اثرات افسردگی مادران بر نتایج کودک در طول سال اول تحصیلات رسمی

Depression among parents, particularly maternal depression, is an important factor in children's lives. Among mental health disorders, depression is one of the most common and can be a debilitating and chronic condition. Little research has examined the effect of maternal depression on children's school-related outcomes during the early years of formal schooling. Using a nationally representative sample of over 16,940 kindergarteners (ages 5–6); we examine the association between the timing and persistence of maternal depression, measured in kindergarten and third grade, and children's school behaviors, academic achievement, and school absences through fifth grade. Results indicate that persistent maternal depression predicts more negative behaviors, smaller achievement gains, and increased school absences. Implications for policy and practice are discussed.

Depression among parents, particularly maternal depression, is an important factor in children's lives. Among mental health disorders, depression is one of the most common and can be a debilitating and chronic condition. Depression is associated with many negative outcomes and difficulties throughout the life course (Miech and Shanahan, 2000 and Teitler and Reichman, 2008). Women are nearly two times more likely than men to experience depression (Kessler, 2003). Over 10 percent of women ages 18–39 suffer from depression at any given time, with lifetime prevalence of Major Depressive Disorder (MDD) for women estimated at over 20 percent (CDC, 2012 and Curry et al., 2014). Nearly 75% of those who have experienced an episode of depression will have a subsequent episode (Keller and Boland, 1998 and Kessler et al., 1997). Individuals with depression can have difficulty maintaining relationships and have increased negative interactions with others (Coyne, 1976 and Segrin and Dillard, 1992). Mothers with depression can have difficulty parenting, have more negative interactions with their children, and have been found to provide less cognitive stimulation compared to mothers without depression (Kiernan and Huerta, 2008 and Turney, 2012). Parental capacity to support children through elementary and middle school is influenced by many factors, including parents’ mental health. Parents play an important role in students’ engagement with and success in school. Students’ perceptions of support from parents and family cohesion have been linked to student competence (Connell et al., 1994 and Wentzel, 1998) and parental neglect or lack of support is linked to school failure (Glasgow, Dornbusch, Troyer, Steinberg, & Ritter 1997). School disengagement is an important predictor of subsequent academic failure (Crosnoe, 2002), and school attendance is one form of school disengagement. Disengagement from school can begin in the early elementary grades. When young children become disconnected from school early on, it may persist and have long-term consequences for academic motivation and achievement (Hamre & Pianta, 2001). Children's early success in school in terms of both achievement and behavior has been linked to subsequent success (Duncan et al., 2007, Durlak et al., 2011 and Raver, 2003). Given the links between children's early achievement, behavior, and attendance for later school-related outcomes, it is important to understand the association between maternal depression and these outcomes across elementary school. While a large body of research explores the association between maternal depression, often measured as depressive symptomology, and child development (Cummings and Davies, 1994, Goodman and Gotlib, 2002, Kiernan and Huerta, 2008 and Turney, 2012), little research has examined the association between maternal depression and school-aged children's classroom behaviors, academic achievement, and school attendance. It is important to note that, like much of the related research (Augustine and Crosnoe, 2010, Goodman and Gotlib, 2002, Kiernan and Huerta, 2008, Turney, 2011 and Turney, 2012), while we refer to maternal depression throughout this study, our measure is an indicator of depressive symptoms rather than an official diagnosis of depression. Among the few studies that have examined the relationship between maternal depression and children's developmental outcomes in elementary school (Ashman et al., 2008, Augustine and Crosnoe, 2010, Bodovski and Youn, 2010 and Essex et al., 2001), little attention has been given to the importance of the timing, persistence, or severity of maternal depression. The current study aims to address this gap in the literature. Using a nationally representative longitudinal sample of kindergarteners, we examine the relationship between the timing, persistence, and severity of maternal depression and children's school behaviors, academic achievement, and attendance in third and fifth grades.

As shown in Table 1, 13 percent of sample mothers reported depression, using the cutoff of 10, during only the kindergarten year and 17 percent reported depression only when their child was in third grade. Eleven percent of sample mothers had CES-D scores indicating depression at both kindergarten and third grade (persistent depression). Table 1 also displays descriptive statistics by the categories we created to indicate the timing and persistence of maternal depressive symptoms. The second column shows results for children whose mothers’ CES-D scores indicated depression in kindergarten but not third grade. The third column represents those children whose mothers were depressed only at third grade. The fourth column includes children whose mothers reported depression at both kindergarten and third grade, and the final column shows descriptive statistics for children whose mothers never reported depressive symptomology. Children whose mothers are persistently depressed are from more disadvantaged backgrounds than their peers whose mothers experience only episodic depression or whose mothers never experience depression. They have lower family income, mothers with less education, and increased likelihood of living with a single parent. In addition, these children were less likely to attend preschool and had fewer children's books in their homes. Table 2 shows means for outcomes of interest for sample students in third and fifth grades for the full sample as well as for the categories we created indicating the timing and persistence of maternal depression. Children whose mothers reported depressive symptoms at any point in time, on average, miss more days of school each year and are more likely to miss more than two weeks of school. Children with persistently depressed mothers miss more school than those in any of the other categories. A similar pattern is observed for our other outcomes of interest; approaches to learning, externalizing and internalizing behavior problems, and achievement. As shown in the table, means are statistically significant across these mutually exclusive groups in most cases. Below, we present regression adjusted mean differences between the groups and discuss the effect size differences for the different independent variables of interest. Table 2. Descriptive statistics for all outcomes for full sample and by timing and persistence of maternal depression (greater than or equal to 10 on CES-D). Full sample M (SD) (1) Depressed kindergarten only M (SD) (2) Depressed 3rd grade only M (SD) (3) Depressed at both timepoints M (SD) (4) Never depressed M (SD) (5) Difference between columns (2) vs. (3) (2) vs. (4) (2) vs. (5) (3) vs. (4) (3) vs. (5) (4) vs. (5) Teacher rated behaviors Spring 3rd learning 3.05 2.99 2.97 2.85 3.11 * * * * * (0.93) (1.22) (1.40) (1.16) (0.96) Spring 3rd externalizing behavior 1.69 1.74 1.73 1.86 1.65 * * * * * (0.82) (1.05) (1.15) (1.20) (0.82) Spring 3rd internalizing behavior 1.63 1.67 1.66 1.77 1.59 * * * * * (0.73) (0.89) (0.97) (1.00) (0.71) Spring 5th learning 3.07 2.99 3.03 2.87 3.12 * * * * * (0.95) (1.13) (1.34) (1.30) (0.97) Spring 5th externalizing behavior 1.64 1.69 1.66 1.78 1.60 * * * * * (0.82) (1.02) (1.15) (1.21) (0.81) Spring 5th internalizing behavior 1.63 1.66 1.69 1.74 1.60 * * * * * (0.77) (1.00) (1.20) (1.05) (0.78) Achievement test scores 3rd grade math 98.70 95.80 94.06 88.26 101.91 * * * * * (30.28) (35.22) (36.64) (33.93) (30.91) 3rd grade reading 126.61 122.56 120.91 113.20 130.71 * * * * * (34.35) (40.28) (44.30) (42.13) (34.72) 5th grade math 123.74 120.87 119.22 113.38 126.71 * * * * * (34.29) (41.82) (45.58) (43.88) (33.91) 5th grade reading 150.06 146.09 144.60 137.22 153.79 * * * * * (36.46) (44.78) (49.31) (47.95) (35.27) Total Absences (mean) 3rd grade 6.80 7.36 7.06 8.21 6.34 * * * * * (10.52) (11.37) (10.97) (12.57) (10.30) 5th grade 6.81 7.09 7.29 8.21 6.35 * * * * * (11.68) (13.31) (11.53) (11.75) (9.52) More than 10 absences (prop.) 3rd grade 0.23 0.26 0.26 0.31 0.21 * * * * * (0.61) (0.59) (0.61) (0.65) (0.59) 5th grade 0.24 0.26 0.27 0.31 0.21 * * * * * (0.55) (0.71) (0.73) (0.65) (0.53) Observations 19,990 2660 3310 2170 11,850 * Statistically significant difference (p < 0.05) for a t-test between the specified columns. Table options To investigate the association between maternal depression and school behavior, achievement, and absence outcomes in third and fifth grades, we estimate models of the form of Eq. (1) presented above. Results showing the relationship between maternal depression using the cutoff score of 10 or above on the CES-D are shown in first set of columns of Table 3, Table 4 and Table 5. To investigate the association between more severe maternal depression and school behavior, achievement, and absences, we estimate models of the form of Eq. (1) using a cutoff of 15 or above on the CES-D. The latter columns of Table 3, Table 4 and Table 5 present results from these models. The independent variables of interest are movement out of depression (depressed at kindergarten but not in third grade), into depression (depressed in third grade, not depressed at kindergarten), and persistent maternal depression (depressed at both time points). Table 3. Coefficients and standard errors from OLS regression models predicting approaches to learning, externalizing, and internalizing outcomes from third and fifth grade using indicators of timing and persistence of maternal depression. Score greater than or equal to 10 on CES-D Score greater than or equal to 15 on CES-D Third grade Fifth grade Third grade Fifth grade Learn β (SE) Extern β (SE) Intern β (SE) Learn β (SE) Extern β (SE) Intern β (SE) Learn β (SE) Extern β (SE) Intern β (SE) Learn β (SE) Extern β (SE) Intern β (SE) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Depressed kindergarten not 3rd grade −0.040 0.062* 0.082** −0.062* 0.057* 0.034 −0.083* 0.142** 0.140** −0.113** 0.103* 0.066 (0.028) (0.027) (0.025) (0.025) (0.025) (0.031) (0.037) (0.035) (0.042) (0.040) (0.043) (0.045) Depressed only at 3rd grade −0.053* 0.050* 0.055 −0.020 0.028 0.088** −0.064 0.072 0.074 −0.037 0.028 0.122** (0.024) (0.024) (0.029) (0.027) (0.026) (0.029) (0.036) (0.039) (0.043) (0.034) (0.034) (0.037) Depressed kindergarten and 3rd grade −0.112** 0.153** 0.174** −0.109** 0.111** 0.123** −0.132* 0.196** 0.223** −0.135* 0.146* 0.140* (0.029) (0.029) (0.037) (0.030) (0.029) (0.034) (0.063) (0.065) (0.072) (0.062) (0.065) (0.067) Controls X X X X X X X X X X X X Observations 19,990 19,990 19,990 19,990 19,990 19,990 19,990 19,990 19,990 19,990 19,990 19,990 t-Tests – ab ab b b – – – – – – – R-squared 0.317 0.301 0.125 0.287 0.256 0.109 0.317 0.301 0.125 0.287 0.255 0.108 Note. Standard errors are clustered at the classroom level. * p < 0.05 ** p < 0.01. a statistically significant p < 0.05 differences between coefficient and the coefficient for depressed kindergarten not 3rd grade for a Welch's t-test. b statistically significant p < 0.05 differences between coefficient and the coefficient for depressed only at 3rd grade for a Welch's t-test. Tests revealed no statistically significant differences between corresponding coefficients using > = 10 and > = 15 cutoffs for a Welch's t-test. Tests revealed no statistically significant differences between depressed kindergarten not 3rd grade and depressed only at 3rd grade for a Welch's t-test. Table 1 contains the full list of control variables. Controls include fall K math, reading, and general knowledge achievement scores, teacher rated measures of learning, externalizing, and internalizing behaviors, child gender, child race, child age at start of kindergarten, parent rated health scale, number of biological siblings, whether the household is led by a single biological parent, whether English is not the primary home language, parent reads to child (days per week), number of children's books in home, maternal education (years), whether mother worked before kindergarten, income, whether the child attended full day kindergarten, center-based pre-kindergarten, or Head Start, region of the country, and urbanicity. Table options Table 4. Coefficients and standard errors from OLS regression models predicting reading and mathematics achievement from third and fifth grade using indicators of timing and persistence of maternal depression. Score greater than or equal to 10 on CES-D Score greater than or equal to 15 on CES-D Third grade Fifth grade Third grade Fifth grade Math β (SE) Reading β (SE) Math β (SE) Reading β (SE) Math β (SE) Reading β (SE) Math β (SE) Reading β (SE) (1) (2) (3) (4) (5) (6) (7) (8) Depressed kindergarten not 3rd grade −0.005 −0.027 −0.003 −0.042* −0.013 −0.031 0.004 −0.059 (0.021) (0.019) (0.022) (0.018) (0.026) (0.030) (0.030) (0.031) Depressed only at 3rd grade −0.031 −0.043* −0.030 −0.051* −0.031 −0.050 −0.038 −0.069* (0.020) (0.019) (0.024) (0.022) (0.029) (0.031) (0.032) (0.030) Depressed kindergarten and 3rd grade −0.063** −0.096** −0.057* −0.113** −0.091* −0.103* −0.079 −0.138** (0.022) (0.026) (0.027) (0.024) (0.040) (0.047) (0.047) (0.048) Controls X X X X X X X X Observations 19,990 19,990 19,990 19,990 19,990 19,990 19,990 19,990 t-tests – a – a – – – – R-squared 0.560 0.558 0.523 0.536 0.56 0.558 0.523 0.525 Note. Standard errors are clustered at the classroom level. * p < 0.05 ** p < 0.01. a statistically significant p < 0.05 differences between coefficient and the coefficient for depressed kindergarten not 3rd grade for a Welch's t-test. Tests revealed no statistically significant differences between depressed only at 3rd grade and depressed kindergarten and 3rd grade for a Welch's t-test. Tests revealed no statistically significant differences between corresponding coefficients using > = 10 and > = 15 cutoffs for a Welch's t-test. Tests revealed no statistically significant differences between depressed kindergarten not 3rd grade and depressed only at 3rd grade for a Welch's t-test. Table 1 contains the full list of control variables. Controls include fall K math, reading, and general knowledge achievement scores, teacher rated measures of learning, externalizing, and internalizing behaviors, child gender, child race, child age at start of kindergarten, parent rated health scale, number of biological siblings, whether the household is led by a single biological parent, whether English is not the primary home language, parent reads to child (days per week), number of children's books in home, maternal education (years), whether mother worked before kindergarten, income, whether the child attended full day kindergarten, center-based pre-kindergarten, or Head Start, region of the country, and urbanicity. Table options Table 5. Coefficients and standard errors from OLS regression models predicting student absences and logistic regressions predicting whether students missed more than two weeks of school in third and fifth grades using indicators of timing and persistence of maternal depression. Maternal depression Score greater than 10 on CES-D Score greater than 15 on CES-D Third grade Fifth grade Third grade Fifth grade Total absences β (SE) Absent > 10 days β (SE) Odds ratio Total absences β (SE) Absent > 10 days β (SE) Odds ratio Total absences β (SE) Absent > 10 days β (SE) Odds ratio Total absences β (SE) Absent > 10 days β (SE) Odds ratio (1) (2) (3) (4) (5) (6) (7) (8) Depressed kindergarten not 3rd grade 0.600* 0.185** 1.203 0.271 0.164** 1.178 0.868* 0.099 1.104 0.181 0.163* 1.177 (0.253) (0.054) (0.262) (0.058) (0.387) (0.077) (0.355) (0.080) Depressed only at 3rd grade 0.319 0.364** 1.439 0.511* 0.379** 1.461 0.426 0.092 1.096 0.705* 0.121 1.129 (0.203) (0.056) (0.212) (0.055) (0.279) (0.077) (0.307) (0.084) Depressed kindergarten and 3rd grade 1.131** 0.441** 1.554 1.052** 0.434** 1.544 1.248* 0.252 1.286 1.672** 0.325* 1.384 (0.298) (0.062) (0.281) (0.068) (0.607) (0.149) (0.641) (0.149) Controls X X X X X X X X Observations 19,990 19,990 19,990 19,990 19,990 19,990 19,990 19,990 t-tests b acd a acd – d a d R-squared 0.050 0.039 0.058 0.041 0.048 0.035 0.058 0.038 Note. Standard errors are clustered at the classroom level. * p < 0.05 ** p < 0.01. a statistically significant p < 0.05 differences between coefficient and the coefficient for depressed kindergarten not 3rd grade for a Welch's t-test. b indicates statistically significant p < 0.05 differences between coefficient and the coefficient for depressed only at 3rd grade for a Welch's t-test. c statistically significant p < 0.05 differences between coefficient and the coefficient for depressed kindergarten not 3rd grade for a Welch's t-test. d statistically significant p < 0.5 differences between the 10 and 15 cutoff estimates for a Welch's t-test. Tests revealed no statistically significant differences between depressed kindergarten not 3rd grade and depressed only at 3rd grade for a Welch's t-test. Table 1 contains the full list of control variables. Controls include fall K math, reading, and general knowledge achievement scores, teacher rated measures of learning, externalizing, and internalizing behaviors, child gender, child race, child age at start of kindergarten, parent rated health scale, number of biological siblings, whether the household is led by a single biological parent, whether English is not the primary home language, parent reads to child (days per week), number of children's books in home, maternal education (years), whether mother worked before kindergarten, income, whether the child attended full day kindergarten, center-based pre-kindergarten, or Head Start, region of the country, and urbanicity. Table options Classroom behaviors Table 3 presents results using maternal depression to predict children's approaches to learning and externalizing and internalizing behaviors in the spring of third and fifth grades using the two different measures of severity of maternal depression (cutoffs of 10 and 15, respectively, on the CES-D). Results for all models shown in Table 3 include the full set of control variables listed in Table 1 and outcomes are standardized to a mean of 0 and a standard deviation of 1 so that coefficients can be interpreted in standard deviation units, effect sizes, or βs. Positive values on the measure of approaches to learning indicate more positive behavior, while positive values on the measures of externalizing and internalizing behaviors indicate more problematic behaviors. Results indicate that, for most outcomes, children whose mothers were depressed at kindergarten but not in third grade had more externalizing problems in both third and fifth grades than those children whose mothers were never depressed. Interestingly, among children whose mothers were depressed only in third grade, we find fewer statistically significant associations, for third grade (concurrent) externalizing and internalizing problems and fifth grade internalizing behavior problems. Persistent depression—having a mother who was depressed in both kindergarten and third grade—relative to those children whose mothers never report depression, is statistically significantly associated with internalizing and externalizing behavioral problems and approaches to learning in third and fifth grades. Coefficients range from approximately (in absolute value) 0.11 standard deviations for approaches to learning measured at fifth grade to 0.17 standard deviations for internalizing problem behaviors measured at the end of third grade. Results of post hoc t-tests, reported in detail in the table, indicate that in many cases, negative effects for children of mothers who are persistently depressed are statistically significantly larger than effects for children whose mothers were depressed at only a single point in time (kindergarten or third grade). Columns 7–12 show results for specifications using a cutoff of 15 or above, as opposed to 10 or above, on the CES-D. While coefficients for the cutoff of 15 or above are consistently larger than those estimated for the cutoff of ten or above, sometimes more than twice as large, t-tests revealed no statistically significant differences across the cutoffs. This is likely due, in part, to the lack of precision in the models using the cutoff of 15 or above, which can be seen in the much larger standard errors that are reported.