Scheffé’s method

210 views
Skip to first unread message

Benjamin Barrett

unread,
Feb 9, 2012, 10:50:44 PM2/9/12
to hon...@googlegroups.com
For the archives.

I just ran across the expression シェッフェの一対比較浦変法. Wikipedia doesn't describe Scheffé’s method in detail and there is no Japanese Wiki page.

As shown at http://blog.taste-technology.com/?p=64, Scheffé’s method is considered in Japan to have four variations:

1. シェッフェの原法(順序考慮する、パネル考慮しない)
2. 芳賀の変法(順序考慮しない、パネル考慮しない)
3. 中屋の変法(順序考慮しない、パネル考慮する)
4. 浦の変法(順序考慮する、パネル考慮する)

浦変法 is written both with and without the の.

More description can be found at http://www.hi.is.uec.ac.jp/rcb/paper/PDF/H19_kuroda.pdf, where it says:

シェッフェの原法では被験者を数組に分けて、1人の被験者が1つの組み合わせのだけを1回比較するが、浦の変法では1人が全ての組み合わせについて一対比較を行い、人を変えて反復する。また、浦の変法は順序効果を考慮するため比較順序を逆にした組み合わせも実験する必要がある。例えば、(Ai,Aj)という組み合わせの比較順序を逆にしたものは(Aj,Ai)となる。順序効果を考慮する場合、試料数が k 個の場合は kC2×2 対の組み合わせが生じる。

Benjamin BArrett
Seattle, WA

Mark Spahn

unread,
Feb 10, 2012, 12:17:59 AM2/10/12
to hon...@googlegroups.com

http://www.hi.is.uec.ac.jp/rcb/paper/PDF/H19_kuroda.pdf
says (at the end of page 14):

シェッフェの原法では被験者を数組に分けて、1人の被験者が1つの組み合わせのだけを1回比較するが、浦の変法では1人が全ての組み合わせについて一対比較を行い、人を変えて反復する。また、浦の変法は順序効果を考慮するため比較順序を逆にした組み合わせも実験する必要がある。例えば、(Ai,Aj)という組み合わせの比較順序を逆にしたものは(Aj,Ai)となる。順序効果を考慮する場合、試料数が
k 個の場合は kC2×2 対の組み合わせが生じる。
==UNQUOTE==

Hoping this will help rather than just belabor the obvious...

In the original document, in the notation kC2 the
k and 2 are written small, like subscripts.
Another form of this notation is C(k,2).
C(k,2) denotes the number of combinations of
k items taken 2 at a time. Let's say there are k=5
subjects, named A, B, C, D, E. From this set of
5 items there are C(5,2) distinct 2-member sets
of items taken from this 5-member set {A,B,C,D,E}.
C(5,2) = 10, and these 10 two-member sets (pairs) are
{A,B}, {A,C}, {A,D}, {A,E}, {B,C},
{B,D}, {B,E}, {C,D}, {C,E}, {D,E}.
The notation {B,E} means a two-member set,
exactly the same set as denoted by {E,B}.
If the order of the elements makes a difference,
then we are dealing not with sets but with
ordered pairs, which are denoted with parentheses.
The ordered pair (E,B) is different from
the ordered pair (B,E).
So from each pair {X,Y} we can derive two
distinct ordered pairs, namely, (X,Y) and (Y,X).

http://www.itl.nist.gov/div898/handbook/prc/section4/prc472.htm and
http://en.wikipedia.org/wiki/Scheff%C3%A9%27s_method
give an explanation of Scheffe's method but require a lot of
background information to follow.
-- Mark Spahn (West Seneca, NY)

Reply all
Reply to author
Forward
0 new messages