Please suggest what analysis method to use for determining how much particular parameter affect the overall testing of a software?
Hi Anureet
You could rank order the mean ratings from highest to lowest (assuming that from the 0-5 the 5 is 'good'). This will tell you the most important parameter.
You might want to consider the fact that different aspects of influence may go together in more meaningful dimensions. For example, if I asked people to rate why they like Apple products they may report that they like (1) the feel of the products, (2) the way they look, (3) the ease of use, and (4) high reliability. But these are not 4 different aspects of Apple products, they probably reflect more meaningful unobservable attitudes. 1 and 2 probably reflect the aesthetic appeal while 3 and 4 represent beliefs about the quality of the products. If you think this sounds reasonable the you may want to do some factor analysis using your data. ResearchGate and Google will be your friends for seeking out more information on the subject of factor analysis.
Best wishes, Mark
It would also be good to do a Singular Value Decomposition. This will tell you whether your respondents cluster: do people who think software attribute X is important also think Y is important.
You can also apply the clustering technique for the opinion response.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Software