Relationship between blood types and personality is real, not imaginary!
Mr. Masayuki Kanazawa's wrote a book "An Introduction to Blood Type Humanics by Using Statistics" in 2014. (統計でわかる血液型人間学入門)
He later wrote an English book "Blood Type and Personality 3.0" in 2018.
In those books he says:
If relationship of blood types and personality is imaginary (i.e. caused by blood-typical briefs), scores of statistical differences must be proportional to those of trait
(blood-typical personality) recognition.
But the reality is, there is little relationship. The correlation coefficient is minus 0.10!
This result suggests that relationship of blood types and personality is real, not imaginary.
Mr. Kanazawa cross-checked blood-typical 28 question items of two surveys conducted by Japanese psychologists (Ms. Yoriko Watanabe and Mr. Shigeyuki Yamaoka *1).
Mr. Yamaoka said that 15 answers showed statistical significance with much blood-typical knowledge subjects of 649 persons (total number of questions is 28, as mentioned). On the contrary, no answer showed statistical significance with little blood-typical knowledge subjects of 651 persons. Many psychologists thought these differences were caused by "blood-typical beliefs" (conformation bias and/or self-fulfilling phenomenon) so that the relationship of blood types and personalities is not real.
Strangely enough, 7 questions out of those significant 15 ones are "wrong" answers. For example, scores of type O people are sometimes higher than A people, even if questions are originally for type "A" -- not O. Wondering, Mr. Kanazawa investigated the phenomenon carefully.
If the relationship of blood types and personality is imaginary (i. e. caused by blood-typical briefs), scores of statistical differences must be proportional to those of trait (blood-typical personality) recognition (Chart A; dots are actual data and a line shows the trend). On the contrary, if the relationship is real, scores of statistical differences have nothing to do with those of trait recongnition (Chart B).
The results are as indicated by dots and a line below (Chart C).
In reality, almost no relationship was observed. As a matter of fact, the correlation coefficient is minus 0.10!
If there are any positive relations, the coefficient of correlation must be plus!
This result suggests that the relationship of blood type and personality is real, not imaginary.
*1 Shigeyuki Yamaoka (2001), A Psychology Book for not to Become a Useless Adult
28 Question Items
Type A Items
Type B Items
Type AB Items
Type O Items