+ Reply to Thread
Results 1 to 4 of 4

Thread: Interaction between binary and continuous variable in OLS

  1. #1
    Points: 142, Level: 2
    Level completed: 84%, Points required for next Level: 8

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Interaction between binary and continuous variable in OLS




    I am trying to examine the association between maternal education and child school test scores using the following equation.

    TestScore = β0 + β1 EduYrMom+ β2 FemaleChild+ β3 EduYrMom * FemaleChild + εij

    where TestScores is the test score of the child, EduYrMom is the years of education of the child’s mother, FemaleChild is a binary value which equals 1 if the child is a girl. The following regression coefficients are obtained using Stata’s regress command.

    HTML Code: 
    Model 1 Test scores of child
    
    (1a)Years of education of mother    0.028
    (1b)Years of education of motherXfemale child   0.031
    Effect of mother's education on female children---- (1a+ 1b)    0.059*
    
    Model 2 
    
    (2a)Years of education of mother    0.039**
    (2b)Years of education of mother *female child  -0.005
    Effect of mother's education on female children---- (2a+ 2b)    0.034
    
    Model 3 
    (3a)Years of education of mother    0.303***
    (3b)Years of education of mother *female child  -0.047
    Effect of mother's education on female children---- (3a+ 3b)    0.256**
    
    note:  *** p<0.01, ** p<0.05, * p<0.1
    For model 1, it seems that mother’s education is positively and significantly associated with test scores of female children, but not for boys. For model 2, I find that mother’s education is significantly associated with test scores of boys only. For model 3, mother’s education is significantly and positively associated with the test scores of both boys and girls.

    I was wondering whether my interpretations are correct. I am also not sure how to interpret B3, the coefficient of the interaction term, which is insignificant for all the 3 models. I look forward to your suggestions.

    Monzur

  2. #2
    TS Contributor
    Points: 5,246, Level: 46
    Level completed: 48%, Points required for next Level: 104
    maartenbuis's Avatar
    Location
    Konstanz
    Posts
    372
    Thanks
    3
    Thanked 146 Times in 123 Posts

    Re: Interaction between binary and continuous variable in OLS

    The fact that your interaction terms are insignificant means that you cannot reject the hypothesis that effect of mother's education is the same for boys and girls. This is a common finding, girls nowadays do better at school, but the effect of social background tends to be similar for boys and girls, see e.g. M.L. Buis (2013) "The composition of family background: The influence of the economic and cultural resources of both parents on the offspring’s educational attainment in the Netherlands between 1939 and 1991", European Sociological Review, 29(3), pp. 593-602. http://maartenbuis.nl/publications/fam_backgr.html

    The fact that in some models the effect of mother's background is significant for boys but not for girls does not mean that the effects of mother's education is different between boys and girls. The test whether or not the interaction effects are 0 is the test for that hypothesis. See: Andrew Gelman and Hal Stern (2006) "The Difference Between “Significant” and “Not Significant” is not Itself Statistically Significant", The American Statistician, 602(4): 328--331. http://www.stat.columbia.edu/~gelman...ed/signif4.pdf

  3. The Following 2 Users Say Thank You to maartenbuis For This Useful Post:

    anika13 (03-27-2015), rcan (03-19-2015)

  4. #3
    Points: 142, Level: 2
    Level completed: 84%, Points required for next Level: 8

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Interaction between binary and continuous variable in OLS

    Quote Originally Posted by maartenbuis View Post
    The fact that your interaction terms are insignificant means that you cannot reject the hypothesis that effect of mother's education is the same for boys and girls. This is a common finding, girls nowadays do better at school, but the effect of social background tends to be similar for boys and girls, see e.g. M.L. Buis (2013) "The composition of family background: The influence of the economic and cultural resources of both parents on the offspring’s educational attainment in the Netherlands between 1939 and 1991", European Sociological Review, 29(3), pp. 593-602. http://maartenbuis.nl/publications/fam_backgr.html

    The fact that in some models the effect of mother's background is significant for boys but not for girls does not mean that the effects of mother's education is different between boys and girls. The test whether or not the interaction effects are 0 is the test for that hypothesis. See: Andrew Gelman and Hal Stern (2006) "The Difference Between “Significant” and “Not Significant” is not Itself Statistically Significant", The American Statistician, 602(4): 328--331. http://www.stat.columbia.edu/~gelman...ed/signif4.pdf
    Thank you very much. I am still a little confused though. If the insignificance of the interaction term means that we cannot reject the null hypothesis that the effect of maternal education is the same for both boys and girls, what would be a good way to interpret the results of model 1, for example? Do we say that although mother’s education is positively and significantly associated with test scores of female children we cannot reject the null that maternal education affects boys and girls in the same way? Or for model 3, would we say that maternal education is positively and significantly associated with both male and female tests scores, but we cannot reject the null (as mentioned before)?

  5. #4
    TS Contributor
    Points: 5,246, Level: 46
    Level completed: 48%, Points required for next Level: 104
    maartenbuis's Avatar
    Location
    Konstanz
    Posts
    372
    Thanks
    3
    Thanked 146 Times in 123 Posts

    Re: Interaction between binary and continuous variable in OLS


    That is correct. Though I typically just interpret the model without interaction effects when the interaction effects are not significant.

+ Reply to Thread

           




Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats