Types of Multigroup Moderation Analysis Hypotheses

on

Here, I compiled different ways of presenting multigroup analysis hypotheses and interpreting these, after reading a lot of moderation papers. I used attitude (ATT) and purchase intention (PUR) as examples. Knowledge (KNO) as moderator (high and low).

Initial steps in doing multigroup analysis in SEM-AMOS

Types of multigroup moderation hypotheses:

  • H1: KNO moderates the relationship between ATT and PUR.
    • Also means that: 
      • KNO has a moderating effect on the relationships between consumers’ purchase intention.
      • KNO has a moderating effect between ATT and PUR.
      • If both groups are significant and path differences are significant
        • The relationship between ATT and PUR varies significantly across the high- and low-knowledge groups.
        • The variation between consumer groups is significant.
      • The influence of ATT and PUR varies significantly across high and low knowledge consumer groups.
        • To conclude that the variation between consumer groups was significant.
    • Requires:
      • Both group and group path differences chi-square test.
    • This is a general statement.
      • Whatever the results are, you just have to explain any form of moderation that occurs.
      • Papers below can be used as references since this is a general moderation statement.
    • Reference papers: 
      • Teng, C. C., & Lu, C. H. (2016). Organic food consumption in Taiwan: Motives, involvement, and purchase intention under the moderating role of uncertainty. Appetite, 105: 95-105. https://doi.org/10.1016/j.appet.2016.05.006
      • Lee, K. H., Bonn, M. A., & Cho, M. (2015). Consumer motives for purchasing organic coffee: The moderating effects of ethical concern and price sensitivity. International Journal of Contemporary Hospitality Management, 27(6): 1157-1180. https://doi.org/10.1108/ijchm-02-2014-0060
      • Huang, X. Q., & Ge, J. P. (2019). Electric vehicle development in Beijing: An analysis of consumer purchase intention. Journal of Cleaner Production, 216: 361-372. https://doi.org/10.1016/j.jclepro.2019.01.231 
      • Kautish, P., Paul, J., & Sharma, R. (2019). The moderating influence of environmental consciousness and recycling intentions on green purchase behavior. Journal of Cleaner Production, 228: 1425-1436. https://doi.org/10.1016/j.jclepro.2019.04.389
        • Did not interpret significant groups (high and low, ex) if no significant group differences.
  • H2: The effect of ATT on PUR is significantly higher for high-knowledge consumers.
    • Also means that:
      • The positive relationship between ATT and PUR is stronger for high-knowledge consumers.
      • The variation between consumer groups is significant.
    • Requires:
      • A significant p-value of the chi-square difference test between two groups or SIGNIFICANT GROUP DIFFERENCES
      • Not necessarily both groups’ p-value is significant.
    • This is more specific:
      • Focused on individual path analysis but comparing two groups if which is higher.
      • Applicable if only one group was found significant.
    • Reference papers: 
      • Lee, H. J., & Hwang, J. (2016). The driving role of consumers' perceived credence attributes in organic food purchase decisions: A comparison of two groups of consumers. Food Quality and Preference, 54: 141-151. https://doi.org/10.1016/j.foodqual.2016.07.011
        • did not interpret significant groups (high and low, example) if no significant group differences.
      • Beza, E., Reidsma, P., Poortvliet, P. M., Belay, M. M., Bijen, B. S., & Kooistra, L. (2018). Exploring farmers' intentions to adopt mobile Short Message Service (SMS) for citizen science in agriculture. Computers and Electronics in Agriculture, 151: 295-310. https://doi.org/10.1016/j.compag.2018.06.015
        • Did not consider insignificant group path differences even if there are groups that are significant
        • Because hypothesis says “stronger for …”
  • H3: The positive effect of ATT on PUR is significant for high-knowledge consumers.
    • Also means that:
      • ATT is a good predictor of PUR for high-knowledge consumers.
      • The positive relationship between ATT and PUR is significant for high-knowledge consumers.
        • High knowledge consumers indicated a significant relationship (moderating effect) between ATT and PUR.
        • or High knowledge of product traceability moderates positively the relationship between ATT and PUR.
        • ATT had a positive significant relationship with PUR for the high-knowledge consumer group.
      • High knowledge consumers have a more positive attitude towards purchasing traceable organic rice.
      • High knowledge of product traceability strengthens the positive relationship between ATT and PUR.
    • Requires:
      • A significant p-value of that group only in the hypothesized path.
      • A not necessarily significant p-value of the chi-square difference test between two groups.
      • Applicable if the variation between consumer groups is not significant, so you chose to focus on the group that is significant on that path.
    • This is also more specific:
      • Focused on individual path analysis but looked at only which group is significant.
    • Reference paper:

Ways of analyzing multigroup moderation results:

  • If there are significant path differences, conclude that one path is stronger than the other.
  • If there are no significant path differences, but one group is significant, you can conclude that there is some form of moderation. Examples:
    • ATT is a good predictor of PUR for high knowledge group consumers (The positive relationship between ATT and PUR is significant for high knowledge group consumers).
    • SNS is a good predictor of purchase intention for high knowledge group
    • PBC is a good predictor of purchase intention for low knowledge group
    • HEC is a good predictor of purchase intention for high knowledge group

In using the Multigroup Magic plugin:

  1. Go to plugin
  2. Multigroup
  3. Refer to this video.
  • If there are significant path differences = the path is stronger for high knowledge than low knowledge group.
  • If there are no significant differences = the positive relationship between these two variables is only significant for high knowledge group consumers. Example:
    • SNS is a good predictor of purchase intention for high knowledge groups.
      • Even if there is no significance, it is only meaningful for high knowledge group consumers.
      • The positive relationship between SNS and purchase intention is only significant for high knowledge groups.
      • You cannot conclude that the relationship is stronger for high knowledge group consumers, but you can conclude that SNS is a good predictor of purchase intention for high knowledge group consumers
Hope this helps. Write down below your comments. Thank you.
King Mongkut's Institute of Technology Ladkrabang

Comments

Post a Comment