Announcement

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

  • Interpretation of interactive variable with opposite sign

    Hi,

    I have ran the following two regressions:

    1: Recognized losses= @ + Monetary incentives + controls

    2: Recognized losses= @ + dummy + Monetary incentives + Monetary incentives * dummy + controls
    Where the dummy is 1 for the years 2018-2022 and 0 for the years 2014-2017.

    Our results are a little hard to interpret since we found a significant negative effect between Monetary incentives and Recognized losses for both regressions. But then we found a significant positive effect between Monetary incentives * dummy and Recognized losses.

    Does this mean that when looking at the period 2018-2022 the correlation between Monetary incentives and Recognized losses is positive or does it mean that the negative correlation we found in regression 1 is stronger in the period 2018-2022?

  • #2
    August:
    welcome to this forum.
    As per FAQ, please share what you typed and what Stata gave you back (via CODE delimiters, please). Thanks.
    Kind regards,
    Carlo
    (Stata 19.0)

    Comment


    • #3
      Hi Carlo,

      Thank you for that. Belove you have a screenshots of our two regressions.The regression at the top is regression 1. Is this enough or do you need something els?

      The variable Monetary incentives is denoted IRNEW, the dummy as PREPOST and Recognized losses as LLPtoYEL.

      Comment


      • #4
        August:
        I'd propose a third hypothesis: taking a look at the within R-sq, the second regression is better than the first one.
        The sign of the coefficient under investigation flipped when you wisely plugged in the interaction, probably because the interactoon was part of the data generating process. I would also check whether the functional form of the regression model is well specified by replicating by hand the procedure detailed in the -linktest- entry, Stata .pdf manual.
        As a (hopefully not too) pedantic aside, in your future post please use CODE delimiters (as recommended by the FAQ) to share what you typed and what Stata gave you back. Thanks.
        Kind regards,
        Carlo
        (Stata 19.0)

        Comment

        Working...
        X