I investigate the relationship between several personal (e.g. extroversion) and organizational (e.g. management support) characteristics and the organizational performance. The personal and organizational characteristics are measured subjectively by means of a questionnaire. In turn, the organizational performance is measured objectively by means of three financial indicators, i.e. gross profit, net income and operating income. I got the values of those variables for three years and calculated mean values. However, the problem is that those mean values seem to be constant values and do not show any correlation with other variables (e.g. personal characteristics) since they do not have any variance.
However, I noticed, for example, that when I calculate an index variable (as a mean) in SPSS (e.g. extroversion) based on a set of items (questions) measuring it, SPSS does not show it as a constant variable, i.e. there is a certain variation in its values.
Dear Denis,
I think you are mixing up several issues here. First start by defining your research question. Then think about what data you have collected and how that can help you to answer your research question by formulating a hypothesis.
For example, if you are interested in the way personal and organisational issues affect organisational performance then you might be looking at a multiple linear regression or a MANOVA, depending upon the nature of your dependent variable or variables.
Several variables explaining the same construct or similar constructs are sometimes combined together using a dimension reduction technique, such as pincipal component analysis or explanatory factor analysis. These can be quite complex but they reduce the complexity of the analysis which is then applied.
Finally, you mentioned an issue to do with measuring your variables over three successive years then taking an average. I would recommend keeping all your data in your model rather than averaging it out. This could lead you into mixed factorial ANOVA or MANOVA.
Just adding to Peter Samuels ' final paragraph. Using aggregate values like the mean when your interest is in individual level hypotheses was discussed in Robinson (1950, Amer Sociological Rev) and often discussed in introductions to multilevel modeling textbooks. However, the main point is that if a variable doesn't vary it can't co-vary with anything.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS