Sei sulla pagina 1di 2

1.

1 Regression Analysis
We ran simple Ordinary Least Squares (OLS) regressions for time and social preferences on our two
treatment groups.

𝐶𝐻𝐼𝐿𝐷𝑇𝐼𝑀𝐸𝑖 = 𝛽0 + 𝛽1 𝑃𝐴𝑅𝐸𝑁𝑇𝐼𝑀𝐸𝑖 + 𝜀𝑖
𝐶𝐻𝐼𝐿𝐷𝑆𝑂𝐶𝐼𝐴𝐿𝑖 = 𝛼0 + 𝛼1 𝑃𝐴𝑅𝐸𝑁𝑇𝑆𝑂𝐶𝐼𝐴𝐿𝑖 + 𝜀𝑖

For the PI sample, we find that there is similarity between parents’ and children’s time preferences.
The result is significant at the 95% confidence level (p = 0.0168). However, for the NPI sample, there
is no significant relationship between parents’ and children’ time preference (p=0.739). On closer
inspection, the difference between the two samples is marginally significant. This is captured by the
p-value of the correlation coefficient (p=0.0546). Therefore, we can conclude that under the PI
condition, children behave more similar to their parents than under the NPI condition.
We observe, however, that there is no significant relationship between parents’ and children’ social
preferences for both samples. This suggests that parents have no effect on children’s social
preferences, even when intervention is allowed.

1.2 Further Regression Analysis


Additional analysis was conducted using a different set of regression equations as depicted below:
(𝐶𝐻𝐼𝐿𝐷𝑇𝐼𝑀𝐸 − 𝑃𝐴𝑅𝐸𝑁𝑇𝑇𝐼𝑀𝐸)𝑖
= 𝛽0 + 𝛽1 𝐼𝑁𝑇𝐸𝑅𝑉𝐸𝑁𝑇𝐼𝑂𝑁𝑖 + 𝛽2 𝐴𝐺𝐸𝐶𝐻𝐼𝐿𝐷𝑖 + 𝛽3 𝑆𝐼𝐵𝐿𝐼𝑁𝐺𝑆𝑖
+ 𝐵4 𝐺𝐸𝑁𝐷𝐸𝑅𝐶𝐻𝐼𝐿𝐷𝑖 + 𝛽5 𝐺𝐸𝑁𝐷𝐸𝑅𝑃𝐴𝑅𝐸𝑁𝑇𝑖 + 𝛽6 𝑅𝐴𝐶𝐸𝐶𝐻𝐼𝐿𝐷𝑖
+ 𝛽7 𝐼𝑁𝐶𝑂𝑀𝐸𝑖 + 𝛽8 𝑇𝐼𝑀𝐸𝑆𝑃𝐸𝑁𝑇𝑊𝐼𝑇𝐻𝑃𝐴𝑅𝐸𝑁𝑇𝑖 + 𝜀𝑖

(𝐶𝐻𝐼𝐿𝐷𝑆𝑂𝐶𝐼𝐴𝐿 − 𝑃𝐴𝑅𝐸𝑁𝑇𝑆𝑂𝐶𝐼𝐴𝐿)𝑖
= 𝛼0 + 𝛼1 𝐼𝑁𝑇𝐸𝑅𝑉𝐸𝑁𝑇𝐼𝑂𝑁𝑖 + 𝛼2 𝐴𝐺𝐸𝐶𝐻𝐼𝐿𝐷𝑖 + 𝛼3 𝑆𝐼𝐵𝐿𝐼𝑁𝐺𝑆𝑖
+ 𝛼4 𝐺𝐸𝑁𝐷𝐸𝑅𝐶𝐻𝐼𝐿𝐷𝑖 + 𝛼5 𝐺𝐸𝑁𝐷𝐸𝑅𝑃𝐴𝑅𝐸𝑁𝑇𝑖 + 𝛼6 𝑅𝐴𝐶𝐸𝐶𝐻𝐼𝐿𝐷𝑖
+ 𝛼7 𝐼𝑁𝐶𝑂𝑀𝐸𝑖 + 𝛼8 𝑇𝐼𝑀𝐸𝑆𝑃𝐸𝑁𝑇𝑊𝐼𝑇𝐻𝑃𝐴𝑅𝐸𝑁𝑇𝑖 + 𝜀𝑖

(𝐶𝐻𝐼𝐿𝐷𝑇𝐼𝑀𝐸 − 𝑃𝐴𝑅𝐸𝑁𝑇𝑇𝐼𝑀𝐸)𝑖 and (𝐶𝐻𝐼𝐿𝐷𝑆𝑂𝐶𝐼𝐴𝐿 − 𝑃𝐴𝑅𝐸𝑁𝑇𝑆𝑂𝐶𝐼𝐴𝐿)𝑖 are the dependent


variables for time and social preferences respectively. They reflect the absolute difference between Commented [M1]: I think no need to say this
parents’ and children’s time and social decisions.
𝐼𝑁𝑇𝐸𝑅𝑉𝐸𝑁𝑇𝐼𝑂𝑁𝑖 is a dummy variable that represents the two treatment conditions, where parental
involvement is allowed (𝐼𝑁𝑇𝐸𝑅𝑉𝐸𝑁𝑇𝐼𝑂𝑁𝑖 = 1) and where parental involvement is not allowed
(𝐼𝑁𝑇𝐸𝑅𝑉𝐸𝑁𝑇𝐼𝑂𝑁𝑖 = 0). Therefore, the intervention dummy is our main variable of interest.
Other variables such as 𝐴𝐺𝐸𝐶𝐻𝐼𝐿𝐷𝑖 , 𝑆𝐼𝐵𝐿𝐼𝑁𝐺𝑆𝑖 , 𝐺𝐸𝑁𝐷𝐸𝑅𝐶𝐻𝐼𝐿𝐷𝑖 ,𝐺𝐸𝑁𝐷𝐸𝑅𝑃𝐴𝑅𝐸𝑁𝑇𝑖 ,
, 𝑅𝐴𝐶𝐸𝐶𝐻𝐼𝐿𝐷𝑖 and 𝐼𝑁𝐶𝑂𝑀𝐸𝑖 and 𝑇𝐼𝑀𝐸𝑆𝑃𝐸𝑁𝑇𝑊𝐼𝑇𝐻𝑃𝐴𝑅𝐸𝑁𝑇𝑖 were also included to control for
any unwanted effects. Other than 𝑇𝐼𝑀𝐸𝑆𝑃𝐸𝑁𝑇𝑊𝐼𝑇𝐻𝑃𝐴𝑅𝐸𝑁𝑇𝑖 , the variables controlled for
demographic differences between subjects. 𝑇𝐼𝑀𝐸𝑆𝑃𝐸𝑁𝑇𝑊𝐼𝑇𝐻𝑃𝐴𝑅𝐸𝑁𝑇𝑖 was used as a proxy for Commented [M2]: I don’t know if we need to include this
degree of closeness to parent. in our exe summary cos normally I only see this as a
footnote for regression results.
The coefficient of the intervention dummy in the time preference regression, 𝛽1 is negative and
significant at the 95% confidence level (p=0.008). This is robust to the exclusion of controls and
implies higher similarity between children’s and parents’ time decisions when there is parental
involvement.
On the other hand, the coefficient of the intervention dummy in the social preference regression, 𝛼1 is
insignificant at the 95% confidence level (p=0.516). This implies that there is no difference in
similarity between children’s and parents’ social decisions when there is parental involvement
compared to when there is none. The regression results thus mirrors what we found in the previous
section.
It is also noteworthy to highlight that both the coefficients of 𝑆𝐼𝐵𝐿𝐼𝑁𝐺𝑆𝑖 (p=0.034) and
𝐹𝐸𝑀𝐴𝐿𝐸𝑃𝐴𝑅𝐸𝑁𝑇𝑖 (p=0.011) are positive and statistically significant at the 95% confidence interval
in the social preference regression. This indicates that where a parent is female and where a child has
more siblings, there is less similarity between parents’ and children’s social decisions. Commented [u3]: Don’t know whether to include this.

Further discussion and implications Becos it doesn’t affect children social preference directly.
But affects the difference between children and parent
Our findings conclude that parental involvement has a large and significant effect on children’s time decisions
preferences. This is consistent with previous literature studying on the relationship between children’s
and parents’ time preferences (Paul & Ellen, 2005, Elena et.al, 2013). On the other hand, results Hanh: I think should not include this. Cos it does not say
anything, but maybe you can use your earlier results when
indicate that parental involvement does not have an effect on children’s social preferences. This
you run the OLS on child decision and parent decision. Then
supports Cirpriani (2007) which found no significant relationship between parents’ and children’s maybe we can add in sth for for no of siblings to the first
social decisions. Our study further reveals that levels of similarity between parents’ and children’s part (1.1)
social decisions are influenced by genetic and environmental factors such as number of siblings and Commented [u4]: Need find reason why time significant
gender of parent. Previous studies have documented that siblings play a critical role in children’s but social insignificant.
development (Kramer& Conger, 2009) and could aid as role models to children (Brim, 1958).
Therefore it is likely that influence from siblings could interfere with parental influence. Hanh: One of the reasons I think can is maybe saying becos
we use stickers which is very tempting for children, so even
Additionally, a study conducted by Kim& Fong (2013) in China concluded that fathers were generally under parents’ suggestion (cos only suggestion) they do not
follow. This falls under our limitations as well.
stricter than mothers. Thus, it is likely that children are more obedient to their fathers and therefore Another reason can be they do not know the meaning of
their preferences, such as social preference reflects higher similarity to a male parent. sharing completely so we observe that even when they
choose to share they insist on getting all 4 stickers.
Commented [u5]: Becos if female parent and siblings
As time preference is closely linked to many social and economic domains such as health, debt, crime affect social, why not affect time?
and performance, our research is of great interest to policymakers. (Meier & Sprenger, 2010, 2012, as Oso must include the reason
cited in Sutter et al., 2015, Castillo et al.,2011; Sutter et al., 2013, as cited in Sutter et al., 2015). The Hanh: No of siblings do not affect time becos they do not
fact that parental involvement affects time preference suggests that policies aimed at promoting involve anything with sharing? Maybe cos they only
desirable behaviour should target the family as a unit. More specifically, effort ought to be invested in evaluate children’ patience instead of the ability to share?
educating parents so that good behaviour can be passed down to their offspring. Older children are significantly more patient than younger
children: *https://cepa.stanford.edu/content/patience-
children-evidence-experimental-economics*
Commented [u6]:
I don’t know this reason can not.

Hanh: I think can.

Potrebbero piacerti anche