Blood type personality theory

In 1927, Takeji Furukawa, a professor at Tokyo Women's Teacher's School, published his paper "The Study of Temperament Through Blood Type" in the scholarly journal Psychological Research. The idea quickly took off with the Japanese public despite Furukawa's lack of credentials, and the militarist government of the time commissioned a study aimed at breeding ideal soldiers.[2] The study used ten to twenty people for the investigation, thereby failing to meet the statistical requirements for generalizing the results to the wider population.

On the other hand, in 1934, Fisher announced the chi-squared test, which is very popular at present, for the first time. Several scholars said that they found statistically significant differences in analyzing Japanese work conducted at that time.[15]

In another study, Furukawa compared the distribution of blood types among two ethnic groups: the Formosans in Taiwan and the Ainu of Hokkaidō. His motivation for the study appears to have come from a political incident:[16] After the Japanese occupation of Taiwan following Japan's invasion of China in 1895, the inhabitants tenaciously resisted their occupiers. Insurgencies in 1930 and 1931 resulted in the deaths of hundreds of Japanese settlers.[16]

The purpose of Furukawa's studies was to "penetrate the essence of the racial traits of the Taiwanese, who recently revolted and behaved so cruelly." Based on a finding that 41.2% of Taiwanese samples had type O blood, Furukawa assumed that the Taiwanese rebelliousness was genetic. His reasoning was supported by the fact that among the Ainu, whose temperament was characterized as submissive, only 23.8% is type O. In conclusion, Furukawa suggested that the Japanese should increase intermarriage with the Taiwanese to reduce the number of Taiwanese with type O blood.[16]

Interest in the theory faded in the 1930s. It was revived in the 1970s with a book by Masahiko Nomi, a journalist with no medical background (he graduated from the engineering department of the University of Tokyo). Few Japanese psychologists criticized him at that time,[17] so he continued to demonstrate statistically significant data in various fields and published several books with these results.[18] Later after his death in 1981, Masahiko Nomi's work was said to be largely uncontrolled and anecdotal, and the methodology of his conclusions was unclear.[19] Because of this, he was heavily criticized by the Japanese psychological community, although his books remain popular.[19] His son, Toshitaka Nomi, continued to promote the theory with a series of books and by running the Institute of Blood Type Humanics.[20] He later established the Human Science ABO Center for further research and publication in 2004.

Kengo Nawata, a Japanese social psychologist, statistically analyzed three data sets of over 10,000 Japanese and American people in total.[10] However, 65 of the 68 items yielded non-significant differences between blood types and the other three items showed relatively slight relationships. Therefore, the blood type explained only 0.3% of the whole differences of these data sets. This result suggests that blood type explained very little of people's personalities. Nawata concluded that there is no relevance of blood type for personality.

However, some academic researchers have shown several statistically significant data in Japan and Korea. Akira Sakamoto and Kenji Yamazaki, Japanese social psychologists, analyzed 32,347 samples of annual opinion polls from 1978 through 1988.[11][12] These results indicated that Japanese blood-typical stereotypes influenced their self-reported personalities—like a self-fulfilling prophecy.

Cosy Muto and Masahiro Nagashima et al. (Nagasaki University) conducted a supplementary survey of Yamazaki and Sakamoto in 2011.[13] They demonstrated that significant and the same difference in personalities between blood-types by using the same database as Samamoto and Yamazaki used. In the 1990s, the difference due to blood types was stabilized and variances became smaller. Then in the 2000s, the difference was statistically significant, too. However, the effect magnitude was extremely small, despite 'significance' in the statistical sense.

Another Japanese social psychologist, Shigeyuki Yamaoka (Shotoku University), announced results of his questionnaires, which were conducted in 1999 (1,300 subjects)[21] and 2006 (1,362 subjects),[22] In both cases, the subjects were university students, and only subjects with enough knowledge of and belief in the "blood-type diagnosis" showed meaningful differences. He concluded that these differences must be the influence of mass media, especially TV programs. Yamaoka later examined 6,660 samples from 1999 through 2009 in total and found the same result.[23]

On the other hand, there are opinions that the statistically meaningful differences according to the blood types are not explained only by beliefs, nor are they a self-fulfilling prophecy. In Japan, the penetration rate of blood-typical personality traits was investigated. Yoriko Watanabe, a Japanese psychologist (then Hokkaido University), chose "well-known" traits and found most traits were known to no more than half of Japanese (subjects were university students).[24] A Japanese writer, Masayuki Kanazawa, analyzed these blood-typical traits in combination with data from Yamaoka (1999)[21] that used the same items of Watanabe's penetration survey.[24] If blood-typical differences are caused by penetration (or their self-recognition), the rate of differences of a trait is proportional to the rate of its penetration. However, Kanazawa was not able to discover any association with blood-type differences and penetration rates.[25] This result raises doubt about the role of beliefs and self-fulfilling prophecy.

Most reports that demonstrated statistical correlation attribute differences to a self-fulfilling prophecy. However, no study directly proved the existence of "self-fulfillment". Therefore, the opinions of researchers are varied at present:

1. Whether there is a statistical correlation or not;
2. Whether any statistical correlations are superficial, being caused by subjects' self-fulfilling prophecy, or if they are truly caused by the blood type.

The five-factor model tests were carried out in several countries, including Japan, Korea, and Taiwan, after the year 2000. These tests were intended to digitize self-ratings of the "big five" personality traits. It was expected that differences in self-reported personalities (a self-fulfilling prophecy) would be detected from the subject who believed in blood-typical stereotypes. As a result, researchers found no meaningful statistical difference.[7][8][9][26][27]

So Ho Cho, a Korean psychologist (Yonsei University), and the others carried out a questionnaire about blood-typical items to subjects and discovered statistical differences as expected.[26] However, the difference was not found when the five-factor model for big five personality traits was administered to the same subjects. Another Korean researcher Sohn (Yonsei University) re-analyzed Cho's data.[28] He found that several independent items of the big five personality test detected differences according to each blood-typical stereotype. However, these differences became extinct in the process of plural items being gathered to five factors (big five).

If these results are correct, the five-factor model test cannot detect differences between the blood types – if such a causal link did indeed exist.[11][12][21][22][23][29] The phenomenon is not reported when using the MBTI personality index.

Kim and Yi (Seoul University of Venture & Information) measured the brain waves of 4,636 adults. They reported that type O people were most stress-resistant.[30] Moreover, an experiment using light topography instruments by Munetaka Haida (Tokai University School of Medicine) suggests the possibility that activated parts of the human brain are different according to blood types. i.e. type A's left brain is superior to the right, while type B's right brain is superior.[31]