Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • II Independence of Irrelevant Alternatives -Multinomial logistic regression

    Hi to everybody,
    i 'm trying to do mlogit. For me it's the first time

    Code:
    mlogit a ITEM3 , baseoutcome(3) rrr

    where a can assume these values (1,2,3 e 4); ITEM3 (0,1)


    I would like to check if it is respected "II Independence of Irrelevant Alternatives". I try this code. Is it correct?
    Code:
    mlogit Diagnosi_num ITEM3 , baseoutcome(3) rrr estimates store allcats mlogit Diagnosi_num ITEM3 if Diagnosi_num!=2 , baseoutcome(3) rrr hausman . allcats, alleqs constant
    and obtain this output
    HTML Code:
    mlogit a ITEM3 , baseoutcome(3) rrr  Iteration 0:  Log likelihood = -86.308789   Iteration 1:  Log likelihood = -85.826252   Iteration 2:  Log likelihood = -85.717961   Iteration 3:  Log likelihood = -85.694522   Iteration 4:  Log likelihood =  -85.68864   Iteration 5:  Log likelihood = -85.687493   Iteration 6:  Log likelihood = -85.687307   Iteration 7:  Log likelihood = -85.687286   Iteration 8:  Log likelihood = -85.687282    Multinomial logistic regression                         Number of obs =     92                                                         LR chi2(2)    =   1.24                                                         Prob > chi2   = 0.5371 Log likelihood = -85.687282                             Pseudo R2     = 0.0072  ------------------------------------------------------------------------------            a |        RRR   Std. err.      z    P>|z|     [95% conf. interval] -------------+---------------------------------------------------------------- 1            |        ITEM3 |   1.58e-06   .0016328    -0.01   0.990            0           .        _cons |   .2962833   .0843346    -4.27   0.000     .1695976    .5176005 -------------+---------------------------------------------------------------- 2            |        ITEM3 |   1.419923    1.78003     0.28   0.780     .1216717    16.57065        _cons |    .351858   .0938537    -3.92   0.000     .2086028    .5934919 -------------+---------------------------------------------------------------- 3            |  (base outcome) ------------------------------------------------------------------------------ Note: _cons estimates baseline relative risk for each outcome.  .  . estimates store allcats  .  . mlogit a ITEM3 if a!=2 , baseoutcome(3) rrr  Iteration 0:  Log likelihood = -38.138846   Iteration 1:  Log likelihood = -37.672198   Iteration 2:  Log likelihood = -37.632994   Iteration 3:  Log likelihood = -37.629057   Iteration 4:  Log likelihood =  -37.62832   Iteration 5:  Log likelihood = -37.628158   Iteration 6:  Log likelihood = -37.628119   Iteration 7:  Log likelihood = -37.628111   Iteration 8:  Log likelihood = -37.628109    Multinomial logistic regression                         Number of obs =     72                                                         LR chi2(1)    =   1.02                                                         Prob > chi2   = 0.3122 Log likelihood = -37.628109                             Pseudo R2     = 0.0134  ------------------------------------------------------------------------------            a |        RRR   Std. err.      z    P>|z|     [95% conf. interval] -------------+---------------------------------------------------------------- 1            |        ITEM3 |   7.37e-07   .0011153    -0.01   0.993            0           .        _cons |   .2962638   .0843303    -4.27   0.000      .169585    .5175707 -------------+---------------------------------------------------------------- 3            |  (base outcome) ------------------------------------------------------------------------------ Note: _cons estimates baseline relative risk for each outcome.  .  . hausman . allcats, alleqs constant  Note: the rank of the differenced variance matrix (1) does not equal the number of coefficients being tested (2); be sure this is what you expect, or there may be problems computing         the test.  Examine the output of your estimators for anything unexpected and possibly consider scaling your variables so that the coefficients are on a similar scale.                   ---- Coefficients ----              |      (b)          (B)            (b-B)     sqrt(diag(V_b-V_B))              |       .         allcats       Difference       Std. err. -------------+----------------------------------------------------------------        ITEM3 |   -14.12059    -13.35799       -.7625974        1105.455        _cons |   -1.216505    -1.216439       -.0000661        .0015597 ------------------------------------------------------------------------------                          b = Consistent under H0 and Ha; obtained from mlogit.           B = Inconsistent under Ha, efficient under H0; obtained from mlogit.  Test of H0: Difference in coefficients not systematic      chi2(1) = (b-B)'[(V_b-V_B)^(-1)](b-B)             =   0.00 Prob > chi2 = 0.9994  .
    p is equal to 0.9994 so was it respected II Independence of Irrelevant Alternatives?


    thanks in advanced to everybody



  • #2
    Sorry i I insert the output again

    mlogit a ITEM3 , baseoutcome(3) rrr

    Iteration 0: Log likelihood = -86.308789
    Iteration 1: Log likelihood = -85.826252
    Iteration 2: Log likelihood = -85.717961
    Iteration 3: Log likelihood = -85.694522
    Iteration 4: Log likelihood = -85.68864
    Iteration 5: Log likelihood = -85.687493
    Iteration 6: Log likelihood = -85.687307
    Iteration 7: Log likelihood = -85.687286
    Iteration 8: Log likelihood = -85.687282

    Multinomial logistic regression Number of obs = 92
    LR chi2(2) = 1.24
    Prob > chi2 = 0.5371
    Log likelihood = -85.687282 Pseudo R2 = 0.0072

    ------------------------------------------------------------------------------
    a | RRR Std. err. z P>|z| [95% conf. interval]
    -------------+----------------------------------------------------------------
    1 |
    ITEM3 | 1.58e-06 .0016328 -0.01 0.990 0 .
    _cons | .2962833 .0843346 -4.27 0.000 .1695976 .5176005
    -------------+----------------------------------------------------------------
    2 |
    ITEM3 | 1.419923 1.78003 0.28 0.780 .1216717 16.57065
    _cons | .351858 .0938537 -3.92 0.000 .2086028 .5934919
    -------------+----------------------------------------------------------------
    3 | (base outcome)
    ------------------------------------------------------------------------------
    Note: _cons estimates baseline relative risk for each outcome.

    .
    . estimates store allcats

    .
    . mlogit a ITEM3 if a!=2 , baseoutcome(3) rrr

    Iteration 0: Log likelihood = -38.138846
    Iteration 1: Log likelihood = -37.672198
    Iteration 2: Log likelihood = -37.632994
    Iteration 3: Log likelihood = -37.629057
    Iteration 4: Log likelihood = -37.62832
    Iteration 5: Log likelihood = -37.628158
    Iteration 6: Log likelihood = -37.628119
    Iteration 7: Log likelihood = -37.628111
    Iteration 8: Log likelihood = -37.628109

    Multinomial logistic regression Number of obs = 72
    LR chi2(1) = 1.02
    Prob > chi2 = 0.3122
    Log likelihood = -37.628109 Pseudo R2 = 0.0134

    ------------------------------------------------------------------------------
    a | RRR Std. err. z P>|z| [95% conf. interval]
    -------------+----------------------------------------------------------------
    1 |
    ITEM3 | 7.37e-07 .0011153 -0.01 0.993 0 .
    _cons | .2962638 .0843303 -4.27 0.000 .169585 .5175707
    -------------+----------------------------------------------------------------
    3 | (base outcome)
    ------------------------------------------------------------------------------
    Note: _cons estimates baseline relative risk for each outcome.

    .
    . hausman . allcats, alleqs constant

    Note: the rank of the differenced variance matrix (1) does not equal the number of coefficients being tested (2); be sure this is what you expect, or there may be problems computing
    the test. Examine the output of your estimators for anything unexpected and possibly consider scaling your variables so that the coefficients are on a similar scale.

    ---- Coefficients ----
    | (b) (B) (b-B) sqrt(diag(V_b-V_B))
    | . allcats Difference Std. err.
    -------------+----------------------------------------------------------------
    ITEM3 | -14.12059 -13.35799 -.7625974 1105.455
    _cons | -1.216505 -1.216439 -.0000661 .0015597
    ------------------------------------------------------------------------------
    b = Consistent under H0 and Ha; obtained from mlogit.
    B = Inconsistent under Ha, efficient under H0; obtained from mlogit.

    Test of H0: Difference in coefficients not systematic

    chi2(1) = (b-B)'[(V_b-V_B)^(-1)](b-B)
    = 0.00
    Prob > chi2 = 0.9994

    .

    Comment


    • #3
      p is equal to 0.9994 so was it respected II Independence of Irrelevant Alternatives?

      I thank anyone who wanted to help me and I apologize if I entered the output incorrectly

      Comment

      Working...
      X