Please have a look on the following documents and please give me your valuable opinion. Scientifically how much it is a correct way or acceptable to get continuous data through accumulation of different statement as follows. I am going to use regression analysis in my study. Most of my independent variables are constructed as mentioned in the following way.
Thanks in advance.
Hope for your best cooperation
Khondokar Humayun Kabir,
The answers do not seem to belong to the same group of questions. E.g. quality of life will be....
You have to do an internal reliability test (Cronbach's alpha) and then compute a variable (a sum) of those items that go together with an alpha of at least .70.
Morning Khondokar,
I agree with Beatrice, you have to make some operations before you proceed with a regression.
1. How many constructs do you you measure with these questions? Are it 3 constructs and for each of them 3 questions? Or do you plan to use each scale as a separate construct?
1.1. You can aggregate from many scales to one super-variable in a regression. Previously, you should check internal consistency, convergent and discriminant validity, reliability. Cronbach's alpha is one of these necessary measures. You should measure also: composite reliability (alternative for Cronbach 's alpha, tolerate different types of scales), average shared variance, maximum shared variance, average variance extracted, and maximum reliability. These are more than enough to legitimise you multi items scales.
1.2. If you want to use each single scale as a separate variable, then it is not a good idea. Statisticians would argue that these variables are not interval scales but ordinal. You will have to make either a generalised (not general) linear modelling with ordinal variables or you will have to dichotomies your scales into high/low, based on the median (not mean). I think this procedure is reliable, but some marketologists would dichotomise based not on the median, but median +/- Standard deviation. In this event you cannot make a regression, only group comparison (for example, ANOVA).
2. You can use the option "dimension reduction " in SPSS. This is for the option 1.1. Add you variables and SPSS will produce a factor which you can use as single interval variable in your analysis.
Sincerely,
Eugene
Please, do what Beatrice said. BTW, the computer directed this question to me -- I am afraid of unnecessary accusations.
Following the young lady, separate your deforestation from those of quality of life and do your questions "Likert scale-like.
I think if you want to be 'kool' with your professors or readers, you can use "factor" loadings. Nevertheless, do not report it if it contradictory to Beatrice suggestion. So you choose!!
Dear Chuck,
Do not fear me, I am actually a rather kind person but sometimes tired that men repeat what I say. Also if one does not understand that n means sample size, what do you then do on Researchgate? I apologize for having stepped on your toes! I do not want to offend people but sometimes in written text I sound harsh. This our supervisor in psychotherapy warned us against.
Dear Beatrice:
I am beginning to truly "love" you. You know how life is "you quarrel and makeup," and it is always sweeter.After, reading yours this night, I felt immediately that you are a nice and wonderful person. Most importantly, very truthful -- I like honest people. It just happened, that we were both tired so whatever happened, we took it wrongly. Then, when you mentioned something about men and women, flames came out of my head.
The reason is that I am 300 hundred % not like that. I am very respectful of the opposite sex; only men with low-self esteem are anti-female and busy looking for something negative to tell a lady. In the end, what do people like that gain? NOTHING! I have three females who work for me and they are like family to me. Please, your apology is accepted. Let us fast new and act as if nothing happened. In RG, students come and some may not know what n is. Some do not even want to know. Some teachers who teach undergraduate courses, require their students to register in RG as part of class requirements. It internationalizes the student, hence, making them aware of what is going on in other parts of the world.Beatrice, I suffer from the same "written text harshness" and what I do is to have my female assistant type my emails. I believe that you are a kind person, not rather kind +:)
Dear Chuck,
Thank you for your answer. It is now snow from yesterday.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Inorganic Materials
- Metals
- Alloys
- Steel