Announcement

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

  • Asking for solutions to a high standard error in contingent valuation study using double bound model in Stata

    I am doing a study on Willingness to Pay for Renewable Electricity using a Double-Bound Contingent Valuation model and analysed it in stata using the guide provided by Alejandro Lopez-Feldman (2012). However, the results showed a high standard error as follows:

    doubleb bid1 bid2 wtp1 wtp2 ngo

    initial: log likelihood = -<inf> (could not be evaluated)
    feasible: log likelihood = -13779.517
    rescale: log likelihood = -892.07509
    rescale eq: log likelihood = -710.1219
    Iteration 0: log likelihood = -710.1219
    Iteration 1: log likelihood = -699.37483
    Iteration 2: log likelihood = -699.01164
    Iteration 3: log likelihood = -699.01147
    Iteration 4: log likelihood = -699.01147

    Number of obs = 400
    Wald chi2(1) = 5.62
    Log likelihood = -699.01147 Prob > chi2 = 0.0178

    ------------------------------------------------------------------------------
    | Coefficient Std. err. z P>|z| [95% conf. interval]
    -------------+----------------------------------------------------------------
    Beta |
    ngo | 141.8417 59.83623 2.37 0.018 24.56488 259.1186
    _cons | 2395.434 41.96212 57.09 0.000 2313.19 2477.678
    -------------+----------------------------------------------------------------
    Sigma |
    _cons | 548.7313 26.17672 20.96 0.000 497.4259 600.0368
    ------------------------------------------------------------------------------

    First-Bid Variable: bid1
    Second-Bid Variable: bid2
    First-Response Dummy Variable: wtp1
    Second-Response Dummy Variable: wtp2

    May I know what does this entail (because the p value is significant)? Will it be a problem when I write the discussion and interpretation part of my study?

  • #2
    The magnitude of the standard errors should be viewed relative to the size of the coefficients. In that regard, your standard errors do not appear inflated to me. Unless you have some specific concern, e.g., the coefficient estimates appear to be out of the range of those found in previous studies, I think you should just proceed as usual.

    Comment

    Working...
    X