Survey instruments help us assess three aspects of the Software Component Engineering Course Sequence:
| Qualitative Response | Proportional
Weighting |
Threshhold
Weighting |
One-sided
Weighting |
|---|---|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The proportional weighting scheme tries to capture the extent to which the respondent disagrees or agrees with the statement. A reported average of, say, 80% for some statement means something like this: "On average, respondents moderately agreed with this statement." Other schemes that weight different responses with different numbers are possible, but this "linear" scale makes about as much sense as any.
The threshhold weighting scheme distinguishes between those who agree and those who disagree with the statement, without attempting to capture the extent of disagreement or agreement. A reported average of, say, 80% for some statement means: "80% of respondents agreed with this statement, and 20% disagreed with it."
The one-sided weighting scheme distinguishes between those who moderately or strongly agree and those who are close to neutral in their agreement or, in fact, disagree with the statement. A reported average of, say, 80% for some statement means: "80% of respondents moderately or strongly agreed with this statement." (Note that of the other 20% in this case, some slightly agreed and others disagreed.)
The proportional weighting scheme tends to smooth out differences in opinion, while the threshhold weighting scheme tends to emphasize them. The one-sided weighting scheme is useful only where the "good" or "desired" response to each statement is agreement, and is used only with the course objectives survey, where students are asked for each objective to select a response for each of the following statements: