Anohito statistical tests

The hunt for Anohito (left: Chapter 4; right: Chapter 9)

If you have read through my posts, you will notice that I have collected and reviewed up to 14 Anohito aspects in Candy’s “present day” monologue. In my latest post, I said that out of the 14 aspects, 11 of them are only attributable to Terry, 3 of them are attributable to both Terry and Albert, and none is attributable to Albert only. Thus… what does it look like if I plug those numbers into a statistical program? Does the proportion of Terry-related aspects (11 out of 14) significantly overwhelm both-Terry-Albert aspects (3 out of 14)?

The fact that there is no Anohito aspect that refers only to Albert in Candy’s monologue already established that Terry is Anohito. However, just to satisfy some statistical curiosity, let me plug those numbers.

I gather that our moderator Nila might find this funny...(source)

I arranged the data as follows, with 14 Anohito aspects as the cases. The variable to test is “Anohito”, a categorical variable with two categories (0 = “Can be Terry or Albert” and 1 = “Terry”).  I did not include a category for Albert because I did not find any Anohito aspect that is solely attributable to Albert. Please read this post to read my analyses of Anohito’s voice, smile and open arms (Aspects 12-14). Read this post for my analysis of Aspects #8 (Anohito laughing at Candy’s mistake) and #10 (the heat of Terry’s chest). Read this post for my summary of Nee-chan's Japanese blog post for the other aspects.

For your convenience, I re-write the aspects here:

1. Anohito obtained Slim's painting (Prologue) -- either Albert or Terry
2. Anohito's heirloom box (Vol 2, p. 148-149) -- either Albert or Terry
3. River Avon (Vol 1, p. 230-235) -- Terry
4. Daffodils (Vol 1, p. 230-235) -- Terry
5. Many painful separations, but reunion is possible when alive, thus Candy no longer fears separations (Vol 1, p. 230-235) -- Terry
6. Candy's reluctance to leave Anohito and vice versa  (Vol 1, p. 230-235) -- Terry

7. Anohito's reaction on Candy being a stowaway (Vol 2, p. 148) -- Terry

8. Anohito laughed at Candy's mistake of judging people by appearance (Vol 2, p. 196) -- Terry
9. Shakespeare books (Vol 2, p. 197) -- Terry
10. The heat of Terry’s chest is still violently pulsating within Candy (Vol 2 p. 237) -- Terry
11. Anohito fixed Stear's music box (Vol 2, p. 238-239) -- either Albert or Terry
12. Epilogue Anohito's voice -- Terry
13. Epilogue Anohito's smile -- Terry
14. Epilogue Anohito's open arms -- Terry


I conducted a Non-Parametric Chi Square to test whether the proportion of “Terry” within the sample (11/14) is statistically the same with the proportion of “Can be Terry or Albert” (3/14). I used IBM SPSS ver 24 to analyse the numbers.

My hypothesis is as follows:

H0: The proportion of Terry and Terry/Albert is 50:50 (p=0.5)
H1: The proportion of Terry and Terry/Albert is not 50:50 (p ≠ 0.5)

The result is as follows:

With p value=0.033 (alpha level 0.05), we can conclude that the proportion of “Terry” (11/14) is significantly different, in this case larger, than the proportion of “Can be Terry or Albert” (3/14). Thus, Anohito is Terry.

I did actually run a t-test (Nila likes to call it “Terry-test”, and I actually agree…), and the t-test yielded p value=0.000 for Terry. However, since our data is categorical, and t-test is valid only for interval data, I did not upload the t-test result here.

I did also calculate the z-score manually to test our hypothesis. This website gives a good example of how to conduct a z-test for proportion. The formula to calculate the z-score for proportion is:

So, we have:

Terry = 11;
Terry or Albert = 3
n=14
Expected proportion (p0) = 0.5
For alpha level 0.05, we reject Ho if |z| > 1.96
Observed proportion for Terry (that's the p with the cap) = 11/14.

Since our z score is 2.138, it is larger than 1.96. Thus, we reject H0. We reject the notion that the number of Terry aspects is the same as the number of Terry/Albert aspects.

Thus… Terry is Anohito… ^_^