When assigning reports in a “Japanese Linguistics” liberal arts course, the author explains and demonstrates to students the format of a thesis paper and instructs them to write reports according to this format. It was found that 12.1% of reports submitted in 2007 lacked thesis-paper structure and were similar to essays. The author hypothesized that those students who were able to properly write a problem proposal as a result of instruction conducted in 2007 would be able to write reports with structures similar to thesis papers. Therefore, in instruction from 2008 to 2010, the author limited all students to a single problem proposal and had them write reports. In 2008, 6.7% of reports lacked an adequate thesis-paper structure. In 2009, the percentage dropped to 6.5%, and to 5.7% in 2010. In order to confirm that this decrease did not occur by chance, the changes in percentages were tested by normal distribution. Comparing 2007 and 2008, the test statistic was Z≒1.79. For 2007 and 2009, it was Z≒2.01. For 2007 and 2010, it was Z≒2.32. In all cases, the values exceeded the one-tailed 5% test index of 1.65 and did not exceed the one-tailed 1% test index of 2.33. Therefore, we can roughly characterize the effects by saying that the percentage of reports lacking structure decreased as a result of the author’s instructional approach.