To learn efficiently, many situations require people to judge what will be easy or difficult to learn, or how well it has been stored in memory. These metacognitive judgments are important to understand because they most likely guide how people behave when they learn, and consequently how much they learn. In this thesis, I focus on what is referred to as ease-of-learning (EOL) judgments, that is judgments about how easy or difficult a material will be to learn. EOL judgments have received relatively limited attention in the metacognitive literature. Therefore, this thesis also considers for comparison the more extensively researched judgments of learning (JOL), which are judgments of how well a studied material has been learned or how likely it is to be remembered on a later memory test. I had two major aims with my research. First, I aimed to investigate how accurate EOL judgments are, that is, how well they can predict the ease of future learning, and what moderates this accuracy. More precisely, I investigated what affects EOL judgment accuracy by varying how much an item-set varies in a predictive item characteristic, as well as varying methodological aspects of the judgment situation. The second major aim was to investigate what sources of information people use to make EOL judgments and how the information is used to make metacognitive judgments. In three studies, participants made EOL judgments for word pairs (e.g., sun – warm), or single words (e.g., bucket), studied the items, and tried to recall them on memory tests. In Study II, participants also made JOLs after studying the items. To estimate the accuracy of the judgments, the judgments were correlated with recall performance on memory tests. The results of the thesis show that EOL judgments can be accurate when they are made on a to-be-learned material which varies in a predictive item characteristic (Study I and II). In some conditions, EOL judgments are even as accurate as JOLs (Study II). Study II also supports the cue competition hypothesis, which predicts that, when people judge memory and learning, they sometimes rely less on one source of information if other information is available. Furthermore, Study III shows that processing fluency (the experience of effort associated with processing information), may be an important source of information for EOL judgments, and that people’s beliefs about available information can moderate how the information is used to make EOL judgments. Overall, the results show when EOL judgments will be accurate and when they will not, and provides evidence that people may use processing fluency to make EOL judgments even when it contradicts their beliefs. Importantly, the results also indicate that when multiple sources of information are available, information may compete for influence over metacognitive judgments.