I pulled these ropes. Then, based on your initial observations, you made a prediction about what would happen if we manipulated the system in a specific way. How accurate was your prediction?"
"We were right," Doug responds.
"And why were you able to predict what would happen before I pulled the rope?"
"I used what I observed in the first few pulls to help me predict what would happen later."
"Basically what each of you did was to speculate about how my tube was working on the basis of some limited observations. Scientists do that type of thing all the time. They make observations and try to explain what's going on, or sometimes they recognize that more than one explanation fits their data. Then they try out their proposed explanations by making predictions that they test. At first I had you draw a picture of how you thought my tube worked and had you each explain your picture. You got to hear each other's view on how the system worked. Doug, did you change your ideas at all based on what you heard from Karen?"
"Well, yes. I first thought that ropes A and C were the two ends of the same rope and B and D were two ends of another rope. Karen had A and B as ends of the same rope and C and D as ends of another rope, and her explanation seemed to fit better than mine."
"Right. Communication about observations and interpretations is very important among scientists because different scientists may interpret data in different ways. Hearing someone else's views can help a scientist revise his or her interpretation. In essence that was what you were doing when you shared your diagrams. Karen, when your model didn't work, what did you do?"
"All I did was adjust the length of one rope, and then it worked fine."
"So as a result of your formal testing of the predictions from your model, you revised your explanation of the system. Your understanding improved. In scientific terms, you revised your model to make it more consistent with your further observations. In science, the validity of any explanation is determined by its coherence with observations in the natural world and by its ability to predict further observations."
"But we still have different models," Karen observes. "How do we know which one is right?"
Doug says: "You told us that, didn't you, Barbara. There can be two possible explanations for the same observation."
"So it's possible for scientists to disagree sometimes," says Karen. "But does that mean that we don't understand evolution because scientists disagree about how evolution takes place?"
"Not at all," Barbara answers, "you both created different models of my tube, but both of your models are fairly accurate. And don't forget there were constraints on