We performed large-scale impression rating experiments to investigate the evaluation of the senses of polysemous words in example sentences. For the impression rating, we used 5,125 example sentences for 530 words in the IPA Lexicon of Basic Japanese Verbs, Adjectives, and Nouns (IPAL). In addition, we annotated the sense tags of Word List by Semantic Principles and their representative senses given by Yamazaki and Kashino (2017). A contrastive survey between the representative senses and impression ratings was performed. Further, linear regression analysis was performed to estimate the degree of representativeness and extract their typical usages. We found that differences in contextual words, including case particles, affect the recognition of word senses.