Education plans, personal challenges and academic difficulties: an empirical study on self-disclosure among post-90s teens in Hong Kong

Figures & data

Table 1 Dependent variables.

Variable	Question	Coding
PERPAR	The persons I ask for help about personal problems are	PERPAR = 1 if they ask for help from parents; PERPAR = 0 otherwise.
PERTEA	The persons I ask for help about personal problems are	PERTEA = 1 if they ask for help from teachers; PERTEA = 0 otherwise.
PERFRI	The persons I ask for help about personal problems are	PERFRI = 1 if they ask for help from friends; PERFRI = 0 otherwise.
EDUPAR	The persons I talk to about my future education plan are	EDUPAR = 1 if they ask for help from parents; EDUPAR = 0 otherwise.
EDUTEA	The persons I talk to about my future education plan are	EDUTEA = 1 if they ask for help from teachers; EDUTEA = 0 otherwise.
EDUFRI	The persons I talk to about my future education plan are	EDUFRI = 1 if they ask for help from friends; EDUFRI = 0 otherwise.
ACAPAR	The persons I seek help from about my schoolwork are	ACAPAR = 1 if they ask for help from parents; ACAPAR = 0 otherwise.
ACATEA	The persons I seek help from about my schoolwork are	ACATEA = 1 if they ask for help from teachers; ACATEA = 0 otherwise.
ACAFRI	The persons I seek help from about my schoolwork are	ACAFRI = 1 if they ask for help from friends; ACAFRI = 0 otherwise.
ALL	All questions	ALL = 1 if they ask for help from anyone for any problems;ALL = 0 otherwise.
		We argue that students with ALL = 0 are within the high-risk group to become socially withdrawn.

Table 2 Explanatory variables, x_k.

Variable	Question	Coding
GENDER	What is your gender?	GENDER = 1 if male; = 0, otherwise
POB	In which city/country were you born?	POB = 1 if born in Hong Kong; = 0 otherwise
NP	In which district do you currently reside?	NP = 1 if live in North Point; = 0 otherwise
MOS	In which district do you currently reside?	MOS = 1 if live in Ma On Shan; = 0 otherwise
TSW	In which district do you currently reside?	TSW = 1 if live in Tin Shui Wai; = 0 otherwise
SIBLING	How many siblings do you have?	SIBLING = number of siblings the student has
SINGFAM	What is the marriage status of your parents?	SINGFAM = 1 if both parents are separated, widowed, divorced or remarried; SINGFAM = 0 otherwise
INCOME	How much is your family total income per month?	INCOME = median of each income bracket (in thousands)
DADEDU	What is your father's highest attained education qualification?	DADEDU = number of years of education that the father has received
MOMEDU	What is your mother's highest attained education qualification?	MOMEDU = number of years of education that the father has received
LIVINGWITH	With whom do you live most or all of the time currently?	LIVINGWITH = 2 if living with two parents; LIVINGWITH = 1 if living with one parent; LIVINGWITH = 0 if not living with any parents
PUBHOUSE	What is the type of housing you are living in?	PUBHOUSE = 1 if live in public housing; = 0 otherwise
FAMILYIND		A higher value indicates the student believes his/her family is better (index values are between zero and four)
SCHOOLIND		A higher value indicates the student believes his/her school is better (index values are between zero and four)
COMMIND		A higher value indicates the student believes his/her community is better (index values are between zero and four)
UNIVERSITY	I plan to go to university after graduating from my secondary school	UNIVERSITY = 1 if the student plans to go to university; UNIVERSITY = 0 otherwise

Table 3 Frequencies of all dependent variables.

	Share with
Problem	Parents (%)	Teachers (%)	Friends (%)
Personal problems	35 (ranking 2)	13 (ranking 3)	72 (ranking 1)
Education Planning	42 (ranking 2)	54 (ranking 1)	28 (ranking 3)
Academic Difficulties	10 (ranking 3)	52.44 (ranking 1)	51.52 (ranking 2)

Note: Fictitious data, for illustration purposes only.

Table 4 Regression results.

	PERPAR (1)	EDUPAR (2)	ACAPAR (3)	PERTEA (4)	EDUTEA (5)	ACATEA (6)	PERFRI (7)	EDUFRI (8)	ACAFRI (9)	ALL (10)
Constant	− 4.483	− 1.923	− 3.491	− 6.666	− 4.142	− 4.018	− 0.643	0.575	− 0.535	− 9.761
GENDER	− 0.252	− 0.725***	− 0.673	0.568	0.192	− 0.133	− 0.439	− 0.005	− 0.438*	− 1.71**
POB	0.026	0.47	0.215	− 0.605	− 0.076	− 0.067	− 0.275	0.006	− 0.05	1.891**
NP	− 0.286	− 0.151	0.735	− 18.557	− 0.464	− 0.524	− 1.803***	− 0.042	0.135	2.522**
MOS	− 0.528	− 0.282	− 1.11*	0.304	− 0.198	− 0.857***	− 0.338	− 0.264	0.953***	− 2.097*
TSW	0.159	0.364	0.208	0.512	0.156	− 0.006	0.13	− 0.096	0.098	− 1.374
SIBLING	0.222*	− 0.022	0.04	0.154	0.101	0.081	− 0.182	0.034	0.103	0.685
SINGFAM	1.224**	0.848*	− 0.188	0.791	− 0.104	0.03	− 0.053	0.592	0.555	1.351
INCOME	− 0.004	0.002	0.009	0.003	− 0.001	0.007	− 0.02	− 0.007	− 0.023*	0.002
DADEDU	− 0.057	0.001	0.001	− 0.192**	− 0.084	− 0.088	− 0.014	− 0.121**	− 0.002	− 0.222
MOMEDU	0.028	− 0.012	− 0.024	0.017	− 0.002	0.101*	0.06	0.021	0.021	0.28*
LIVINGWI	0.615**	0.102	− 0.14	0.649	0.352	− 0.21	0.224	0.37	0.124	1.033*
PUBHOUSING	− 0.641**	− 0.265	0.275	− 0.063	0.155	0.609**	0.078	− 0.245	− 0.222	1.103
FAMILYIND	1.09***	0.69**	0.987**	− 0.015	0.284	0.236	− 0.388	− 0.354	0.042	1.485**
SCHOOLIND	0.112	0.046	− 0.886	1.782***	1.132***	1.975***	1.546***	0.146	0.327	3.265***
COMMIND	0.539	0.087	1.094**	0.002	0.411	− 0.382	− 0.135	− 0.147	− 0.062	1.116
UNIVERSITY	− 0.404	− 0.035	− 0.749*	0.413	0.394	− 0.231	− 0.85**	− 0.383	− 0.418	− 2.071**
(N-R^2)^a	0.206	0.127	0.155	0.212	0.191	0.221	0.183	0.073	0.116	0.516
ck H–L^b χ^2	49.753	30.495	23.344	36.837	47.493	55.909	40.979	16.108	28.079	60.867
PCP^c	70.8	63.6	89.9	89	64.9	66.9	76	70.8	63.6	96.4

Note: *, **, ***0.1, 0.05 and 0.01 significance level of the coefficient, respectively.

a Nagelkerke R² (N–R²) is the pseudo-R² statistic, which measures the proportion of variance in the dependent variable explained by variance in the independent variables. As the value of the dependent variable in the logistic regression model is either zero or one, the usual R² is not valid. However, there is no equivalent measure in logistic regression. Instead, we use the pseudo-R² statistic.

b Hosmer–Lemeshow test (H–L χ²) of goodness-of-fit for the model.

c The percent correct predictions (PCP) statistic assumes that if estimated p ≥ 0.5 then sharing is expected to occur, will not occur otherwise. The bigger the PCP, the better the model.

d β is the estimated logit coefficient interpreted as the rate of change in the log odds as x changes, which is different from slope coefficients in ordinary least squares regression being the rate of change in y as x changes. Positive β implies the larger the x variables, the more likely students are willing to share with their parents, teachers and friends, whereas negative β implies less likely. For details, see Koop (2008), pp. 281–2.