picture feels like the absence of a memory.

Meredith Coles, Radomsky and Betty Horng found something in a real-stove study that fits this. After repeated checking, people relied more on a sense of knowing that they'd checked and less on remembering it. Knowing is "of course I did". Remembering is "I did, and here's the click". The repeated checkers had lost the second and kept the first. That's your experience at the lights. You know you turned the straighteners off. You can't remember it. The gap is the difference between those two, and it's a gap that checking makes, not one that checking fills.

The same study found something else worth having. The changes appeared after a relatively small number of checks on a single occasion. You don't need nineteen. A few will do it. So the person who checks the door three times in the morning is well into the range where the picture is going.

There's one more thing to notice about the gap. It's a gap in the memory of the check, not in the door. Accuracy in every one of these experiments was fine. People who couldn't picture the stove still knew, correctly, what state it was in. So when you say you can't remember, what's true is that you can't picture it. What isn't true is that you don't know. You know. You locked the door, the way you lock it every morning, and the knowing is the reliable part.