The thing about this, is what we cover observations and inferences in science so I'm always expecting my students will make the connection and a few of them do. I also know that we introduce students to this strategy in third grade, so it shouldn't be brand new. Nevertheless, each year, when I mention that we are going to begin to talk about inferences, I don't get the familiarity that I am hoping for, so I go back to the basics.
To begin my introduction, I use the chart above with Inference Riddles from Phil Tulga. He has a lot of great riddles on his site. I sometimes mix up the order to build from the less obvious to the totally obvious, and I add my own clues in here and there, but his site is a great place to get riddles so you don't have to make up your own. Here's an example of how we might use one:
Notice that I list all of the riddle clues along the left column. I do not list anything in the schema column until I have read all of the clues. This is because making inferences is about reading all of the clues, adding it to what you already know, and concluding the answer based on the combination of clues and your knowledge. Students are going to naturally guess based on the first clue, but those guesses are really more like predictions because they aren't putting all of the clues together yet and thus will end up with a lot of guesses that don't make sense as they add more clues. Imagine what students would think it was if the only clue were round. They could predict a ton of items, but they would be basing that on one clue. That's why, when we "picture walk", we ask students to predict. We are giving them limited information to "guess" about the story and then confirm that guess with a yes or no at the end. Now, don't get me wrong, I think we should encourage students to make predictions in their mind as they read, because we want them to use all of the strategies as they read. We just want them to understand that making an inference happens when they put all of the clues together, and students use their background knowledge (they love to say "schema") with those clues to make a decision about the text. A discussion about their background knowledge and how it may differ from others can be really helpful here. For example, if I lived in China, this could be a very challenging riddle for me because I won't have any schema about a penny. That is also why some questions that require a child to infer can be very tricky, as some students have much more limited background knowledge than others.
After students understand the process of making an inference, I apply our learning to our current read-aloud (not usually in the same day, but in the next day or two). My students always have their own copy of my read-aloud so that they can follow along or listen, and so that we can do guided practice of the strategies that they are learning with a familiar story.
We're currently reading my all-time favorite fifth grade read-aloud, Maniac Magee. Here are some of the inferences that students made as we read on our first day of applying our strategy. Please excuse the writing residue, as I'm in great need of a new eraser!
I was quite impressed by what they come up with for their first round and though it's not reflected in the chart, we talked about the fact that we want more than one text clue to make an inference (to again be sure that we are making an inference and not a prediction) and they did orally add more to their clues, I just ran out of room! I did need to guide them a bit more at first, but they are really getting the hang of it now and they are on their way to inference success!
If you'd like a copy of my inference chart, click on the picture below.