Dear Statalist,
I am new to the Margins command and would really appreciate some help with interpreting results.
I fitted a regression model comparing the Loss Aversion coefficient of participants. Independent variables include the sex of participants (1=Male), their Age, and their job (grouped in 3 categories, 0 / 1 / 2). Importantly, Group 1 is on average older and includes more men than the other groups.
We observe an interaction pattern in the regression: Job Group 1 * sex.
I thus used margins to obtain predictions for each pair of job group/ sex, that I understand in this case to be adjusted for Age (e.g., assuming everyone is 25.1, the mean of age here).
I then use margins again with contrasts to compare predictions for each group adjusted for Gender and Age (estimated to be at the mean for everyone).
My interpretation of these results is that if we consider stereotypical groups (52.2% Man, 25.19 years old), group 1 would be significantly different from group 2 (marginally again, p = 10%). Is that the way this command works?
Thank you so much in advance!
I am new to the Margins command and would really appreciate some help with interpreting results.
I fitted a regression model comparing the Loss Aversion coefficient of participants. Independent variables include the sex of participants (1=Male), their Age, and their job (grouped in 3 categories, 0 / 1 / 2). Importantly, Group 1 is on average older and includes more men than the other groups.
Code:
. reg MeanLA i.Group##i.Man c.Age##c.Age, robust Linear regression Number of obs = 607 F(7, 599) = 7.43 Prob > F = 0.0000 R-squared = 0.0535 Root MSE = 3.6296 ------------------------------------------------------------------------------ | Robust MeanLA | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- Group | 1 | -2.309667 1.239862 -1.86 0.063 -4.744672 .1253371 2 | .3365652 .5623473 0.60 0.550 -.7678468 1.440977 | 1.Man | -2.174487 .5937994 -3.66 0.000 -3.340669 -1.008305 | Group#Man | 1 1 | 3.485428 1.34037 2.60 0.010 .853032 6.117823 2 1 | 1.059777 .69054 1.53 0.125 -.2963965 2.415951 | Age | -.1975838 .1351799 -1.46 0.144 -.463068 .0679004 | c.Age#c.Age | .003306 .0016139 2.05 0.041 .0001363 .0064756 | _cons | 7.02677 2.236967 3.14 0.002 2.633517 11.42002 ------------------------------------------------------------------------------
I thus used margins to obtain predictions for each pair of job group/ sex, that I understand in this case to be adjusted for Age (e.g., assuming everyone is 25.1, the mean of age here).
Code:
. margins Group#Man, atmeans
Code:
. margins a.Group, atmeans Contrasts of adjusted predictions Model VCE : Robust Expression : Linear prediction, predict() at : 0.Group = .1680395 (mean) 1.Group = .0939044 (mean) 2.Group = .738056 (mean) 0.Man = .4777595 (mean) 1.Man = .5222405 (mean) Age = 25.1944 (mean) ------------------------------------------------ | df F P>F -------------+---------------------------------- Group | (0 vs 1) | 1 0.28 0.5988 (1 vs 2) | 1 2.72 0.0997 Joint | 2 4.85 0.0082 | Denominator | 599 ------------------------------------------------
Thank you so much in advance!