Discussion

Issues with Personality Tests

In Analysis 1, out of all 20 blood type characteristic items, 13 of them (after the Bonferroni’s correction) clearly output the differences as predicted. This study measured self-reported personality, and since many Japanese believe in the relationship between blood type and personality, conventional personality psychology theory suggests the differences among the blood types will certainly appear, as prior mentioned. For the same reason, personality psychology hardly explains that no difference in blood type, or inconsistent results, appeared theoretically (Cattell et al., 1964; Cho et al., 2005; Cramer & Imaike, 2002; Flegr et al., 2013; Furukawa, 1927 & 1930; Gupta, 1990; Jogawar, 1983; Kim, Lee & Lee, 2007; Lester & Gatto, 1987; Mao et al., 1991; Nawata, 2014; Rogers & Glendon, 2003; Sharifi et al., 2015; Sato & Watanabe, 1992; Shimizu & Ishikawa, 2011; Wu et al., 2005).
Ryu and Sohn (2007) re-analyzed Cho, Suh and Ro’s result of the Big Five test of 40 items (2005), and found statistically significant differences which matched blood type characteristics in 10 individual items. This means that in the case of a “personality factor” composed of multiple items, the difference by the blood type decreases ‒ few significant differences appear. Wada (1996) constructed the Big Five Scales (BFS) of 60 personality trait terms, which matched blood type characteristics in only 3 individual items (Appendix B).
This corresponds to the inconsistent results of many preceding studies conducted by psychologists. In addition, the reason why psychological personality tests did not show differences in blood type (Cattell et al., 1964; Cho et al., 2005; Cramer & Imaike, 2002; Flegr et al., 2013; Furukawa, 1927 & 1930; Gupta, 1990; Jogawar, 1983; Kim et al., 2007; Lester & Gatto, 1987; Mao et al., 1991; Nawata, 2014; Rogers & Glendon, 2003; Sharifi et al., 2015; Sato & Watanabe, 1992; Shimizu & Ishikawa, 2011; Wu et al., 2005) was probably due to the halo effect. Kamise and Matsui (1994) said “there exist core characteristics of each blood type, and the entire contents are formed around them” (Appendix D). Thus, the differences by the blood type are heavily depended upon “original” words. In many cases, differences did not appear if “similar” words were selected.

Consistency with Preceding Studies

Preceding research by psychologists had concluded that there appeared no difference in personality among blood types, and that even if there was a difference predicted by the blood type, it was assumed to be the result of self-fulfillment prophecy phenomena (Kim et al., 2007; Sakamoto and Yamazaki, 2004; Yamaoka, 2009; Yamazaki and Sakamoto, 1991 & 1992). However, the same differences in scores were found in the group who reported no blood type personality knowledge (“no-knowledge” group), although the values were smaller; all four items showed the same results as those shown for blood type characteristics in the preceding psychology papers (Sato et al., 1991; Watanabe 1994; Yamazaki & Sakamoto, 1991& 1992 – Tables 2-4). Therefore, it is highly likely that differences by blood type are real, not caused by self-fulfilling prophecy phenomena; no difference in the “no-knowledge” groups were caused by Type II errors.
In addition, backed on this study’s AI result, it was suggested that these phenomena occurred not only because of the wording of question items, but also because gender and age were not taken into account (Chart 1). The experimental blood type predictions by AI (Amazon Machine Learning) in this study found that adding non-blood type variables, such as gender or age, to the training data considerably increased the accuracy. Moreover, when performing blood type prediction, AI sometimes failed to build its machine learning models, if gender or age of the training and prediction data were different. In this respect, AI technology may suggest that factors such as gender and age affect the characteristics of blood type. Therefore, gender, age, and other factors may offer a better explanation, even if past data were inconsistent.