Hello Rookantha,

I think you're more likely to get constructive responses if you could: (a) indicate the specific research question(s) you're trying to answer; and (b) elaborate on the variables/measures, sample, and data collection (procedure/s) that you used.

Good luck with your work.

It looks like you are using a Cronbach’s alpha, α. It was developed by Lee Cronbach in 1951 to measure reliability, or internal consistency. “Reliability” means is how well a test measures what it should.

Cronbach’s alpha usually tests if multiple-question Likert scale surveys are reliable. These questions measure latent variables — hidden or unobservable variables like: a person’s conscientiousness or openness. These are very difficult to measure in real life. Cronbach’s alpha will tell you if the test you have designed is accurately measuring the variable of interest.

Based on your data, the Cronbach's alpha is in the range 0.6 to 0.77. It means that the internal consistency of your test is, at least, questionable or worse.

At the same time, a high level for alpha may mean that the items in the test are highly correlated. However, α is also sensitive to the number of items in a test. A larger number of items can result in a larger α, and a smaller number of items in a smaller α. If alpha is high, this may mean redundant questions (i.e. they’re asking the same thing). A low value for alpha may mean that there aren’t enough questions on the test. Adding more relevant items to the test can increase alpha. Poor interrelatedness between test questions can also cause low values, so can measuring more than one latent variable.

Confusion often surrounds the causes for high and low alpha scores. This can result in incorrectly discarded tests or tests wrongly labeled as untrustworthy.

Also, keep in mind that Cronbach’s alpha assumes the questions are only measuring one latent variable or dimension. If you measure more than one dimension (either knowingly or unknowingly), the test result may be meaningless. You could break the test into parts, measuring a different latent variable or dimension with each part.

Based on your presented plots the use of a linear correlation of the paired variables are in the range ~ 0.2 - 0.3. It means that there is practically no linear correlation between your paired variables, or it's very weak.

In summary, I believe that your test (if it is indeed a Likert scale test) was poorly designed, and its results are inconclusive and questionable.

David Morse :Thank you. I've updated the question

Alexander Kolker : Is there anyway that I can justify the low Cronbach's alpha values? However as per the coefficient I can accept two hypothesis? isn't it?

There is a relationship between II and EGB , IC and EGB.

Thank you

Hello again Rookantha,

Presuming that the four IVs represent the target dimensions of ETL, then:

1. If this RQ is more concerned with the zero-order correlations, then your table of correlations for IVs with DV (the 7th one in your post) give the answers to this. The individual correlations with Green Behavior (DV) vary from .37 (for the IV of IM) to .46 (for II). Note that the other two IVs have nearly the same (zero-order) relationship with the DV as does II.

2. As there is correlation among the IVs, the preferred way to address this question would be via partial correlations or the standardized regression coefficients (from the regression model, given in the last table in your post): From these, II (with value of .324) is given more emphasis in the model than the next closest IV (IC, with value .270), and far more than the other two IVs (IM, IS). As only II and IC have estimated regression coefficients that are significantly different from zero (at the .05 level), the IM and IS scores could be said to not be contributing to the explanatory power of the model, given that II and IC scores were part of that model.

3. For your sample, the four IVs collectively explain (R-squared = ) 30.7% of the observed variance in Green Behavior scores; that is an amount that is statistically distinguishable from zero: F(4, 76) = 8.402, p < .001.

4. I think that this RQ calls for inference beyond the data!

Good luck with your work.