A median split was conducted to divide the subjects into HNFC and LNFC groups based on their NFC score.
For testing H3a, H3b and H3d, four independent samples t-tests were conducted to compare whether a significant difference in means exists for HNFC and LNFC consumers across website loyal and website non-loyal consumers.
[mean.sub.HIGH WEBSITE LOYALTY] = 6.176; t = 4.41, mean difference = 2.09, p < .01) than for LNFC consumers ([mean.sub.LOW WEBSITE LOYALTY] = 5.444 vs.
Based on the rationale of the ELM and the need for cognition construct, it would be reasonable to expect that LNFC subjects would be more ready to rely on source characteristics to make a judgement about the job under consideration than would HNFC subjects, since the latter are more likely to consider specific job attributes or claims before finalizing judgement.
LNFC, M = 5.36), and significant interaction effects for source credibility by NFC and message framing by NFC.
The simple effects analysis reveals that LNFC subjects (M = 4.68) rated job attitudes significantly lower than HNFC subjects (M = 5.80) when RJP messages were positively framed and given by non-experts and when RJP messages were negatively framed and given by a low credibility source (HNFC, M = 5.41 vs.