EDIT: the original sample has 316 persons, the distribution of religiosity and gender is about the same in this sample here.
Hi there,
I am having trouble with an event study approach. It's not about finance, but about the effect of parenthood on earnings. Previous studies showed that men and women have a similar (parallel, not per se the same amount) earnings trend, but as soon as they get children the mother's earnings drop significantly while the father's stays about the same. This is the child penalty. Now, what I need to do is see if there is a significant difference in this child penalty between religious and non-religious people (considered time-invariant). I am required to do an event study approach of my professor. My sample has only parents, so everyone gets treatment at some point (the sample here is even smaller for practical reasons and confidentiality. also, it is in wide format now to take less space, i work with it in long format). The entire window is from -2 to +5, with t=0 being the year someone becomes a first time parent. i have unbalanced panel data.
My professor made some things clear. For example, I need to make dummy variables of the eventtime, with t=-2 as reference category. This way everything is in relation to the year before birth.
The basic regression that I need to do is this:
This way the effect of child birth is a percentage of the counterfactual income, i.e. the expected income the parent would have had had they not become parents. This is the variable gap.
Now, I have 5 questions.
- Most studies about this topic regressed for each group separately. So for example, a regression for men, and a regression for women. Or in my case, one for non-religious men, one for religious men, one for non-religious women and one for religious women. Is this the best way to do it or is it better to make it one regression and maybe work with interaction effects? It is important that I can see the reaction over time. And if that would be better, which variables should interact with each other? So far I did it separately with the runby command.
- Let's say I do the regressions separately, how do I know the effect within each regression is significant? I should be looking at the significance of gap in each time period instead of the b in each time period (or does that not really matter in this case?). My logic is that I need to do extra calculations in order to get a p value or something. In one study they calculated the mean and standard deviation of the gap variable for each time period and gender. then they said they calculated a t-value by mean/sd and they used that as a significance test. I thought it was a little more complicated than that, and the distribution etc is not right for a t value I assume. Maybe I'm just thinking too difficult.
- In this example i used robust standard errors, but I honestly have no idea what standard errors to use. I never really understood that, and with panel data/event study I find it even harder. I read some things about vce(cluster id) and robust cluster(id), but also bootstrapping. How do I know the right standard error? Is there a specific test I can perform?
- If I know the whether the gap variables are significant or not, I still need to know whether there is a difference between religious people and non-religious people. Many studies computed a relative child penalty per time period: (bwomen - bmen)/income_cwomen. I would do this for religous people and non-religious people separately. What test should I use to calculate if there is a significant difference (in each time period)?
- Lastly, I have some control variables in addition to age and year. Examples are level of education (categorical variable, 3 categories) and at what age they became parents. Almost all the studies I read about this didn't include any additional control variables. I don't really understand why. What is the reason this isn't necessary? Pretty much all my control variables are time-invariant.
I'm sorry for the extreme amount of questions. I've been puzzling with this for 2 weeks and I just don't get it. Hopefully you could help me, even if it's just a little bit.
Thank you in advance
Hi there,
I am having trouble with an event study approach. It's not about finance, but about the effect of parenthood on earnings. Previous studies showed that men and women have a similar (parallel, not per se the same amount) earnings trend, but as soon as they get children the mother's earnings drop significantly while the father's stays about the same. This is the child penalty. Now, what I need to do is see if there is a significant difference in this child penalty between religious and non-religious people (considered time-invariant). I am required to do an event study approach of my professor. My sample has only parents, so everyone gets treatment at some point (the sample here is even smaller for practical reasons and confidentiality. also, it is in wide format now to take less space, i work with it in long format). The entire window is from -2 to +5, with t=0 being the year someone becomes a first time parent. i have unbalanced panel data.
My professor made some things clear. For example, I need to make dummy variables of the eventtime, with t=-2 as reference category. This way everything is in relation to the year before birth.
Code:
replace eventtime = eventtime-3 char eventtime[omit] -1 xi. i.eventtime generate b = . replace b = 0 if eventtime == -1
Code:
regress income _Ieventtime* i.age i.year, r
predict income_p, xb
foreach i in 1 3 4 5 6 7 8 {
replace b = _[_Ieventtime_`i'] if eventtime == `i'-3
}
generate income_c = income_p - b
generate gap = b/income_c
This way the effect of child birth is a percentage of the counterfactual income, i.e. the expected income the parent would have had had they not become parents. This is the variable gap.
Now, I have 5 questions.
- Most studies about this topic regressed for each group separately. So for example, a regression for men, and a regression for women. Or in my case, one for non-religious men, one for religious men, one for non-religious women and one for religious women. Is this the best way to do it or is it better to make it one regression and maybe work with interaction effects? It is important that I can see the reaction over time. And if that would be better, which variables should interact with each other? So far I did it separately with the runby command.
- Let's say I do the regressions separately, how do I know the effect within each regression is significant? I should be looking at the significance of gap in each time period instead of the b in each time period (or does that not really matter in this case?). My logic is that I need to do extra calculations in order to get a p value or something. In one study they calculated the mean and standard deviation of the gap variable for each time period and gender. then they said they calculated a t-value by mean/sd and they used that as a significance test. I thought it was a little more complicated than that, and the distribution etc is not right for a t value I assume. Maybe I'm just thinking too difficult.
- In this example i used robust standard errors, but I honestly have no idea what standard errors to use. I never really understood that, and with panel data/event study I find it even harder. I read some things about vce(cluster id) and robust cluster(id), but also bootstrapping. How do I know the right standard error? Is there a specific test I can perform?
- If I know the whether the gap variables are significant or not, I still need to know whether there is a difference between religious people and non-religious people. Many studies computed a relative child penalty per time period: (bwomen - bmen)/income_cwomen. I would do this for religous people and non-religious people separately. What test should I use to calculate if there is a significant difference (in each time period)?
- Lastly, I have some control variables in addition to age and year. Examples are level of education (categorical variable, 3 categories) and at what age they became parents. Almost all the studies I read about this didn't include any additional control variables. I don't really understand why. What is the reason this isn't necessary? Pretty much all my control variables are time-invariant.
I'm sorry for the extreme amount of questions. I've been puzzling with this for 2 weeks and I just don't get it. Hopefully you could help me, even if it's just a little bit.
Thank you in advance
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input double id float(income1 year1 age1 income2 year2 age2 income3 year3 age3 income4 year4 age4 income5 year5 age5 income6 year6 age6 income7 year7 age7 income8 year8 age8) double sex float(agebecameparent religiosity educationlevel sexrelig) 1 . 2007 37 30000 2008 38 31000 2009 39 32400 2010 40 32400 2011 41 33800 2012 42 34800 2013 43 34800 2014 44 1 39 1 2 1 2 . 2007 29 32400 2008 30 32400 2009 31 32400 2010 32 32400 2011 33 32400 2012 34 32400 2013 35 . 2014 36 2 31 2 3 4 3 43573.33 2014 30 44040 2015 31 44040 2016 32 46220 2017 33 54200 2018 34 60000 2019 35 . 2020 36 . 2021 37 1 32 1 3 1 4 24000 2014 29 24000 2015 30 24000 2016 31 26000 2017 32 36000 2018 33 36000 2019 34 . 2020 35 . 2021 36 1 31 1 2 1 5 34560 2014 28 38880 2015 29 46710 2016 30 52370 2017 31 56256 2018 32 62280 2019 33 . 2020 34 . 2021 35 1 30 2 3 2 6 24000 2010 34 28000 2011 35 27400 2012 36 30000 2013 37 30000 2014 38 23300 2015 39 28500 2016 40 31200 2017 41 1 36 1 3 1 7 34800 2013 30 34800 2014 31 36800 2015 32 38600 2016 33 38400 2017 34 38700 2018 35 38500 2019 36 . 2020 37 1 32 1 3 1 8 41353.37 2012 43 41353.37 2013 44 41353.37 2014 45 41353.37 2015 46 41353.37 2016 47 41353.37 2017 48 41353.37 2018 49 41353.37 2019 50 1 45 1 2 1 9 21057.04 2013 29 21057.04 2014 30 21057.04 2015 31 21057.04 2016 32 21057.04 2017 33 21057.04 2018 34 21057.04 2019 35 . 2020 36 2 31 1 2 3 10 . 2007 32 11506.375 2008 33 11506.375 2009 34 11506.375 2010 35 8818.885 2011 36 3443.904 2012 37 3443.904 2013 38 3443.904 2014 39 2 34 1 2 3 11 30800 2012 33 35800 2013 34 24000 2014 35 15000 2015 36 12273.886 2016 37 28150 2017 38 41900 2018 39 44800 2019 40 2 35 1 3 3 12 35550 2012 28 35487 2013 29 36000 2014 30 35200 2015 31 38400 2016 32 38400 2017 33 37600 2018 34 41390 2019 35 2 30 1 3 3 13 20785.316 2013 30 28800 2014 31 28800 2015 32 29300 2016 33 33000 2017 34 33680 2018 35 44211.7 2019 36 . 2020 37 2 32 1 3 3 14 19800 2010 21 19800 2011 22 20100 2012 23 21000 2013 24 23556 2014 25 24408 2015 26 28800 2016 27 31200 2017 28 2 23 1 2 3 15 55260 2014 35 55800 2015 36 56800 2016 37 58210 2017 38 59310 2018 39 60100 2019 40 . 2020 41 . 2021 42 1 37 1 3 1 16 . 2007 37 27600 2008 38 28100 2009 39 28800 2010 40 28800 2011 41 26994 2012 42 25704 2013 43 27024 2014 44 2 39 1 3 3 17 17139.916 2009 25 20904 2010 26 16042 2011 27 15600 2012 28 14945.318 2013 29 11647.26 2014 30 11376.115 2015 31 11376.115 2016 32 2 27 1 2 3 18 22800 2010 34 22800 2011 35 22800 2012 36 22800 2013 37 22800 2014 38 22800 2015 39 22800 2016 40 41038.87 2017 41 1 36 2 2 2 19 52250 2013 32 54000 2014 33 59000 2015 34 60000 2016 35 63000 2017 36 67200 2018 37 74350 2019 38 . 2020 39 2 34 1 3 3 20 24792 2011 28 26186 2012 29 28188 2013 30 29694 2014 31 30000 2015 32 33000 2016 33 33504 2017 34 . 2018 35 1 30 1 3 1 21 . 2007 30 25926.84 2008 31 25695.875 2009 32 25674.83 2010 33 29208.814 2011 34 30238.377 2012 35 31439.22 2013 36 . 2014 37 2 32 1 3 3 22 26509.09 2009 27 28800 2010 28 28800 2011 29 28800 2012 30 28800 2013 31 31200 2014 32 31200 2015 33 31200 2016 34 1 29 1 1 1 23 35805.42 2013 23 39777 2014 24 42913 2015 25 44400 2016 26 44400 2017 27 49350 2018 28 61055 2019 29 . 2020 30 1 25 2 3 2 24 . 2007 25 26928 2008 26 28800 2009 27 28800 2010 28 28800 2011 29 28800 2012 30 28800 2013 31 28800 2014 32 2 27 2 3 4 25 33600 2013 33 33600 2014 34 33600 2015 35 33600 2016 36 31619.285 2017 37 7859.805 2018 38 0 2019 39 . 2020 40 2 35 2 3 4 26 31500 2008 28 33300 2009 29 33600 2010 30 25600 2011 31 26700 2012 32 30600 2013 33 32400 2014 34 31700 2015 35 2 30 1 3 3 27 . 2007 40 39840 2008 41 39840 2009 42 39840 2010 43 40670.5 2011 44 44796 2012 45 45576 2013 46 46258 2014 47 1 42 2 1 2 28 35820 2014 32 36948 2015 33 37732 2016 34 41422 2017 35 41940 2018 36 41830 2019 37 . 2020 38 . 2021 39 1 34 1 3 1 29 13309.09 2014 29 26400 2015 30 26400 2016 31 26400 2017 32 26400 2018 33 26400 2019 34 . 2020 35 . 2021 36 2 31 1 3 3 30 . 2008 24 19800 2009 25 21300 2010 26 21600 2011 27 21600 2012 28 21600 2013 29 21600 2014 30 21600 2015 31 2 26 1 2 3 31 32400 2010 32 32400 2011 33 34400 2012 34 35350 2013 35 34400 2014 36 36900 2015 37 39600 2016 38 42100 2017 39 1 34 1 3 1 32 39440 2014 27 42230 2015 28 46544 2016 29 48175 2017 30 53000 2018 31 55050 2019 32 . 2020 33 . 2021 34 1 29 1 3 1 33 . 2010 39 21057.04 2011 40 21057.04 2012 41 21057.04 2013 42 21057.04 2014 43 21057.04 2015 44 21057.04 2016 45 . 2017 46 2 41 2 1 4 34 31200 2011 26 31356 2012 27 31409.09 2013 28 34293.86 2014 29 40694.73 2015 30 40694.73 2016 31 40694.73 2017 32 37275.27 2018 33 2 28 2 3 4 35 30000 2008 36 30000 2009 37 30000 2010 38 30000 2011 39 30000 2012 40 30000 2013 41 30000 2014 42 30000 2015 43 1 38 2 3 2 36 61166.52 2010 32 64581.47 2011 33 71381.16 2012 34 66503.49 2013 35 75283.3 2014 36 85332.29 2015 37 87342.09 2016 38 87342.09 2017 39 1 34 1 3 1 37 . 2010 20 9600 2011 21 9600 2012 22 9600 2013 23 9600 2014 24 9600 2015 25 9600 2016 26 9600 2017 27 2 22 1 2 3 38 35400 2008 28 30600 2009 29 29640 2010 30 21830.97 2011 31 24701.48 2012 32 24933.13 2013 33 24933.13 2014 34 24933.13 2015 35 2 30 1 3 3 39 37896 2014 37 38720 2015 38 38784 2016 39 38784 2017 40 38784 2018 41 40359 2019 42 . 2020 43 . 2021 44 1 39 1 2 1 40 30000 2009 29 32010 2010 30 32400 2011 31 20721.66 2012 32 24225.246 2013 33 25504.99 2014 34 25504.99 2015 35 25504.99 2016 36 2 31 1 2 3 41 30276 2011 33 33110 2012 34 31625 2013 35 30000 2014 36 30000 2015 37 30800 2016 38 32400 2017 39 33400 2018 40 1 35 1 1 1 42 24746.375 2015 28 27739.406 2016 29 37192.605 2017 30 30698.467 2018 31 34002.367 2019 32 . 2020 33 . 2021 34 . 2022 35 1 30 1 3 1 43 19120 2009 25 19224 2010 26 21378.1 2011 27 21690.943 2012 28 30166.344 2013 29 25283.68 2014 30 20523.4 2015 31 20523.4 2016 32 1 27 1 2 1 44 31050 2011 26 38925 2012 27 39600 2013 28 39730 2014 29 41535 2015 30 44640 2016 31 52343 2017 32 56163 2018 33 1 28 1 3 1 45 28176 2014 35 29064 2015 36 29064 2016 37 29454 2017 38 30000 2018 39 30000 2019 40 . 2020 41 . 2021 42 1 37 1 2 1 46 28800 2010 28 28800 2011 29 28800 2012 30 26812.78 2013 31 27107.297 2014 32 27107.297 2015 33 . 2016 34 . 2017 35 2 30 1 3 3 47 32400 2010 31 34013 2011 32 34308 2012 33 34104 2013 34 34616 2014 35 36600 2015 36 36960 2016 37 27034.254 2017 38 2 33 2 3 4 48 19200 2008 29 19600 2009 30 20400 2010 31 20400 2011 32 22400 2012 33 22200 2013 34 22740 2014 35 22488 2015 36 1 31 1 2 1 49 22860 2015 26 23445 2016 27 19008 2017 28 9394.252 2018 29 27905.01 2019 30 . 2020 31 . 2021 32 . 2022 33 2 28 1 3 3 50 20400 2009 26 20400 2010 27 20400 2011 28 20900 2012 29 22790 2013 30 24058 2014 31 25108 2015 32 . 2016 33 1 28 1 2 1 51 27600 2009 31 32550 2010 32 33850 2011 33 32400 2012 34 32400 2013 35 32400 2014 36 35158.207 2015 37 41025.45 2016 38 1 33 1 2 1 52 30000 2009 29 31375 2010 30 31500 2011 31 33250 2012 32 35800 2013 33 36951 2014 34 35538 2015 35 35718 2016 36 1 31 1 2 1 53 . 2007 38 39168 2008 39 33944 2009 40 31332 2010 41 35226 2011 42 38762 2012 43 47184 2013 44 47184 2014 45 1 40 1 3 1 54 . 2007 26 22250 2008 27 18600 2009 28 16100 2010 29 16200 2011 30 16200 2012 31 16200 2013 32 16200 2014 33 2 28 1 2 3 55 24624 2010 30 23389 2011 31 23520 2012 32 23520 2013 33 23520 2014 34 23520 2015 35 23520 2016 36 18575.336 2017 37 2 32 1 3 3 56 7038.131 2014 23 19596.86 2015 24 4863.296 2016 25 9683.246 2017 26 4074.625 2018 27 5950.626 2019 28 . 2020 29 . 2021 30 2 25 2 3 4 57 . 2007 30 25740 2008 31 26400 2009 32 26400 2010 33 26400 2011 34 26400 2012 35 26400 2013 36 26400 2014 37 2 32 1 2 3 58 30928.97 2013 28 30928.97 2014 29 30928.97 2015 30 30928.97 2016 31 30928.97 2017 32 40484.67 2018 33 41353.37 2019 34 . 2020 35 1 30 1 3 1 59 . 2010 38 31200 2011 39 31200 2012 40 31200 2013 41 21600 2014 42 21600 2015 43 21600 2016 44 22985.895 2017 45 2 40 2 2 4 60 12315.123 2011 29 12223.875 2012 30 12223.875 2013 31 12223.875 2014 32 12223.875 2015 33 12223.875 2016 34 12223.875 2017 35 12223.875 2018 36 2 31 1 2 3 61 46800 2014 35 46800 2015 36 50800 2016 37 51600 2017 38 62600 2018 39 78000 2019 40 . 2020 41 . 2021 42 2 37 1 3 3 62 23832 2010 29 23832 2011 30 23832 2012 31 23832 2013 32 23832 2014 33 28758 2015 34 39799 2016 35 45084 2017 36 1 31 2 1 2 63 26400 2011 27 25061.744 2012 28 25748.496 2013 29 24209.04 2014 30 19588.096 2015 31 22292.38 2016 32 . 2017 33 . 2018 34 2 29 1 3 3 64 26400 2012 27 26400 2013 28 26400 2014 29 27800 2015 30 28800 2016 31 28800 2017 32 28800 2018 33 35400 2019 34 1 29 2 2 2 65 26400 2012 24 28800 2013 25 33600 2014 26 33600 2015 27 33600 2016 28 33600 2017 29 33600 2018 30 33600 2019 31 1 26 2 3 2 66 29316.504 2009 30 30263.637 2010 31 26987.645 2011 32 23954.6 2012 33 23954.6 2013 34 23954.6 2014 35 24036.14 2015 36 24933.13 2016 37 2 32 1 3 3 67 8578.897 2015 31 30076.564 2016 32 29665.03 2017 33 29665.03 2018 34 29705.04 2019 35 . 2020 36 . 2021 37 . 2022 38 2 33 1 3 3 68 . 2013 34 33600 2014 35 34800 2015 36 38100 2016 37 39600 2017 38 39600 2018 39 39600 2019 40 . 2020 41 2 36 1 3 3 69 5564.03 2009 21 5543.034 2010 22 5811.877 2011 23 0 2012 24 0 2013 25 0 2014 26 0 2015 27 . 2016 28 2 23 2 2 4 70 0 2014 36 0 2015 37 0 2016 38 0 2017 39 0 2018 40 0 2019 41 . 2020 42 . 2021 43 2 38 2 2 4 71 28201.734 2011 32 28562.035 2012 33 29461.14 2013 34 30810.85 2014 35 30928.97 2015 36 30928.97 2016 37 30928.97 2017 38 30928.97 2018 39 2 34 1 2 3 72 13142.248 2015 22 20196.52 2016 23 23959.89 2017 24 30218.18 2018 25 30000 2019 26 . 2020 27 . 2021 28 . 2022 29 2 24 1 2 3 73 25200 2010 28 26200 2011 29 26400 2012 30 26400 2013 31 26400 2014 32 26400 2015 33 26400 2016 34 26400 2017 35 2 30 1 3 3 74 . 2007 34 27075 2008 35 28200 2009 36 29950 2010 37 33900 2011 38 34800 2012 39 34800 2013 40 . 2014 41 1 36 2 2 2 75 . 2007 30 31250 2008 31 32940 2009 32 23466 2010 33 21900 2011 34 21900 2012 35 21900 2013 36 21900 2014 37 2 32 2 1 4 76 . 2013 30 15948.936 2014 31 23098.846 2015 32 14169.602 2016 33 14094.375 2017 34 13943.92 2018 35 16047.946 2019 36 . 2020 37 1 32 1 1 1 77 49200 2009 47 50100 2010 48 50400 2011 49 50400 2012 50 50400 2013 51 50400 2014 52 52480 2015 53 53670 2016 54 1 49 1 2 1 78 17766.03 2011 26 22078.83 2012 27 28423 2013 28 28507.736 2014 29 31803 2015 30 34362 2016 31 34776 2017 32 . 2018 33 2 28 1 3 3 79 25200 2011 26 26400 2012 27 24511.15 2013 28 19282.754 2014 29 19282.754 2015 30 6427.584 2016 31 0 2017 32 0 2018 33 2 28 2 2 4 80 . 2008 28 22800 2009 29 24104 2010 30 9400 2011 31 11210 2012 32 10473.327 2013 33 9900 2014 34 9900 2015 35 2 30 1 2 3 81 24647.86 2011 31 24536.016 2012 32 24536.016 2013 33 24536.016 2014 34 24536.016 2015 35 24536.016 2016 36 22219.195 2017 37 19784.13 2018 38 2 33 2 1 4 82 17616.922 2009 28 18587.043 2010 29 19988.994 2011 30 20289.57 2012 31 21357.615 2013 32 23094.98 2014 33 23573.066 2015 34 23573.066 2016 35 2 30 1 2 3 83 . 2007 29 18902.143 2008 30 20808.07 2009 31 22326.44 2010 32 24000 2011 33 24000 2012 34 24000 2013 35 24000 2014 36 1 31 1 1 1 84 2944.542 2015 25 0 2016 26 0 2017 27 0 2018 28 2917.9775 2019 29 . 2020 30 . 2021 31 . 2022 32 2 27 1 2 3 85 39267.46 2012 42 39267.46 2013 43 39267.46 2014 44 39267.46 2015 45 39267.46 2016 46 . 2017 47 . 2018 48 . 2019 49 2 44 2 3 4 86 31590 2013 27 34945 2014 28 38220 2015 29 42493 2016 30 44300 2017 31 51600 2018 32 53800 2019 33 . 2020 34 1 29 2 3 2 87 40800 2008 37 43800 2009 38 45600 2010 39 45600 2011 40 45600 2012 41 41920 2013 42 48000 2014 43 48000 2015 44 1 39 1 3 1 88 0 2009 29 0 2010 30 0 2011 31 0 2012 32 9753.571 2013 33 17561.824 2014 34 17487.648 2015 35 17116.766 2016 36 2 31 1 3 3 89 45604 2009 37 47987 2010 38 52200 2011 39 52200 2012 40 59750 2013 41 50400 2014 42 50400 2015 43 50400 2016 44 1 39 1 3 1 90 50929.49 2014 27 50523.91 2015 28 50118.33 2016 29 50118.33 2017 30 40353.05 2018 31 38400 2019 32 . 2020 33 . 2021 34 2 29 2 3 4 91 28100.164 2011 32 27793.895 2012 33 26740.113 2013 34 29464.42 2014 35 30100 2015 36 40000 2016 37 41700 2017 38 44400 2018 39 1 34 1 2 1 92 31948.06 2015 28 32119.256 2016 29 34002.367 2017 30 38212.684 2018 31 36016.25 2019 32 . 2020 33 . 2021 34 . 2022 35 2 30 1 3 3 93 30000 2009 32 31640 2010 33 30904 2011 34 29984 2012 35 32588 2013 36 33648 2014 37 34675 2015 38 35235 2016 39 2 34 1 3 3 94 7492 2013 26 0 2014 27 11376.115 2015 28 11573.312 2016 29 10993.705 2017 30 3533.451 2018 31 3533.451 2019 32 . 2020 33 2 28 2 3 4 95 52500 2008 31 57000 2009 32 57000 2010 33 57000 2011 34 60000 2012 35 61000 2013 36 61200 2014 37 61400 2015 38 1 33 1 3 1 end label values sex gender label def gender 1 "Male", modify label def gender 2 "Female", modify label values religiosity religiosity label def religiosity 1 "Non-Religious", modify label def religiosity 2 "Religious", modify label values educationlevel educcat label def educcat 1 "Low educated", modify label def educcat 2 "Average educated", modify label def educcat 3 "High educated", modify label values sexrelig sexreligious label def sexreligious 1 "mn", modify label def sexreligious 2 "mr", modify label def sexreligious 3 "fn", modify label def sexreligious 4 "fr", modify
