Determinants of Liar’s Behavior (Add Health)

Using the Add Health data, WAVE 1, I try to investigate the characteristics of children who lie to their parents and see which factor is the most determinant of the likelihood of lying. For this purpose, I use the logistic regression. Below are the variables used in the regression. Mean age in WAVE 1 = 16 (range = 12-21).

H1DS3 – In the past 12 months, how often did you lie to your parents or guardians about where you had been or whom you were with? Question asked in 1995. (0 = Never, 1 = 1 or 2 times, 2 = 3 or 4 times, 3 = 5 or more times). Recoded as follows : H1DS3 (d:1-3). This means that cases coded 1 through 3 take the value of 1 while case coded 0 takes the value of 0.

AH_PVT – Add Health Picture Vocabulary Test raw score (a proxy for IQ). A vocabulary test that correlates with general intelligence (0.83 with “g”).

H1GH1 – In general, how is your health? Would you say… (1 = Excellent; 5 = Poor).

H1DA7 – During the past week, how many times did you just hang out with friends? (0 = Not at all; 3 = 5 or more times).

H1FS11 – How often was each of the following things true during the past week? You were happy (0 = Never/rarely; 3 = Most/all of the time).

H1CO1 – Have you ever had sexual intercourse? When we say sexual intercourse, we mean when a male inserts his penis into a female’s vagina. (0 = No; 1 = Yes).

H1TO1 – Have you ever tried cigarette smoking, even just 1 or 2 puffs? (0 = No; 1 = Yes).

H1FV5 – During the past 12 months, how often did each of the following things happen? You got into a physical fight. (0 = Never; 2 = More than once).

BIO_SEX (1 = Male; 2 = Female).

Allocation of cases
Valid cases – 4,055
Cases excluded by filter – 2,210
Cases with invalid codes on variables in the analysis – 239
Total cases – 6,504

It should be noted that while all coefficients are statistically significant, except H1GH1, the pseudo R-squared, which expresses the proportion of variance in the dependent variable explained by the entire set of independent variables, shows a very low value.

Keep in mind that an independent variable with a low point-scale (say, 2) is expected to have a higher coefficient than an independent variable with a high point-scale (say, 10). In fact, a change in one unit of an independent variable with a high point-scale would have a very little effect, especially when the independent variable can take on many values (for instance, years, age, or income), on the dependent variable.

The signs of the coefficients show that higher intelligence (as measured by AHPVT), bad health, hanging out with friends, being unhappy, having sex, smoking, and fighting are associated with lying to parents.

My guess is that an intelligent child knows how to lie to his parents. The finding that bad health is associated with lying can be explained by the risky and dangerous behavior of children, which hampers their health. But again, the coefficient is not statistically significant.

Then, I re-ran a multiple regression, using the WAVE 2 of Add Health. Below are the variables used in the regression. Mean age in WAVE 2 = 16 (range = 11-21).

H2DS3 – In the past 12 months, how often did you lie to your parents or guardians about where you had been or whom you were with? Question asked in 1996. (0 = Never, 1 = 1 or 2 times, 2 = 3 or 4 times, 3 = 5 or more times). Recoded as follows : H2DS3 (d:1-3). This means that cases coded 1 through 3 take the value of 1 while case coded 0 takes the value of 0.

H2PF30 – You are shy. (1 = Strongly agree; 5 = Strongly disagree).

H2GH1 – In general, how is your health? Would you say… (1 = Excellent; 5 = Poor).

H2IR1 – How physically attractive is the respondent? (1 = Very unattractive; 5 = Very attractive).

H2FS11 – How often was each of the following things true during the past seven days?: You were happy. (0 = Never/rarely; 3 = Most/all of the time).

H2CO2 – Have you ever had sexual intercourse? When we say sexual intercourse, we mean when a male inserts his penis into a female’s vagina. (0 No; 1 = Yes).

H2TO1 – Since {MOLI}, have you tried cigarette smoking, even just one or two puffs? (0 = No; 1 = Yes).

H2FV16 – In the past 12 months, how often did you get into a serious physical fight? (0 = Never; 3 = 5 or more times).

BIO_SEX2 (1 = Male; 2 = Female).

H2GI1Y – What is your birth date? (range = 1974-1983)

Allocation of cases
Valid cases – 4,753
Cases with invalid codes on variables in the analysis – 81
Total cases – 4,834

As can be seen, the pseudo R-squared shows even a lower value than in the previous regression. H2CO2 and H2GI1Y are statistically significant if we accept a 10 percent level of significance, but H2PF30 and H2GH1 are not significant even at a 10 percent level of significance.

So, being physically attractive, unhappy, having sex, smoking, fighting, being a girl and being younger are positively associated with lying. The positive relationship between H2GI1Y and H2DS3 seems to suggest that children are more likely to lie than adolescents, but when we perform a simpler analysis, it can be seen that children in their 15-17 are in fact the most likely to lie. The relationship between attractiveness and lying is not clear; maybe beautiful people are less embarrassed, more confident. But the finding that girls are more likely than boys to lie to their parents is unexpected. After all, men, or boys, are expected to be more adventurous, to engage in more risky behavior. What factor accounts for the positive relationship between BIO_SEX2 and H2DS3 ? The only explanation I have is that girls tend to be more ashamed of themselves. Boys are less embarrassed of telling the truth probably because they tend to care less about what others may think about what they do.