ally ended up as CS majors. But the CS1 students reasonably represent the students available as input to the program. A finding that students late in the curriculum agree less with faculty could be due to many factors but likely shows that the output from the program is not improved with respect to the available input—a challenge to faculty charged with managing the program. The table does not present student responses about how faculty would like them to respond to each item. In some cases we bring this information into the discussion of individual items.

Among the 27 items for which we found strong faculty consensus, seniors were generally in line with faculty (75% or more agreement) on only seven items. For 16 items, we found mixed agreement (50%–75%), and for four items we found less than 50% of seniors endorsed the faculty position. Though this level of agreement was not good, it is better than for CS1 students. For 22 of the 27 items, we found seniors were more likely to agree with the faculty consensus than CS1 students. For seven of the items the difference in agreement was statistically significant. Among the five items on which seniors agreed less with the faculty consensus than CS1 students, the difference was statistically significant for one item ( 44).

Results by subcluster and theme

We now examine the data for the thematic groups of items and the subclusters to see what it says about student beliefs and how these beliefs differ between beginning and more advanced students.

Don’t learn just from examples. In this statistical cluster, seniors agreed with faculty that there is more to CS than memorizing solutions or learning a single best approach to problems. They agreed less strongly that reading is important, and it may be they value reading less than, say, CS1 students.

The end justifies the means. Most
seniors agreed with faculty that how
a program is written is important
(item 52) and were significantly more
likely to share this attitude than CS1
students. This is good news. But only
68% of seniors agreed with faculty that
doing things right is more important
than just getting a solution, though
this response represents a significant
improvement over the position of CS1
students. Less than half of the seniors
felt that how they do the work on an
assignment is more important than
getting “the desired answer” (item
20). This is a little better, though not
much, than the response from CS1
students. Most seniors (76%) correctly
indicated that faculty would disagree
with the item, so this was not a case
of students not knowing where faculty
stood on an issue.