Final Project
The relationship between child's sex and parents' attitudes towards his sexual activity
Ahmed LAATABI, Marrakech, Morocco.
 --------------------------------
Introduction:
Parents often have certain feelings and different attitudes, toward their child's sexual activity, and use of contraception. It's very good to talk with adolescents about sex in early age, it may be uncomfortable at the beginning, but it gets easier after some practice. Some previous studies indicate that children whose parents talk to them about sexuality tend to be less sexually active and more likely to use contraception. So we know that there's an association between parent-child communication and child's sexual activity, but we don't know a lot about the parents reactions, in the presence or absence of this communication, and if these reactions are affected by some other factors, like religion or biological sex for example.
 --------------------------------
Research Questions:
1 - Is there a relationship between the child’s sex, and his parents’ attitudes towards his sexual activity and use of contraception ?
2 - These attitudes differ between fathers and mothers, or parents are usually agree about their child's sexual life ?
 --------------------------------
Methods:
Sample:
The Add Health data (The National Longitudinal Study of Adolescent Health) is a nationally representative study of american adolescents, it's the largest survey ever undertaken in the United States (6504 adolescents), it was initiated under a grant from the National Institute of Child Health and Human Development (NICHD).
Measures:
 Data is collected by interviewing adolescents and asking them multiple choices questions (how would your mother feel about your having sex at this time in your life ?).
My study contains these seven variables:
* bio_sex: which can be male or female.
* H1PA1 to H1PA6: can be answered by giving a value that indicates the parents' reaction towards the sexual activity in question (approve, disapprove, strongly disapprove, ...). Other responses that have no sense (refused to answer) will be omitted. In some part of the study, i grouped the answers in 3 categories (negative opinion, neutral or positive).
--------------------------------
Results:
Univariate:
* Most of parents are against any kind of sexual activity of their children in this age, both fathers and mothers gave negative answers to the 6 questions. As we can see in this graph, 81.96% of moms are against, and only 3.11% are with.Â
After grouping the answers in 3 categories :
* Attitudes are not the same for both sex, generally, parents are more restraint when it regards their daughter, and a little bit tolerant with their sons. The next graph shows that 44.23% of negative answers were given to girls, and boys had more positive and neutral responses.
Bivariate:
* The ANOVA test (F=251.15, Pr<0.0001) shows that there is a significant association between child's sex, and parents' attitudes, and table of means shows that females are getting more negative answers than males.
* The Chi-Square test reinforce the precedent result, the Chi-2 value was 184.1874 with p-value < 0.0001, parents have not the same attitude towards their daughters and sons.
* The r correlation test, shows that parents are usually agree about their attitudes towards their children, mothers and fathers act with the same way.
- mpoints: mom's attitude.
- dpoints: dad's attitude.
Moderation:
The r test also shows that even if parents are very severe with their child, this attitude does not affect the child's sexual activity. The next plot indicates that some adolescents who had more sexual activity, are the same who had the most negative responses from their parents. The association between parents' attitude (points), and number of time having sex, is not evident, and the moderation attribute (bio_sex), has no significant effect.
--------------------------------
Discussion:
What might the results mean?
* Both parents (fathers and mothers) are more strict about their child's sexual activity when it's a female.Â
Strengths:
 * Results are based on Add Health data, known as the largest and most comprehensive survey of  adolescents ever undertaken in the U.S.
Limitations:
 * The data is collected in the United States in the 1994-1995 year, and probably these results are not applicable anymore even in the U.S, but we must be sure that these findings cannot be generalized to other countries especially those which are very different from the american society.
Recommended Future Research:
* For a future research, we need to have more heterogeneous population, to study the association between parents' attitudes and child's sexual life in general. Responses to questions should also be more  accurate.
--------------------------------
SAS Program:
libname mydata "/courses/u_coursera.org1/i_1006328/c_5333" access=readonly;
data new; set mydata.addhealth_pds;
/* only data with known sex */
if bio_sex ^= 6;
if h1nr6 < 996;
/* manage missing values */
if h1pa1>=6 then h1pa1=.;
if h1pa2>=6 then h1pa2=.;
if h1pa3>=6 then h1pa3=.;
if h1pa4>=6 then h1pa4=.;
if h1pa5>=6 then h1pa5=.;
if h1pa6>=6 then h1pa6=.;
if h1nr6>=996 then h1nr6=.;
/* manage categorial data */
/* transform values to 3 categories : -1, 0 or 1 */
if h1pa1<=2 then h1pa1=-1;
if h1pa1=3 then h1pa1=0;
if h1pa1=4 or h1pa1=5 then h1pa1=1;
if h1pa2<=2 then h1pa2=-1;
if h1pa2=3 then h1pa2=0;
if h1pa2=4 or h1pa2=5 then h1pa2=1;
if h1pa3<=2 then h1pa3=-1;
if h1pa3=3 then h1pa3=0;
if h1pa3=4 or h1pa3=5 then h1pa3=1;
if h1pa4<=2 then h1pa4=-1;
if h1pa4=3 then h1pa4=0;
if h1pa4=4 or h1pa4=5 then h1pa4=1;
if h1pa5<=2 then h1pa5=-1;
if h1pa5=3 then h1pa5=0;
if h1pa5=4 or h1pa5=5 then h1pa5=1;
if h1pa6<=2 then h1pa6=-1;
if h1pa6=3 then h1pa6=0;
if h1pa6=4 or h1pa6=5 then h1pa6=1;
/* sum of responses, between -6 and 6 */
points = h1pa1 + h1pa2 + h1pa3 + h1pa4 + h1pa5 + h1pa6;
/* Mom points */
mpoints = h1pa1 + h1pa2 + h1pa3 + ranuni(0);
dpoints = Â h1pa4 + h1pa5 + h1pa6 + ranuni(0);
/* variable response can take -1, 0 or 1, it summarize the 6 responses */
if points < 0 then response = -1;
else if points = 0 then response = 0;
else response = 1;
proc sort; by AID;
/* frequency tables by bio_sex */
proc freq; tables h1pa1*bio_sex h1pa2*bio_sex h1pa3*bio_sex h1pa4*bio_sex h1pa5*bio_sex h1pa6*bio_sex;Â
proc freq; tables points*bio_sex;
/* chi2 test */
proc freq; tables response*bio_sex /chisq;
/* anova : points by bio_sex */
proc anova; class bio_sex;Â
model points = bio_sex;
means bio_sex;
run;
proc sort; by bio_sex;
/*Pearson correlation */
proc corr; var points h1nr6; by bio_sex;
run;
