I am doing a study of the relationship between two items say A and B. The data consists of information printed on the annual reports from say 2000 - 2010 (10 years). For each item, i.e. A and B there are sub variables. Item A has 3 variables with 9 indicators and item B has 1 variable with 2 indicators.

The data for item A are in words. I used content analysis to compute the number of times each indicator was mentioned in each report for each year. For the item B, the data are in percentages so i have 2 different percentage values for each year (10 years).

I want to find the relationship (positive, negative or neutral) between these items with item A as my independent variable and item B my dependent variable. How do I go about this? I have been told to use SPSS, but I don't know what method I should use.

How should I go about my analysis to determine the relationship between these two items?


Throwing simple thoughts together and restating your problem: you have a latent variable B with 2 observables (each clamped between 0 and 1), and a latent vector A of 3 variables with 9 observables (all count data), and are trying to find the relationship between A and B.

The first hurdle to overcome is how to formulate the relationships between latent and observable variables. What in your opinion is the process that generates word counts? What is the correlation among observables? Are you really sure vector A includes THREE variables and not, say, five?

The second step would be estimating the model in its entirety; however, judging by your post I suppose this would be overkill, and would suggest that you should run linear regressions between observable items influencing latent stuff (A and B).

Since there are many companies and many years, you would have to start reading on panel data before actually running the regressions...

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.