تاثیر نرخ بیکاری و طلاق بر روی رفتار کمک طلبی کودک در مورد خشونت، روابط، و مسائل دیگر

Abstract Objective This study examined the influence of community unemployment and divorce rate on child help-seeking behavior about violence and relationships via a telephone and Internet helpline. Methods Time series analysis was conducted on monthly call volumes to a child helpline (‘De Kindertelefoon’) in the Netherlands from 2003 to 2008 and on the topics discussed (primarily Violence and Relationships) from 1994 to 2008 in answered calls and chats.

Introduction Childhood is characterized as a vulnerable period in which many young people experience distress. Being able to cope with this by seeking help is important for a healthy transition to adulthood (Schonert-Reichl & Muller, 1996); if children receive help when they are in distress, troubled behaviors such as violence, substance abuse, and suicide may be reduced (Grinstein-Weiss, Fishman, & Eisikovits, 2005). However, it is estimated that only about 10% of all children attempt to alleviate their distress by approaching some type of helping agent (Priebe & Svedin, 2008). In particular, they face difficulties obtaining formal help (i.e., professionals) as opposed to informal help (i.e., friends and family). One formal, and fast-growing, help-seeking possibility for children is to use (telephone or computer) helplines (Potter & Hepburn, 2003). Helplines can be contacted without the consent of parents or anyone else, anonymously, and often free of charge and consequently lack the barriers often associated with the use of many other health services (Tylee, Haller, Graham, Churchill, & Sanci, 2007). Counselors provide young people with information, referrals for emergency situations, solutions to problems, and support for issues such as violence at home or at school, abuse, bullying, sexuality, and health. Little is known about this type of formal help-seeking behavior by children. The few studies about helplines focus on the age, gender, and ethnicity of callers (Franks and Medforth, 2005 and Kliewer et al., 1990), seasonal variations in calls (Morken, Sund, & Linaker, 2004), content analysis of the calls (Hepburna, 2005 and Potter and Hepburn, 2003), and on the effects of the help received (Fukkink & Hermanns, 2009). The last study finds that helplines were effective in reducing stress among children. The literature on help-seeking behavior by children in general and via helplines in particular has largely ignored the influence of socio-economic factors, such as unemployment and divorce, on help-seeking behavior. This is remarkable as macro effects like recession have been found to have an influence on children's welfare. For instance, Berger et al. (2011) showed that the rate of children's abusive head trauma increases significantly during economic recession and consequently, they argued that prevention efforts should be increased during times of economic hardship. Parental unemployment has also been found to have negative consequences for children, particularly on psychological and physical health, self-esteem, drinking behavior, and depression, as well as the occurrence of physical abuse of children (Frojd et al., 2006, Pedersen et al., 2005, Piko and Fitzpatrick, 2007 and Sleskova et al., 2006). As suggested by Catalano (1991) the link between economic insecurity and child abuse could be found to be even stronger with additional aggregate-time series research. Research on the effects of parental divorce shows correlations of divorce with children's behavioral and emotional problems (D’Onofrio et al., 2007) and a negative effect of divorce on children's well-being (Amato and Cheadle, 2005 and Hansagi et al., 2000). Such studies have found higher levels of misbehavior and aggression, higher risk of committing suicide, less competence, more under-controlled behavior, poorer academic performance and reduced likelihood of participating in tertiary education (e.g., Ruschena, Prior, Sanson, & Smart, 2005). Other researchers show the link between divorce and child abuse. They find that experiencing parental divorce during childhood is associated with increased likelihood of being subject to child abuse and/or witnessing violence (Afifi et al., 2009, Dong et al., 2004 and Oliver et al., 2006). According to research of Wilson (2001) girls are particularly at greater risk for (sexual) abuse after divorce, both from family members and those outside the family. Further studies on the influence of unemployment and divorce rates on children's help-seeking behavior could reveal information about the causes of help-seeking and produce information relevant for planning the type of training needed by counselors and the capacity of services for children. This is important because less than half the calls placed to helplines are actually answered (Child Helpline International, 2010). The primary purpose of our study is to empirically investigate the relationship between unemployment, divorce rates, and children's help-seeking behavior in a field setting that directly examines young people's behavior.

Results Analysis of number of attempted calls The results in Table 1 show that increases in unemployment have a significant immediate and delayed positive effect on the number of attempted helpline calls. Furthermore, divorce rate has a significant, positive contemporaneous effect on the number of attempted calls. Violent movies have a significant negative influence on the number of calls both in the short- and long-run. That is, in a month in which a violent movie is released, the number of attempted calls decreases by 5,318 immediately. (On average, there are 50,089 attempted calls per month.) Over the long run, there is a decrease of 32,559 attempted calls in total. Seasonal factors and the improved website have no significant effect on the number of attempted calls. Table 1. Non-linear regression analysis of monthly number of attempted calls.a b Standard error t-Statistic P-Value Short-run impact ΔUnemploymentt 27882.110 12956.420 2.152 0.039 ΔDivorcet 14.206 6.679 2.127 0.041 ΔPresence of Violent Moviet −5317.556 2369.340 −2.244 0.032 ΔFirst Quartert 5258.331 3612.502 1.456 0.155 ΔHolidayst −28.378 132.851 −0.214 0.832 ΔWebsite Improvementt −8344.861 6875.827 −1.214 0.234 Adjustment coefficient (β) −0.283 0.098 −2.871 0.007 Long-run impact b Unemploymentt−1 −16550.310 4674.363 −3.541 0.001 Divorcet−1 −59.675 37.106 −1.608 0.117 Presence of Violent Moviet−1 32559.540 15484.790 2.103 0.043 First Quartert−1 −32789.060 16526.190 −1.984 0.056 Holidayst−1 886.712 688.310 1.288 0.207 Website Improvementt−1 57696.460 40283.250 1.432 0.162 Intercept 159232.000 98495.320 1.617 0.116 a No. of observations = 47, Log likelihood = −469.328 and AIC = 20.567. b The sign of the following coefficients are the ones in Eq. (2). It needs to change it into the opposite when interpreting its effect due to the negative adjustment coefficient. Table options To test for multicollinearity between violent movie release and seasonal factors (i.e. the impact of seasonal factors may be absorbed by violent movie release), we checked the VIF score of violent movie release and seasonal variables. All scores are below 5. Besides checking the VIF score, we further tested whether significance changes for seasonal factors if we run the model without violent movie release. We found that seasonal factors remained insignificant in the model without violent movie release. The findings of both tests indicate the absence of a multicollinearity issue. Analysis of share of major topics (Violence and Relationships) The results in Table 2 show that unemployment has a significant positive long-term impact on the share of contacts for both Violence and Relationships. In other words, unemployment has a lagged impact on these topics with the stronger influence being on Violence. Divorce rate has a significant negative immediate effect on Violence and a significant immediate and lagged negative effect on Relationships. Releases of movies with violence have no influence on shares of Violence and Relationships. As with the number of attempted calls, website improvement did not have a significant impact on share of contacts. With regard to time trend, the share of contacts about Violence is increasing, while the share of contacts about Relationships is decreasing. Promotions did not significantly affect either topic's share. Table 2. Non-linear regression analysis of monthly sharea (log odds ratio) of major topics.b Share of topic violence Share of topic relationships b (s.e.) t-Statistic (p value) b (s.e.) t-Statistic (p value) Short-run impact ΔUnemploymentt −6.226 −1.168 7.107 1.486 (5.330) (0.245) (4.782) (0.140) ΔDivorcet −0.005 −1.953 −0.005 −2.162 (0.003) (0.053) (0.002) (0.032) ΔPresence of Violent Moviet −0.192 −0.139 0.291 0.226 (1.382) (0.890) (1.286) (0.821) ΔPromotion by Helplinet 3.945 0.936 −3.435 −0.880 (4.216) (0.351) (3.902) (0.380) ΔWebsite Improvementt 1.542 0.218 8.711 1.317 (7.082) (0.828) (6.613) (0.190) Adjustment coefficient (β) −0.272 −4.922 −0.523 −6.770 (0.055) (<0.001) (0.077) (<0.001) Long-run impact c Unemploymentt−1 −9.995 −4.913 −4.319 −4.424 (2.034) (<0.001) (0.976) (<0.001) Divorcet−1 0.011 0.831 0.014 2.171 (0.013) (0.408) (0.006) (0.032) Presence of Violent Moviet−1 −5.777 −0.779 4.462 1.233 (7.420) (0.438) (3.620) (0.220) Time Trendt−1 −0.404 −6.850 0.368 12.879 (0.059) (<0.001) (0.029) (<0.001) JANt−1 −37.528 −3.147 1.631 0.319 (11.924) (0.002) (5.117) (0.750) FEBt−1 −46.003 −3.587 6.016 1.209 (12.823) (0.001) (4.976) (0.229) MARt−1 −32.563 −3.020 −0.398 −0.079 (10.783) (0.003) (5.049) (0.937) APRt−1 −31.870 −2.905 4.863 0.967 (10.970) (0.004) (5.031) (0.335) MAYt−1 −24.861 −2.429 −4.294 −0.825 (10.235) (0.016) (5.205) (0.411) JUNt−1 0.893 0.087 −16.161 −2.749 (10.291) (0.931) (5.879) (0.007) AUGt−1 −41.433 −3.037 14.525 2.882 (13.641) (0.003) (5.040) (0.005) SEPt−1 −35.039 −2.903 4.624 0.938 (12.070) (0.004) (4.931) (0.350) OCTt−1 −26.470 −2.478 10.209 2.113 (10.683) (0.014) (4.830) (0.036) NOVt−1 −6.401 −0.637 −0.559 −0.101 (10.046) (0.525) (5.522) (0.920) DECt−1 −28.249 −2.314 5.743 1.047 (12.208) (0.022) (5.487) (0.297) Promotion by Helplinet−1 −15.039 −0.658 0.940 0.088 (22.859) (0.512) (10.651) (0.930) Website Improvementt−1 −34.259 −0.912 −7.885 −0.437 (37.561) (0.363) (18.029) (0.663) Intercept 222.829 4.989 −50.102 −2.314 (44.667) (<0.001) (21.648) (0.022) Log likelihood −522.706 −511.444 AIC 6.791 6.651 Note: Significant variables are in bold and are significant at the level of 0.05. a Share refers to the fraction of a certain topic that is talked about via telephone or chat. b No. of observations = 161. c The sign of the following coefficients are the ones in Eq. (3). It needs to change it into the opposite when interpreting its effect due to the negative adjustment coefficient. Table options The monthly results are interesting. We chose July as our base case month and we see that July has the lowest share for Violence, while February is the highest. The coefficients for June and November are not significant, suggesting that these are also months in which the lowest share of Violence is discussed. By contrast, the share of conversations about Relationships peaks in June.