Skip to main content

View Diary: More on Research 2000 (474 comments)

Comment Preferences

  •  I found another problem (4+ / 0-)

    Using the numbers supplied by the report, namely the following:

    6/3/10       Favorable     Unfavorable    Undecided
    Question    Men  Women     Men  Women    Men  Women

    Obama        43    59       54    34        3     7
    Pelosi       22    52       66    38       12    10
    Reid         28    36       60    54       12    10
    McConnell    31    17       50    70       19    13
    Boehner      26    16       51    67       33    17
    Cong. (D)    28    44       64    54        8     2
    Cong. (R)    31    13       58    74       11    13
    Party (D)    31    45       64    46        5     9
    Party (R)    38    20       57    71        5     9

    If you add up the numbers, Boehner's male favorables (26) his unfavorables (51) and his undecideds (33) add up to 110.  That's just sloppy...

    I have also figured out a way in which the odd/even matches could have arisen.  N.b., this is not exculpatory.

    All of the numbers have to add up to 200, since the M and F numbers add up to 100 separately.  Let's say you had, from whatever source, Fav, Unfav, and Undec numbers: e.g., for Obama 51/44/5.  Now you want to concoct M/F numbers for your crosstabs.  Multiply each of the numbers by two, i.e. 102/88/10.  Because they are multiplied by two, they will always be even.  Then, according to some formula or other, split these numbers into M and F numbers that add up to 102, 88, and 10 respectively.  Because the numbers are always even, they can only be split into two odd or two even numbers, no matter which way you split them.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site