Serie de Documentos de Discusión
Serie de Documentos de Discusión
Serie de Documentos de Discusión
Jason M. Fletcher
Stefanie Schurer
enero de 2017
Serie de Documentos de Discusión
Jason M. Fletcher
Universidad de Wisconsin-Madison e IZA
Stefanie Schurer
Universidad de Sydney e IZA
enero de 2017
Todas las opiniones expresadas en este documento son las del autor(es) y no las de IZA. La investigación
publicada en esta serie puede incluir puntos de vista sobre políticas, pero IZA no toma posiciones políticas
institucionales. La red de investigación de IZA está comprometida con los Principios rectores de integridad
de la investigación de IZA.
El Instituto IZA de Economía Laboral es un instituto de investigación económica independiente que realiza
investigaciones en economía laboral y ofrece asesoramiento sobre políticas basado en evidencia sobre
cuestiones del mercado laboral. Con el apoyo de la Fundación Deutsche Post, IZA dirige la red de
economistas más grande del mundo, cuya investigación tiene como objetivo brindar respuestas a los
desafíos del mercado laboral global de nuestro tiempo. Nuestro objetivo clave es tender puentes entre la
investigación académica, los responsables políticos y la sociedad.
Los documentos de debate de IZA a menudo representan un trabajo preliminar y se distribuyen para
fomentar el debate. La cita de tal documento debe dar cuenta de su carácter provisional. Una versión
revisada puede estar disponible directamente del autor.
Resumen
Orígenes de Personalidad en la edad adulta:
el papel de las experiencias infantiles adversas1
Probamos si las experiencias infantiles adversas (exposición al maltrato de los padres y
su efecto indirecto en la salud) están asociadas con los rasgos de personalidad de los
30 años. Utilizamos abundantes datos longitudinales de una gran cohorte
representativa de jóvenes estadounidenses y explotamos las diferencias entre
hermanos para controlar las influencias de confusión de los factores ambientales y
genéticos compartidos. Encontramos que las experiencias de maltrato están
significativa y fuertemente asociadas con el neuroticismo, la escrupulosidad y la
apertura a la experiencia, pero no con la amabilidad y la extraversión. Los altos niveles
de neuroticismo están relacionados con el abuso y la negligencia sexuales; Los bajos
niveles de escrupulosidad y apertura a la experiencia están relacionados con la
negligencia de los padres. Las asociaciones estimadas se reducen significativamente
en magnitud cuando se controla la salud física o mental, lo que sugiere que la salud de
los adolescentes podría ser una vía importante a través de la cual el maltrato afecta la
personalidad en la edad adulta. Las experiencias de maltrato, en combinación con sus
efectos en la salud, explican una fracción significativa de la relación entre la conciencia
de la edad adulta y los ingresos o el capital humano. Nuestros hallazgos brindan una
posible explicación de por qué los rasgos de personalidad son importantes predictores
de los resultados del mercado laboral en la edad adulta.
Autor correspondiente:
Stefanie Schurer
School of Economics
The University of Sydney
Sydney
Australia
E- mail: [email protected]
1
Los autores agradecen el apoyo financiero de una subvención del Programa de Descubrimiento Temprano
de Carreras del Australian Research Council (DE140100463) y del Australian Research Council Centre of
Excellence for Children and Families over the Life Course (número de proyecto CE140100027 ). Esta
investigación utiliza datos de Add Health, un proyecto de programa dirigido por Kathleen Mullan Harris y
diseñado por J. Richard Udry, Peter S. Bearman y Kathleen Mullan Harris en la Universidad de Carolina del
Norte en Chapel Hill, y financiado por la subvención P01-HD31921 del Instituto Nacional de Salud Infantil y
Desarrollo Humano Eunice Kennedy Shriver, con financiamiento cooperativo de otras 23 agencias y
fundaciones federales. Se debe un reconocimiento especial a Ronald R. Rindfuss y Barbara Entwisle por su
ayuda en el diseño original. La información sobre cómo obtener los archivos de datos de Add Health está
disponible en el sitio web de Add Health (http://www.cpc.unc.edu/addhealth). No se recibió apoyo directo de
la subvención P01-HD31921 para este análisis.
1. Introducción
de calificación postsecundaria. Sin embargo, en los últimos años, los economistas han
(NCS), habilidades blandas o rasgos de carácter, son una forma alternativa importante de
capital humano (Almlund et al., 2011). Aunque en la literatura se han considerado numerosos
indicadores de los rasgos de personalidad de la edad adulta, los rasgos de personalidad de los
Cinco Grandes son uno de los más utilizados (p. ej., Mueller & Plug, 2006; Heineck & Anger,
2010; Heckman y Kautz, 2012; Fletcher, 2013; Lundberg, 2013; Cameron et al., 2014;
Gensowski, 2014).
En este estudio exploramos los factores que dan forma a los rasgos de personalidad de
los Cinco Grandes entre la niñez y la edad adulta joven para una población estadounidense
grande y representativa. Específicamente, buscamos arrojar luz sobre los efectos a largo plazo
estudiar la exposición de los niños al maltrato, nuestro estudio se distingue, pero contribuye a,
sobre la economía del desarrollo humano que busca comprender mejor la función de
en educación), el sector de la educación y los ingresos. Nuestros hallazgos podrían ser útiles
los responsables políticos que están preocupados por las ventanas de oportunidad para dar
.
el 150% de la variación en los rasgos de personalidad puede atribuirse a experiencias
personales (Turkheimer, 2000; Turkheimer et al., 2003; Krueger et al., 2008; Borkenau et al,
2011). Los psicólogos han sugerido que las variaciones en la personalidad de la edad adulta
(Caspi & Roberts, 2001; Caspi et al., 2005). En este estudio, por lo tanto, centramos nuestra
Los niños que han experimentado abuso y negligencia tienen un mayor riesgo de una
externalización (Petersen & Feit, 2014 para una revisión). Como adultos, estos niños tienen
más probabilidades de sufrir problemas de salud física, tener conductas de riesgo (Felitti et al.,
1998, Widom et al., 2012) y desarrollar trastornos del estado de ánimo (Kaufman et al., 2004,
Fletcher, 2009, Widom et al., 2012) y psicopatologías (Putnam, 2006; Spila et al., 2008; Tyrka
et al., 2009; Widom et al., 2009; Clark et al., 2010); también es más probable que se
involucren en actividades delictivas (Currie & Tekin, 2012), lo que es consistente con la
observación de que los niños con experiencias de maltrato a menudo se ven afectados en su
desarrollo de comportamiento prosocial y ético (Koenig et al., 2004). . Es por estas razones
(Roberts et al., 2006, p. 1) que definen los rasgos de personalidad. Un estudio reciente de
Hengartner et al. (2015) enfatizaron la importancia del vínculo, pero concluyeron que es un
Para investigar el impacto a largo plazo del maltrato en la personalidad, utilizamos datos
2
entre las experiencias infantiles adversas y las respuestas a un cuestionario de personalidad en
la edad adulta que se recopiló años después de la exposición (antes de los 12 años) y el
que contiene información sobre el maltrato y la personalidad de los hermanos, que explotamos
en nuestro marco empírico para controlar algunos de los factores de confusión que pueden
sesgar los efectos del tratamiento de interés (Bound & Solon, 1999, Conley et al. , 2007,
Moffit et al., 2011). Exploramos más a fondo si las trayectorias de salud de la adolescencia,
que se informaron entre los 15 y los 22 años, median el efecto del maltrato infantil en la
personalidad. Finalmente, evaluamos hasta qué punto se reduce la relación estimada entre la
personalidad de la edad adulta y los ingresos de la edad adulta (p. ej., Fletcher, 2013) y el
nivel educativo (p. ej., Lundberg, 2013) cuando se controlan las experiencias infantiles
adversas.
2. Revisión de la literatura
Los psicólogos han desarrollado una serie de diferentes inventarios de personalidad, pero el
consistente para describir las diferencias humanas por parte de los psicólogos (McCrae y
Costa, 2008).2 Los psicólogos de la personalidad han demostrado fuertes vínculos entre los
desempeño laboral (p. ej., Judge et al., 1999), el rendimiento académico (Poropat, 2009) o los
estilos de vida saludables (p. ej., Hampson et al. ., 2006; Roberts et al., 2007). Diversos
estudios han establecido que bajos niveles de neuroticismo y altos niveles de escrupulosidad
2
El Diccionario de la Asociación Estadounidense de Psicología (2007) los describe de la siguiente manera: 1. Apertura
a la experiencia (intelecto): la tendencia a estar abierto a nuevas experiencias estéticas, culturales o intelectuales. 2.
Escrupulosidad—La tendencia a ser organizado, responsable y trabajador. 3. Extraversión—Una orientación de los
propios intereses y energías hacia el mundo exterior de las personas y las cosas en lugar del mundo interior de la
experiencia subjetiva; caracterizado por el afecto positivo y la sociabilidad. 4. Amabilidad: la tendencia a actuar de
manera cooperativa y desinteresada. 5. Neuroticismo (vs. Estabilidad emocional): un nivel crónico de inestabilidad
emocional y propensión a la angustia psicológica.
3
2006), y la esperanza de vida está asociada con la conciencia de los jóvenes (Kern &
emocional y la apertura a la experiencia están fuertemente asociadas con los ingresos del
mercado laboral (p. ej., Muller & Plug, 2006, Heineck & Anger, 2010, Fletcher, 2013),
incluso para personas con mucho talento (Gensowski, 2014). Tanto los altos niveles de
obtener un título universitario (Lundberg, 2013, Schurer et al., 2015), mientras que la
la disposición genética están menos investigados. Trabajos anteriores sostenían que los
individuos nacen con un temperamento fijo y que los cambios entre la niñez y la edad adulta
ocurren de manera determinista, un proceso que a menudo se denomina maduración (para una
descripción general, consulte McCrae y Costa, 2000). La dependencia del camino entre el
Caspi & Silva 1995, Caspi et al., 2003, Deal et al., 2005, Asendorpf et al., 2008, McAdams &
Olson, 2010, Moffitt et al. al., 2011). El trabajo anterior sobre el modelo de cinco factores
también asumió que los rasgos de personalidad se estabilizan en la edad adulta joven, pero la
evidencia más reciente ha demostrado que los rasgos de personalidad en la edad adulta pueden
incluso cambiar más allá de los 50 años (Roberts et al., 2008; Fraley & Roberts, 2005, Roberts
& DelVecchio, 2000, Roberts.et al., 2006; Roberts & Mroczek, 2008). Algunos estudios se
centran en el impacto de los roles sociales (Roberts et al., 2005), los acontecimientos de la
vida experimentados en la edad adulta (Specht et al., 2011, Cobb-Clark y Schurer, 2012,
Luhmann et al., 2014) o la adolescencia (Elkins et al., 2017), educación secundaria (Dahmann
y Anger, 2014) o educación terciaria (Lüdtke et al., 2011, Schurer et al., 2015).
4
se ha reconocido el papel que juegan los padres en el procesoUna literatura emergente en la
formación de capital humano (ver Cobb-Clark et al., 2016 para una discusión teórica). La
mayor parte del trabajo se centra en el desarrollo de las habilidades cognitivas y no cognitivas
de los niños (por ejemplo, Cunha y Heckman, 2008; Cunha et al., 2010; Del Bono et al., 2014;
Attanasio et al., 2015). Sin embargo, se ha realizado poca investigación empírica sobre el
efecto de las conductas parentales negativas, cuando los padres no ayudan a sus hijos a regular
personalidad. Esta falla a menudo se resume como maltrato y comprende abuso o negligencia
sexual, psicológica o física. El abuso se refiere al daño activo a través de un comportamiento
Hay muchas vías a través de las cuales las experiencias de maltrato pueden impactar
en la personalidad de la edad adulta. Ante todo, el maltrato tiene un efecto directo sobre el
desarrollo del cerebro. Varios estudios han demostrado las diferencias cerebrales biológicas
entre niños maltratados y no maltratados (ver Petersen & Feit, 2014 para referencias). Estos
incluyen diferencias en las áreas del cerebro que están involucradas en procesos cognitivos de
orden superior: ejecutivo función – que ayuda en el seguimiento y control de las emociones y
ejemplo, Kern et al., 2009). Esto sugeriría que los niños maltratados mostrarían niveles más
bajos de escrupulosidad y niveles más altos de neuroticismo temprano en la vida, facetas del
temperamento de un niño que pueden tener el potencial de madurar en rasgos fijos de la edad
adulta.
5
Las experiencias de maltrato infantil podrían vincularse con los rasgos de personalidad
aparición de un trastorno de personalidad (ver Galaif et al., 2001, Spila et al., 2008, Tyrka et
al., 2009). Si bien los rasgos de personalidad son distintos de los trastornos de personalidad,
ahora hay un cuerpo considerable de investigación que entiende los trastornos de personalidad
Samuel y Widenger (2008) y Widenger et al. (2005), que mapeó con éxito los trastornos del
Eje II en variantes desadaptativas de las 30 facetas del modelo de cinco factores, mostró que
también tienen más probabilidades de sufrir dificultades de atención y aprendizaje (p. ej.,
Koenen et al., 2003, Krueger et al., 2008). De acuerdo con las sugerencias hechas por Caspi &
Roberts (2001) y Caspi et al. (2005) de que las diferencias de personalidad en la edad adulta
pueden estar precedidas por problemas de salud en la niñez, esperaríamos que el maltrato
correlacionadas porque tienen las mismas causas subyacentes. Por ejemplo, las experiencias
sostenidas de pobreza
6
y los factores ambientales estresantes pueden hacer que ambos padres, de forma
pueden dar forma de forma independiente a la capacidad cognitiva, la salud física o mental,
Hasta donde sabemos, somos los primeros en explorar en detalle la relación entre los
rasgos de personalidad de la edad adulta y las experiencias infantiles adversas relacionadas
con la crianza. La única excepción es Hengartner et al. (2015), quienes muestran fuertes
asociaciones no ajustadas entre los rasgos de personalidad de la edad adulta y las experiencias
de maltrato autoinformadas (abuso y negligencia emocional, física y sexual) para 1170 sujetos
de una población basada en encuesta comunitaria. Los autores enfatizan que la relación entre
los rasgos de personalidad y el maltrato es un “campo muy poco estudiado” (Hengartner et al.,
2015, p. 1).
análisis con datos de una cohorte representativa a nivel nacional; (2) Somos capaces de
controlar un gran conjunto de factores de confusión que incluyen la pobreza, los compañeros
y la disposición genética que se comparten entre hermanos; (3) Somos capaces de comentar
sobre las posibles vías a través de las cuales las experiencias adversas de la infancia impactan
otros factores que influyen en el maltrato, la personalidad y los mediadores; y (4) Aunque
también fue autoinformado, nuestras medidas de maltrato se recopilaron 12 años antes de que
3
Es posible que los estilos de crianza en general sean consecuencia del temperamento del niño, lo que significa que
los padres ajustan sus estilos de crianza a las necesidades y el temperamento del niño (p. ej., Deal et al., 2005).
Debido a esta simultaneidad de crianza, temperamento y salud, modelar el efecto de la conducta de crianza es un
desafío empírico. Algunos estudios explotaron el orden de nacimiento, una variación exógena en el trato diferencial
de los padres, para probar si el comportamiento de los padres afecta la personalidad de la edad adulta, pero
encontraron poca evidencia a favor (p. ej., Marini & Kurtz, 2011; Sulloway, 1996). La investigación sobre la política del
hijo único de China (OCP, por sus siglas en inglés), un experimento institucional natural que condujo a que todos los
cuidadores concentraran la atención en un niño, descubrió que los niños nacidos justo después de la introducción de
la OCP en 1979 tendían a ser menos conscientes, más neuróticos y menos optimistas que los niños nacidos justo
antes (Cameron et al. 2013).
7
3. Datos: Estudio Longitudinal Nacional de Salud Adolescente
Los datos de este estudio provienen de la versión confidencial del Estudio Longitudinal
Nacional de Salud Adolescente (Add Health). Add Health, una de las encuestas de
edad adulta joven (Udry, 2003). Comenzando con un cuestionario en la escuela administrado
a una muestra representativa a nivel nacional de estudiantes en los grados 7 a 12 en 1994-95
(Etapa 1: edad promedio de 15 años), el estudio siguió con una serie de entrevistas en el hogar
de los estudiantes aproximadamente un año (Etapa 2 en 1996: edad media 16), seis años
(Etapa 3 en 2001/2002: edad media 22) y 12 años después (Etapa 4 en 2007/2008: edad media
29).
estudios y administradores escolares. Por diseño, la encuesta Add Health incluyó una muestra
estratificada por región, urbanidad, tipo de escuela, mezcla étnica y tamaño. Las bases de
datos preexistentes (por ejemplo, los datos del censo) se han vinculado con las personas de la
muestra y brindan información sobre los vecindarios y las comunidades. De los 20 745
estudiantes encuestados durante el Relevamiento 1, casi 15 000 (75 %) han sido seguidos
del Relevamiento III disponible. Hay 5470 hermanos en la muestra completa del
pertenecen a una familia donde también se siguió a otro hermano en los Relevamientos 3 y 4.
De estos 3098 encuestados elegibles, 2453 tienen información sobre maltrato y medidas de
datos), dejando nuestra muestra de análisis de 2319. La mayoría de las familias de la muestra
8
Evaluamos si la muestra de hermanos es sistemáticamente diferente en características
muy pocas diferencias entre muestras, especialmente con respecto a los niveles de maltrato y
nuestra muestra final de análisis de hermanos con las características familiares de Wave 1, no
Encontramos una asociación muy pequeña con el ingreso familiar, lo que sugiere que una
ítems del Modelo internacional de cinco factores de grupo de ítems de personalidad (IPIP-
FFM) de 50 ítems conocido como el Mini IPIP (Donnellan et al. 2006). Baldasaro et al.
(2013) sugieren que el Mini-IPIP tiene una estructura de cinco factores que representan
y apertura a la experiencia. La mayoría de las escalas tienen una confiabilidad aceptable, todas
las escalas tienen invariancia métrica parcial o total y las escalas exhiben suficiente validez de
4
En nuestra muestra de estimación completa, el alfa de Cronbach para cada dimensión es: escrupulosidad 0,64,
apertura a la experiencia 0,61, extraversión 0,70, amabilidad 0,68 y neuroticismo 0,85.
9
5
dominio y estandarizarlo a media 0 y desviación estándar 1 (ver Almlund et al. 2011, p. 32).
3.2. Indicadores
En el Relevamiento 3, a los encuestados se les hicieron cuatro preguntas sobre cómo los
trataban sus padres (o cuidadores adultos) antes de que estuvieran en sexto grado (12 años).
Específicamente, se les preguntó si y con qué frecuencia:
(1) Los padres (u otros cuidadores adultos) no habían atendido sus necesidades
(2) Los padres (u otros cuidadores adultos) los abofetearon, golpearon o patearon.
(3) Los padres (u otros cuidadores adultos) los habían tocado de manera sexual, los
habían obligado a tocarlo a él o ella de manera sexual o los habían obligado a tener
relaciones sexuales. (4) Los padres (u otros cuidadores adultos) los dejaron solos en
Usamos dos medidas para usar la información de maltrato. Primero, derivamos del análisis
factorial sobre las cuatro categorías de respuesta una medida continua de maltrato. Esta
medida captura la intensidad del maltrato, sin especificar las causas subyacentes del maltrato.
En segundo lugar, para identificar por separado los efectos y la intensidad de las cuatro
variantes de maltrato, seguimos a Currie & Tekin (2012) para construir para cada evento un
indicador binario que toma el valor 1 si el encuestado informa que experimentó el respectivo
maltrato más de 10 veces, y 0 en caso contrario. Una razón para considerar solo la mayor
los eventos únicos, tendrá un impacto a largo plazo en el comportamiento. Currie y Tekin
(2012)
5
Las mismas medidas se han utilizado en Fletcher (2013) y Lundberg (2013).
10
y Fletcher (2009) brindan mayores detalles de que la información proporcionada en el informe
experimentado algún tipo de abuso durante su infancia. Cualquier abuso sexual fue reportado
por el 5% de la muestra (535), pero de manera regular (> 10 veces) alrededor del 1% (106) de
la muestra reportó abuso sexual. Mientras que casi un tercio de la muestra reporta haber sido
azotado, golpeado o pateado alguna vez, más del 6% (642) experimentó abuso físico de
manera más regular (> 10 veces). El descuido regular de las necesidades básicas o haber sido
abuso sexual frecuente: los miembros femeninos de la cohorte tienen cuatro veces más
sobre su personalidad que se pueden mapear en tres de los Cinco Factores: neuroticismo,
Beaujean (2011) compararon estas preguntas con ítems de la versión IPIP (Goldberg et al.,
2006) del NEO-PI-R (Costa & McCrae, 1992). Sometieron todos los elementos disponibles a
un nivel de elemento
6
Las medidas de maltrato autoinformadas son propensas a errores. Currie y Tekin (2012) analizan los peligros
potenciales de estas medidas, pero se refieren a documentos metodológicos que han demostrado que, "si se
recopilan correctamente, estos datos son válidos" (p. 515). Se pidió a los participantes del estudio AddHealth que
escucharan preguntas pregrabadas sobre temas delicados a través de auriculares y que ingresaran sus respuestas
directamente en las computadoras portátiles. Este proceso aseguró la confidencialidad y minimizó la posibilidad de
que el entrevistador u otra influencia de terceros. Con el fin de obtener respuestas precisas sobre el momento de los
eventos, se solicitó a los miembros del estudio un calendario que proporcionaba las fechas de muchos eventos
importantes. Si bien el sesgo de recuerdo es una consideración importante para estas medidas, el sesgo podría ser
pequeño porque los encuestados eran adultos jóvenes cuando se les preguntó sobre el maltrato infantil. Esto tiene la
ventaja de que los adultos jóvenes son lo suficientemente maduros para comprender e informar sobre tales eventos
(ver Perkonigg et al., 2000). Otra ventaja es que la ventana de tiempo durante la cual los encuestados recuerdan
eventos pasados es relativamente corta (10 años en promedio).
11
análisis factorial para determinar qué ítems conservar, así como la dimensionalidad de los
dominios en los que se midieron los ítems. Llegaron a la conclusión de que 13 de los 21 ítems
originales se pueden usar de manera confiable para generar medidas de neuroticismo infantil
Construimos medidas estándar de problemas de salud física y mental que se informaron entre
que reflejen los problemas de salud que probablemente se hayan presentado después de la
experiencia del maltrato. La Tabla A5 (Apéndice Técnico) describe estas variables en detalle.
La Tabla 1 presenta una breve descripción de todas las variables utilizadas para el
7
La estructura factorial de los ítems indica que cada uno mide solo un constructo y las puntuaciones de los tres
factores tienen estimaciones de consistencia interna que oscilan entre 0,76 y 0,86, que son similares a los
coeficientes de confiabilidad para los instrumentos de personalidad NEO en la edad adulta medidos con 3 a 10 veces
más ítems (Young & Beaujean, 2011, Tabla 5).
12
4. Marco empírico
Para probar si la experiencia de maltrato está asociada con la personalidad de la edad adulta,
ecuación. (1) el rasgo de personalidad k (��������) es una función lineal del maltrato
��� ��� captura todos los choques no observables que afectan el rasgo de personalidad k,
��,�� y ���� son parámetros a estimar. El vector ������ contiene una medida
resumen continua de maltrato o cuatro variables binarias que indican si el individuo
experimentó maltrato regularmente antes del grado 6 (abuso sexual, ser golpeado, dejado solo,
descuidado). The vector ����′ controls for pre treatment characteristics including age,
being female, family characteristics, and birth weight.
The above outlined model considers only variation between families and does not
control for unobserved family factors that may potentially confound the treatment effects of
interest. To control for some of these confounding factors, we exploit differences between
siblings (S-FE model). In Eq. (2) each individual has now subscript f, which represents the
family, and subscript i which represents sibling i within family f. On the right-hand side of Eq.
(2) we include only variables that vary between siblings (��������, ������).
The error term is now broken down into two components: ������ is a family fixed effect
and �������� the error specific to each sibling i in family f. The family fixed effect
could represent, for instance, level of familial conflict, family-specific behavioral styles, or a
genetic proneness to disease. To eliminate this family-fixed effect, we difference across the
13
����1���� − ����2���� = �����1�� − ����2���
′���
�
+
���
1�� −
��2�
��
′
�� +
(���
��
� −
���
��
� )
+
(��1
��
��
��2�
��
). (3)
The S-FE approach improves upon the OLS model because it controls for difficult-to measure
shared background components. However, as with most empirical models used to analyze
observational data, it has its limitations. On the one hand, the S-FE approach exploits only
variation within families, and therefore is an inefficient estimator (eg Conley et al., 2007, p.
��
1095). More important is that the estimated coefficients may still be biased if the −
��2��
differences in both maltreatment and personality. This approach would fail to identify a causal
effect of maltreatment on personality, if eg both siblings were maltreated, but only one sibling
will report the abuse due to being more willing to share maltreatment experiences (eg being
more extraverted or open to experiences) or due to inflating negative experiences (eg being
more neurotic).As we cannot rule out such possibility with certainty, the S-FE
We have sufficient variation in our data between siblings on most variables, which is
the main requirement for this approach to yield efficient estimates (Bound and Solon, 1999).
In Table 2 we report the number of siblings who differ in outcomes and treatment. Between
84% (Agreeableness) and 87% (Extraversion) of the sibling-pairs differ in their personality
scores. The numbers of sibling pairs who report differences in maltreatment are in ascending
order: 41 for sexual abuse (1.8%), 160 for neglect (7.0%), 237 for slapped or beaten (10.2%),
14
4.2. Mediation analysis
To better understand the potential pathways via which childhood maltreatment affects
separately that capture: (1) adolescent personality8; (2) adolescent cognitive skills; (3)
The mediation analysis is conducted on the S-FE model, as described in Eq. (4). Here
���1�� − ��2��� describes the difference in the block of control variables capturing
−
�
�
�
′
�
�
+
(
�
�
−
�
�
)
′
�
�
+
(
�
�
−
�
�
�
�
)
+
(
�
�
�
�
�
��
). (4)
− ��2��
The parameters in vector �� capture the association of the difference in these control
variables with adulthood personality. If adding the block of variables reduces the estimated
association between maltreatment and adulthood personality (���� ), then this would be
evidence in favor of the hypothesis that maltreatment affects adulthood personality via this
particular channel (McKinnon et al. 2007, p. 597). Three assumptions must hold to interpret
the results obtained from mediation analysis using our siblings-fixed effects model as causal
(see Judd and Kenny, 2010; McKinnon et al., 2007 for a review): (1) there is no reverse
causality between adult personality and the mediator (eg adolescent mental health) or the
mediator and
8
Instead of investigating the role of adolescent personality as possible mediator via which maltreatment affects
adulthood personality, we could have included it as standard control variable in the benchmark model. This is a
commonly used strategy in the literature on non-cognitive skill formation referred to as the value added model. Such
model assumes that adolescent personality is a valid proxy for previous inputs, for instance parenting behavior
including maltreatment, and educational opportunities (see Cunha and Heckman, 2008 for an overview of these
models).
15
maltreatment experiences; (3) There are no unobserved factors that cause both mental health
First, in our empirical setting we are likely to exclude the case of reverse causality
because of temporal precedence. Maltreatment experiences refer to a period before grade 6,
the potential mediators are measured after grade 6 but before adulthood, and personality
outcomes are measured strictly after maltreatment experiences and potential mediators.
Second, it is likely that measurement error occurs both in the report of maltreatment
experiences and the potential mediators, although of unknown magnitudes; hence we cannot
exclude the possibility that our mediation analysis yields bias estimates. Finally, although the
siblings-fixed effects model controls for all family-specific background variables that could
affect both maltreatment and potential mediators, there is still the possibility that unobservable
factors that are not shared between siblings impact independently on maltreatment
experiences and the mediators. In this case it would be inappropriate to interpret reduced
Because we cannot assume with certainty that all three modelling assumptions are
satisfied, we do not interpret our mediation analysis as causal. However, to make the flow of
the discussion easier, we will use the term "is mediated by" or "is not mediated by" to indicate
To complete our analysis, we also include all blocks of control variables simultaneously to
assess whether maltreatment affects adulthood personality, over and above the influence of all
early to late adolescence mediation variables. Finally, we add to this full specification
9
Furthermore, it also possible that the mediator variables measured in adolescence and personality measured in
adulthood are strongly correlated only because they measure imperfectly the same underlying trait. This is possible if
maltreatment experiences cause eg mental health problems in adolescence, which is an imperfect proxy of
personality in adolescence, while adolescent and young adulthood personality are strongly correlated (see Elkins et
al., 2017 for evidence of the latter). We thank an anonymous referee for having pointed out this possibility.
16
adulthood SES to assess whether childhood maltreatment is correlated with adulthood
5. Estimation Results
In this section we discuss the estimated associations between childhood exposure to parental
The OLS model that includes all control variables – but not adulthood SES - yields an
experience (10.5%), (3) agreeableness (10.2%), (4) conscientiousness (7.3%), and (5)
extraversion (5.1%). Each block of variables adds significantly (p-value < 0.01) to the
(standardized to mean 0, SD 1) on each of the five personality traits obtained from a basic and
a full estimation model, and its 90% confidence intervals. The dark-grey spike refers to the
OLS estimates, whereas the light-grey spikes refer to the S-FE models. In the OLS model with
basic control variables, maltreatment experiences are significantly positively associated with
neuroticism (0.15 SD) and extraversion (0.03 SD, not significant) and negatively associated
with agreeableness (-0.07 SD), conscientiousness (-0.10 SD), and openness to experience (-
0.05 SD, not significant). When controlling for family fixed effects, the associations for
neuroticism are reduced to 0.09 SD, but they remain significant at the 1% level. The
associations with openness to experience increase in magnitude to -0.06 SD and are now
significant at 5% level. In contrast, family fixed effects explain to a large degree the
10
We alert that this last estimation model cannot exclude the possibility of reverse causality, because adult
personality traits are measured in the same time period as adulthood SES.
17
by over 50% from -0.10 SD (significant at the 1% level) to -0.05 SD (not significant), and
The association between early-life maltreatment experiences and neuroticism is also robust to
physical or mental health problems. When controlling for the potential pathways via which
for sibling-comparisons is even slightly stronger when controlling for these potential
maltreatment experiences and extraversion when controlling for childhood and adolescent
developmental outcomes, independent of the estimation method (OLS: 0.06 SD, significant at
We have therefore shown that maltreatment experiences are robustly associated with
neuroticism and openness to experience, and to some extent with conscientiousness and
abuse or neglect – are the driving forces in the maltreatment-personality nexus. In Table 3, we
therefore report the associations between adult personality traits and four different types of
frequent maltreatment experiences using the siblings fixed effects (S-FE) model only.
11
We have tested for the possibility that maltreatment experiences have differential effects for children from low
and high SES background. For this reason, we have interacted the maltreatment measure with indicators for high and
low levels of education of the mother. We find no statistically significant interaction effect of maltreatment on any of
the five personality traits. The exception is for agreeableness. Children from high SES backgrounds who experience a
higher level of maltreatment than their sibling tend to be less agreeable in adulthood than their siblings (-0.18 SD,
significant at the 1% level). Full results are provided upon request.
18
Each type of maltreatment is coded as a dummy variable that takes the value 1 if the
individual reported that sexual abuse (or being slapped, left alone, or neglect of basic needs)
occurred more than ten times before grade 6. Column (1) reports the estimation results for the
S-FE model which controls for baseline characteristics. In column (2) to (5) we add
individually blocks of control variables that capture adolescent information on personality (2),
cognitive ability (3), physical health (4), and mental health (5). In column (6) we add all
blocks of control variables simultaneously. In column (7) we additionally add to the full
maltreatment, we obtain significant and robust associations between neuroticism and sexual
abuse or neglect, and between conscientiousness and openness to experience and neglect. In
the baseline model the association between neuroticism and sexual abuse is almost 0.50 SD
(significant at the 5% level), while its association with neglect is 0.32 SD (significant at the
1% level). The relationship between neuroticism and sexual abuse remains the same when
including cognitive ability, but the association is reduced by 20% when controlling for
physical health (0.42 SD) or mental health problems (0.39 SD). Interestingly, the association
increases by 16% when controlling for adolescent personality (0.58 SD). Including all control
5% level. Less than 10% of this association is explained by differences in adulthood SES
(0.41 SD).
The association between neglect and neuroticism is stable across the various control
variable specifications, ranging between 0.32 SD (baseline) and 0.25 SD (all controls).
19
Mental health problems have the largest influence on the association, as they explain 13% of
the relationship (a change from 0.32 SD in column (1) to 0.28 SD in column (5)). Controlling
for all potential pathways simultaneously reduces the association between neglect and
between -0.25 SD (baseline) and -0.24 SD (all controls), and they are significant at the 5%
level. The largest reduction in this association is driven by mental health problems, but they
explain less than 10% of the association (a change from -0.25 SD in column (1) to -0.23 SD in
column (5)). Similar to the relationship between neuroticism and neglect, adulthood SES does
not explain the association between conscientiousness and neglect. We find identical stable
associations between openness to experience and neglect, which range between -0.28 SD
(baseline) and -0.26 SD (all controls), and they are statistically significant at the 1% level.
Finally, there is tentative evidence that siblings who felt that they were frequently left alone,
relative to their sibling, are more extraverted than their sibling in young adulthood, although
the association is smaller in magnitude, ranging between 0.13 SD (baseline) and 0.12 SD (all
As a consequence, we conclude that sexual abuse experiences are only associated with
young adulthood neuroticism, while experiences of neglect are associated with higher levels
of neuroticism, and lower levels of conscientiousness and openness to experience. The most
important pathway via which these experiences affect young adulthood personality is mental
health problems. In the next section, we therefore assess whether these adverse experiences –
maltreatment and subsequent mental or physical health problems – explain the strong
outcomes that have been documented in eg Fletcher (2013) and Lundberg (2013) using Add
Health data.
20
6. Do Adverse Childhood Experiences Mediate the Relationship between Adult
We re-estimate the same OLS models as reported in Fletcher (2013, Table 4) and Lundberg
(2013, Table 1) to predict log earnings (models (1a), (1b)) or the probability to obtain a
college degree (models (2a), (2b)) with the Big Five personality traits, both measured in Wave
4. In models (a) we only control for family SES, ethnicity, education, cognitive ability,
number of siblings, and geographic region. In models (b) we control additionally for adverse
childhood experiences (maltreatment and physical and mental health problems). The
obtain a college degree do not change when controlling for adverse childhood experiences.
One important exception is that the estimated coefficient of the effect of conscientiousness on
both earnings and educational attainment is significantly reduced when moving from models
(a) to (b). In the base model for log earnings (1a), a 1 SD increase in conscientiousness is
associated with a 4.2 percent increase in earnings. In the full model (1b), this effect drops
significantly by 45 percent to 2.3 percent, which is not statistically different from zero. A
similar result is obtained for educational attainment. In the base model for educational
increase in the probability of obtaining a college degree. When also controlling for childhood
maltreatment experiences and adolescent health (2b), this association drops significantly by 38
21
attainment is also significantly reduced by 26%. This suggests that the productivity and
human capital boosting effects of conscientiousness are fully and partially, respectively,
7. Discussion
This study tests to which degree the Big Five personality traits, a commonly used domain to
proxy non-cognitive skills, have their origins in childhood maltreatment experiences. We use
effects models to control for the potentially confounding effects of shared environments and
genes. Some adverse childhood experiences predict significantly and robustly neuroticism,
agreeableness and extraversion. Neuroticism is the only trait significantly associated with
experience of frequent sexual abuse. Frequent parental neglect experiences are positively
associated with neuroticism, and negatively associated with conscientiousness and openness
to experiences.
Our findings must be considered in light of some important limitations. The amount of
variation in sexual abuse between siblings may be too small to detect statistically significant
effects, which is a general limitation of siblings-fixed effects models (Conley et al., 2007,
Bound & Solon, 1999). For instance, the associations between sexual abuse and agreeableness
or openness are sizable (0.14 SD and -0.18 SD), but their standard errors are too large to
detect effects with certainty (0.19 and 0.20).
We are also not able to fully control for reporting differences that are linked to
personality. This leaves open the possibility that some of the estimated effects are driven by
differences in the willingness to report maltreatment that also correlate with personality. It is
impossible to say for certain that the treatment effect of parental neglect on neuroticism,
22
siblings. More neurotic personalities may be more likely to inflate negative experiences such
as neglect - leading to false positives - while more open personalities may be more likely to
share harmful experiences. In both cases, we would overestimate the effect of maltreatment on
personality. On the other hand, frequent sexual abuse is likely to be underreported because of
its stigma and criminal nature. Memories of sexual abuse may even be repressed, a
phenomenon that is referred to as “dissociative amnesia”, which finds strong scientific support
(Schefflin & Brown, 1996). Due to these false negatives we are likely to underestimate the
true effect of sexual abuse on neuroticism. The alternative to self-reported maltreatment data
is administrative data on substantiated abuse and neglect from child protection services.
Although more accurate, such data tend to severely under-report the occurrence of
maltreatment (Petersen & Feit, 2014). For this reason, most major studies rely on self-reported
data (eg Felitti et al., 1998, Currie & Tekin, 2012, Hengartner et al., 2015).
Although our results cannot be given a causal interpretation, they provide a clearer
picture of what adulthood personality traits may capture. Importantly, the strong associations
are partially explained by these adverse childhood experiences. Our findings are useful to
applied researchers who seek to explore the meaning of the estimated associations between
adulthood personality and labor market outcomes. These findings complement the knowledge
we have already about the correlates of openness to experience with intelligence (see Almlund
et al., 2011 for an overview). Our results also emphasize the important role of the earlier-life
family environment in shaping personality. Thus, the findings are useful to policy makers who
search for windows of opportunity to boost children's non-cognitive skills through family
policy.
23
References
Almlund, M., Duckworth, AL, Heckman, JJ & Kautz, TD (2011). Personality psychology and
economics. In: Handbook of the economics of education, vol. 4, ed. Eric A.Hanushek, Stephen
Machin, and Ludger Woessmann, 1–182. Amsterdam: North-Holland, Elsevier Science.
American Academy of Child Adolescent Psychiatry (2009). ADHD - A Guide for Families.
http://www.aacap.org/cs/adhd_a_guide_for_families/what_is_adhd, Accessed February 10,
2010.
Asendorpf JB, Denissen JJA & van Aken MAG (2008). Inhibited and aggressive preschool
children at 23 years of age: personality and social transitions into adulthood. Developmental
Psychology 44, 997–1011.
Attanasio, O., Cattan, S., Fitzsimons, E., Meghir, C. & Rubio-Codina, M. (2015). Estimating
the Production Function for Human Capital: Results from a Randomized Control Trial in
Colombia. NBER Working Paper Nr. 20965.
Baldasaro RE, Shanahan MJ & Bauer DJ (2013). Psychometric properties of the mini IPIP in
a large, nationally representative sample of young adults. Journal of Personality Assessment
95(1), 74-84.
Borghans, L., Golsteyn, BHH, Heckman, JJ & Humphries, JE (2011). Identification Problems
in Personality Psychology. Personality and Individual Differences 51(3), 315–320.
Borkenau P., Riemann R., Spinath FM & Angleitner A. (2001). Genetic and environmental
influences on observed personality: Evidence from the German Observational Study of Adult
Twins. Journal of Personality and Social Psychology 80, 655–668.
Bound, J. & Solon, G. (1999). Double trouble: on the value of twins-based estimation of the
return to schooling. Economics of Education Review 18(2), 169-182.
Cameron, L., Erkal, N., Gangadharan, L. & Meng, X. (2013). Little emperors: Behavioral
impacts of china's One-Child Policy. Science 339(6122), 953–957.
Caspi, A., Harrington, HL, Milne, B., Amell, JW, Theodore, RF & Moffitt, TE (2003).
Children's behavioral styles at age 3 are linked to their adult personality traits at age 26.
Journal of Personality 71(4), 1467-6494.
Caspi, A. & Silva, P. (1995). Temperamental qualities at age 3 predict personality traits in
young adulthood: Longitudinal evidence from a birth cohort. Child Development 66, 486-
498.
Caspi, A., & Roberts, BW (2001). Personality development across the life course: the
argument for change and continuity. Psychological Inquiry 12, 49–66.
24
Caspi, A., Roberts, BW, & Shiner, RL (2005). Personality development: stability and change.
Annual Review of Psychology 56, 453–484.
Clark, C., Caldwell, T., Power, C., & Stansfeld, SA (2010). Does the influence of childhood
adversity on psychopathology persist across the lifecourse? A 45-year prospective
epidemiologic study. Annals of Epidemiology 20(5), 385-394.
Conley, D., Pfeiffer, KM, Velez, M. (2007). Explaining sibling differences in achievement
and behavioral outcomes: The importance of within- and between-family factors. Social
Science Research 36, 1087-1104.
Costa, PT Jr., McCrae, RR (1992). NEO Personality Inventory (NEO-PI-R) and NEO Five-
Factor Inventory (NEO-FFI) Professional Manual. Odessa, FL: Psychological Assessment
Resources Inc.
Cunha, F. & Heckman, JJ (2008). Formulating, identifying and estimating the technology of
cognitive and noncognitive skill formation, Journal of Human Resources 43(4), 738–82.
Cunha, F., Heckman, JJ & Schennach, SM (2010). Estimating the technology of cognitive and
noncognitive skill formation, Econometrica78(3), 883–931.
Currie, J. & Tekin, E. (2012). Understanding the cycle: childhood maltreatment and future
crime? The Journal of Human Resources 47(2), 509-549.
Deal, JE, Halverson, CF, Jr., Havill, V. & Martin, RP (2005). Temperament factors as
longitudinal predictors of young adult personality. Merrill-Palmer Quarterly 51, 315–334.
Del Bono, E., Francesconi, M., Kelly, Y. & Sacker, A. (2014). Early Maternal Time
Investment and Early Child Outcomes. IZA Discussion Paper Nr 8608. IZA Bonn.
Donnellan, MB, Oswald, FL, Baird, BM & Lucas, RE (2006). The mini-IPIP scales: Tiny-yet-
effective measures of the Big Five factors of personality. Psychological Assessment 18(2),
192–203.
Elkins, R., Kassenboehmer & Schurer, S. (2017). The stability of personality traits during
adolescence and young adulthood. Journal of Economic Psychology. Available online 6
January 2017. DOI 10.1016/j.joep.2016.12.005.
Fletcher, JM (2011). The medium term schooling and health effects of low birth weight:
Evidence from siblings. Economics of Education Review 30, 517–527.
Felitti, VJ, Anda, RF, Nordenberg, D., Williamson, DF, Spitz, AM, Edwards, Koss, MP &
Marks, JS (1998). Relationship of Childhood Abuse and Household Dysfunction to
25
Many of the Leading Causes of Death in Adults: The Adverse Childhood Experiences (ACE)
Study. American Journal of Preventive Medicine 14(4), 245-258.
Filer, RK (1986). The role of personality and tastes in determining occupational structure.
Industrial Labor Relations Review XX, 412–424.
Fletcher, JM (2013). The effects of personality traits on adult labor market outcomes:
Evidence from siblings. Journal of Economic Behavior & Organization 89(C), 122-135.
Fletcher, JM (2009). Childhood mistreatment and adolescent and young adult depression.
Social Science & Medicine 68, 799–806.
Fletcher, JM, Greena, JC &Neidell, MJ (2010). Long term effects of childhood asthma on
adult health. Journal of Health Economics 29, 377–387.
Fletcher, JM & Richards, MR (2012). Lower Wages And Employment In Young Adults
Diabetes's 'Health Shock' To Schooling And Earnings: Increased Dropout Rates And Lower
Wages And Employment In Young Adults. Health Affairs 31(1), 27-34.
Fraley, C., & Roberts, BW (2005). Patterns of Continuity: A Dynamic Model for
Conceptualizing the Stability of Individual Differences in Psychological Constructs Across
the Life Course. Psychological Review, 112, 60-74.
Gensowski, M. (2014). Personality, IQ, and lifetime earnings. IZA Discussion Paper Series
8235. June. IZA Bonn.
Gerson, R. & Rappaport, N. (2013). Traumatic stress and posttraumatic stress disorder in
youth: Recent research findings on clinical impact, assessment, and treatment. Journal of
Adolescent Health 52, 137–143.
Goldberg, LR, Johnson, JA, Eber, HW, Hogan, R., Ashton, MC, Cloninger, CR, &Gough, HG
(2006). The international personality item pool and the future of public domain personality
measures. Journal of Research in Personality 40, 84–96.
Goodwin, RG & Friedman, HS (2006). Health status and the Five Factor personality traits in a
nationally representative sample. Journal of Health Psychology 11, 643–654.
Hampson, SE, Goldberg, LR, Vogt, TM, & Dubanoski, JP (2006). Forty years on: teacher's
assessments of children's personality traits predict self-reported health behaviors and
outcomes at midlife. Health Psychology 25, 57–64.
Heckman, JJ & Kautz, T. (2012). Hard evidence on soft skills. Labour Economics 19, 451–
464.
Heineck, G. & Anger, S. (2010). The returns to cognitive abilities and personality traits in
Germany. Labour Economics 17(3), 535-546.
26
Hengartner, MP, Cohen, LJ, Rodgers, S, Müller, M, Rössler, W, Ajdacic-Gross, V (2015).
Association Between Childhood Maltreatment and Normal Adult Personality Traits:
Exploration of an Understudied Field. Journal of Personality Disorders 29(1), 1-14.
Hoffman, ML (2000). Empathy and moral development: Implications for caring and justice.
New York, NY: Cambridge University Press.
Judd, CM & Kenny, DA (2010). Data analysis. In D. Gilbert, ST Fiske, & G. Lindzey (Eds.),
The Handbook of Social Psychology (5th ed., Vol. 1, pp. 115-139), New York. Judge, TA,
Higgins, CA, Thoresen, CJ, and Barrick, MR (1999). The Big Five personaltiy traits,
general mental ability, and career success across the life span. Personality Psychology 52,
621–652.
Kaufman J., Yang B., Douglas-Palomberi H., Houshyar S., Lipschitz D., et al. 2004. Social
supports and serotonin transporter gene moderate depression in maltreated children.
Proceedings of the National Academy of Sciences 101, 17316–21.
Kern, ML, Friedman, HS, Martin, LR, Reynolds, CA, Luong, G. (2009). Conscientiousness,
Career Success, and Longevity: A Lifespan Analysis. Annals of Behavioral Medicine 37, 154–
163.
Koenen, K., Moffitt, T., Caspi, A., Taylor, A., Purcell, S. (2003). Domestic violence is
associated with environmental suppression of IQ in young children. Development and
Psychopathology 15, 297-311.
Koenig AL, Cicchetti D., Rogosch FA (2004). Moral development: The association between
maltreatment and young children's prosocial behaviors and moral transgressions. Social
Development 4(13), 97–106.
Krueger, RF & Eaton, N. (2010). Personality traits and the classification of mental disorders:
Toward a more complete integration in DSM-V and an empirical model of psychopathology.
Personality Disorder: Theory, Research, and Treatment 1(2), 97–118.
Krueger, RF, South, S., Johnson, W., Iacono, WG (2008). The heritability of personality is not
always 50%: gene-environment interactions and correlations between personality and
parenting. Journal of Personality 76(6): 1485–1521.
Lüdtke, O., Roberts, BW, Trautwein, U., & Nagy, G. (2011). A random walk down university
avenue: life paths, life events, and personality trait change at the transition to university life.
Journal of Personality and Social Psychology, 101(3), 620.
27
Luhmann, M., Orth, U., Specht, J., Kandler, C. and Lucas, RE (2014), Studying Changes in
Life Circumstances and Personality: It's About Time. European Journal of Personality, 28:
256-266.
Lundberg, S. (2013). The College Type: Personality and Educational Inequality, Journal of
Labor Economics 31(3), 421-441.
Marini, VA, & Kurtz, JE (2011). Birth order differences in normal personality traits:
Perspectives from within and outside the family. Personality and Individual Differences 51,
910–914.
McAdams, DP & Olson, BD (2010). Personality development: Continuity and change over
the life course. Annual Review of Psychology 61, 517-542.
McCrae, RR, Costa, PT, Jr., Ostendorf, F., Angleitner, A., Hrebickova, M., Avia, MD, et al.
(2000). Nature over nurture: Temperament, personality, and life span development. Journal
of Personality and Social Psychology 78, 173–186.
McCrae, RR, & Costa, PT Jr. (2008). A five-factor theory of personality, in Handbook of
Personality: Theory and Research, Vol. 3, eds LA Pervin, OP John, and RW Robins (p. 159–
181) New York: Guilford Press.McKinnon, DP, Fairchild, AJ & Fritz, MS (2007). Mediation
analysis. Annual Review of Psychology, 58, 593-614.
Moffitt, TE, Arseneault, L., Belsky, D., Dickson, N., Hancox, RJ, Harrington, H., . . ., &
Caspi, A. (2011). A gradient of childhood self-control predicts health, wealth, and public
safety. Proceedings of the National Academy of Sciences 108, 2693–2698.
Mueller, G., & Plug, E. (2006). Estimating the effects of personality on male and female
earnings. Industrial and Labor Relations Review 60 (1), 3–22.
Perkonigg, A., Kessler, R., Storz, S., & Wittchen, H. (2000). Traumatic events and post
traumatic stress disorder in the community: prevalence, risk factors and comorbidity. Acta
Psychiatrica Scandinavica 101(1), 46–59.
Petersen AC, Joseph J. & Feit M.,(2014) . New Directions in Child Abuse and Neglect
Research. Washington (DC): National Academies Press (US), 25(4), Consequences of Child
Abuse and Neglect.
Putnam, F. (2006). The impact of trauma on child development. Juvenile and Family Court
Journal. Winter, 1-11.
Roberts, BW (2009). Back to the future: personality and assessment and personality
development. Journal of Research in Personality 43 (2), 137-145.
28
Roberts, BW, & DelVecchio, WF (2000). The rank-order consistency of personality traits
from childhood to old age: A quantitative review of longitudinal studies. Psychological
Bulletin 126, 3-25.
Roberts, BW, Wood, D., & Smith, JL (2005). Evaluating the five factor theory and social
investment perspective on personality trait development. Journal of Research in Personality
39, 166-184.
Roberts, BW, Kuncel, NR, Shiner, R., Caspi, A., Goldberg, LR (2007). The power of
personality: The comparative validity of personality traits, socioeconomic status, and
cognitive ability for predicting important life outcomes. Perspectives on Psychological
Science 2(4), 313-45.
Roberts, BW, Wood, D, & Caspi, A. (2008). The development of personality traits in
adulthood. In OP John, RW Robins, & LA Pervin (Eds.), Handbook of personality: theory and
research (3rd edition, Ch 14, pp. 375-398). New York, NY: Guilford.
Samuel DB, & Widiger TA (2008). A meta-analytic review of the relationships between the
five-factor model and DSM-IV-TR personality disorders: a facet level analysis. Clinical
Psychology Review 28, 1326-42.
Scheflin, AW & Brown, D. (1996). Repressed memory or dissociative amnesia: What the
science says. Journal of Psychiatry and Law 24, 143–188.
Schurer, S., Kassenboehmer, S., Leung, F. (2015). Do universities shape their students'
personality? IZA Discussion Paper Nr. 8873, IZA Bonn.
Specht, J., Bleidorn, W., Denissen, JJA, Hennecke, M., Hutteman, R., Kandler, C., Luhmann,
M., Orth, U., Reitz, AK & Zimmermann, J. (2014), What Drives Adult Personality
Development? A Comparison of Theoretical Perspectives and Empirical Evidence. European
Journal of Personality 28, 216-230.
Spila, B., Makara, M., Kozak, G., Urbanska, A. (2008). Abuse in Childhood and Mental
Disorder in Adult Life. Child Abuse Review 17(2), 133-138.
Sulloway, FJ (1996). Born to rebel: Birth order family dynamics, and creative lives. New
York: Pantheon Books.
29
Trull, TJ, & Widiger, TA (2013). Dimensional models of personality: the Five-Factor Model
and the DSM-5. Dialogues in Clinical Neuroscience 15(2), 135-146.
Turkheimer E. (2000). Three laws of behavior genetics and what they mean. Current
Directions in Psychological Science 9, 160–164.
Turkheimer E., Haley A., Waldron M., D'Onofrio B., Gottesman, II (2003). Socioeconomic
status modifies heritability of IQ in young children. Psychological Science 14(6):623–628.
Tyrka, AR, Wyche, MC, Kelly, MM, Price, LH & Carpenter, LL (2009). Childhood
maltreatment and adult personality disorder symptoms: Influence of maltreatment type.
Psychiatry Research 165(3), 281-287.
Udry, JR (2003). The National Longitudinal Study of Adolescent Health (Add Health), Waves
I, II & III, 1994–2001 [Machine-readable data file and documentation]. Carolina Population
Center, University of North Carolina at Chapel Hill, Chapel Hill.
Widiger, AT, Simonsen, E., Krueger, R., Livesley, WJ & Verheul, R. (2005). Personality
disorder research agenda for the new DSM–V. Journal of Personality Disorder 19(3), 315–
338.
Widiger TA, & Trull TJ (2007). Plate tectonics in the classification of personality disorder:
shifting to a dimensional model. American Psychologist 62, 71-83.
Widom CS, DuMont K. & Czaja SJ (2007). A prospective investigation of major depressive
disorder and comorbidity in abused and neglected children grown up. Archives of General
Psychiatry 64(1), 49–56.
Widom CS, Czaja SJ & Paris J. (2009). A prospective investigation of borderline personality
disorder in abused and neglected children followed up into adulthood. Journal of Personality
Disorders 23(5), 433–446.
Widom CS, Czaja SJ, Bentley T. & Johnson MS (2012). A prospective investigation of
physical health outcomes in abused and neglected children: New findings from a 30-year
follow-up. American Journal of Public Health, 102(6),1135–1144.
Young, JK, & Beaujean, AA (2011). Measuring personality in Wave 1 of the national
longitudinal study of adolescent health. Frontiers in Psychology. doi:
10.3389/fpsyg.2011.00158, 13 July.
30
Table 1: Control variables for regression models
Childhood Maltreatment indicators (before grade 6, or age
12) Sexual abuse more than 10 times (0,1)
Slapped more than 10 times (0,1)
Left alone when shouldn't have more than 10 times (0,1)
Neglected basic needs more than 10 times (0,1)
or
Continuous measure of maltreatment obtained from factor analysis Note: Data on
maltreatment experiences was collected in Wave 3, with reference to experiences
that occurred before grade 6 (age 12)
31
Table 2: Number of sibling-pairs in the Siblings-Fixed
Effects model with differences in outcomes and treatment
% Outcome in Wave 4 ∆≠0
Nr Sample size
Sexual abuse -0.025 -0.104 -0.000 -0.039 -0.011 -0.089 -0.033 (0.206) (0.202) (0.206) (0.207) (0.208) (0.204) (0.203)
Being slapped 0.027 0.045 0.015 0.032 0.024 0.031 0.022 (0.088) (0.087) (0.088) (0.088) (0.088) (0.087) (0.087)
Left alone 0.129* 0.119 0.135* 0.113 0.136* 0.120 0.118 (0.074) (0.073) (0.074) (0.074) (0.075) (0.073) (0.073)
Neglect of needs -0.130 -0.112 -0.130 -0.120 -0.120 -0.112 -0.120 (0.106) (0.105) (0.106) (0.106) (0.107) (0.105) (0.104)
R-squared 0.003 0.026 0.009 0.012 0.004 0.036 0.044 Neuroticism
Sexual abuse 0.499** 0.578*** 0.491** 0.419** 0.394* 0.446** 0.410** (0.212) (0.208) (0.211) (0.211) (0.212) (0.207)
(0.207)
Being slapped 0.000 -0.005 -0.005 -0.010 0.028 -0.001 0.002 (0.090) (0.089) (0.090) (0.090) (0.090) (0.088) (0.088)
Left alone 0.050 0.060 0.061 0.062 0.039 0.067 0.068 (0.077) (0.075) (0.076) (0.076) (0.076) (0.075) (0.074)
Neglect of needs 0.319*** 0.291*** 0.300*** 0.305*** 0.277** 0.253** 0.257** (0.110) (0.107) (0.109) (0.109) (0.109)
(0.106) (0.106)
R-squared 0.028 0.055 0.041 0.050 0.043 0.083 0.086 Agreeableness
Sexual abuse 0.138 0.101 0.145 0.088 0.170 0.073 0.119 (0.190) (0.187) (0.190) (0.190) (0.191) (0.189) (0.187)
Being slapped 0.003 0.025 0.000 -0.012 -0.006 -0.002 0.003 (0.081) (0.080) (0.081) (0.081) (0.081) (0.081) (0.080)
Left alone 0.031 0.019 0.032 0.020 0.026 0.009 0.007 (0.068) (0.068) (0.068) (0.068) (0.069) (0.068) (0.067)
Neglect of needs -0.158 -0.130 -0.157 -0.176* -0.145 -0.148 -0.148 (0.098) (0.097) (0.098) (0.098) (0.098) (0.097) (0.096)
R-squared 0.104 0.112 0.105 0.115 0.107 0.123 0.136 Conscientiousness
Sexual abuse -0.028 -0.096 -0.014 -0.000 0.061 0.001 0.025 (0.212) (0.209) (0.212) (0.212) (0.213) (0.209) (0.209)
Being slapped -0.021 -0.029 -0.022 -0.012 -0.051 -0.050 -0.048 (0.090) (0.089) (0.090) (0.090) (0.090) (0.089) (0.089)
Left alone -0.007 -0.010 -0.005 -0.018 -0.001 -0.007 -0.009 (0.077) (0.075) (0.077) (0.076) (0.077) (0.075) (0.075)
Neglect of needs -0.249** -0.253** -0.246** -0.236** -0.229** -0.242** -0.242** (0.109) (0.108) (0.109) (0.109) (0.109)
(0.107) (0.107)
R-squared 0.013 0.031 0.015 0.030 0.023 0.054 0.056 …to be continued on next page.
33
Table 3: Estimation Results Siblings Fixed Effects Model (N=2,319), continuing
+All
control (7) Baseline (1) (2) (3) (4) (5)
+Adult Open to + Cogn. Ability + Phys. Health + Mental Health + All
SES Experience controls (6)
+ Child Temp
Sexual abuse -0.182 -0.224 -0.193 -0.146 -0.140 -0.171 -0.120 (0.203) (0.200) (0.202) (0.203) (0.203) (0.200) (0.199)
Being slapped -0.035 -0.054 -0.035 -0.022 -0.049 -0.054 -0.057 (0.086) (0.085) (0.086) (0.086) (0.086) (0.085) (0.085)
Left alone 0.077 0.070 0.062 0.064 0.065 0.043 0.041 (0.073) (0.072) (0.073) (0.073) (0.073) (0.072) (0.072)
Neglect of needs -0.275*** -0.271*** -0.261** -0.273*** -0.257** -0.262** -0.267*** (0.105) (0.103) (0.104) (0.104)
(0.104) (0.103) (0.102)
R-squared 0.011 0.024 0.026 0.023 0.018 0.047 0.055 Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1.
Estimated coefficient on binary variable of type of maltreatment is interpreted as standard-deviation change in personality
trait k. In each model is estimated with a Siblings-Fixed Effects model (2,319 siblings). The Baseline model (1) includes
controls for age, gender, family characteristics, and birth-weight. Subsequent columns (2)-(5) add separately a block of
variables - adolescent personality, adolescent cognitive ability, adolescent physical health problems, and adolescent mental
health problems - respectively. The all-controls model (6) includes all control variables simultaneously. Model (7) adds adult
socioeconomic status variables (education, earnings) to model (6). Full estimation results for the full model reported in
column (7) is reported in Table A7 Technical Appendix.
34
Table 4: Estimated relationship between earnings or education and personality with
and without controlling for childhood health and maltreatment indicators
(Reported: marginal effects)
Dependent Variable
Log Log % College degree College degree %
Earnings (1a) Earnings (1b) Diff (2a) (2b) Diff
35
D a
S n
n
o
i
t
s
i
e
a
r
r
il
2
e l e l e l e l e
Baselin Full contro Baselin Full contro Baselin Full contro Baselin Full contro Baselin Full control
37
Table A2: Determinants of probability to
be in the siblings sample
VARIABLES Coef.
(SE)
Observations 5,470
R-squared 0.004
Note: Linear probability model. Standard
errors in parentheses. *** p<0.01, ** p<0.05,
* p<0.1
38
Table A3: The Big Five Personality Traits Measured in Wave 4
Conscientiousness (C) H4PE11 11. I often forget to put things back in their proper place
H4PE19 19. I like order
H4PE27 27. I make a mess of things
Neuroticism (N) Characteristics are related to anxiety and emotional liability versus being placid and emotionally
stable
Note: Factor analysis is used to predict the first principal factor from the four questions each. Scores are standardised to mean
0 and standard deviation 1.
39
Table A4. Personality Questionnaire Wave 1 (as suggested by Young and Beaujean, 2011)
Neuroticism You have a lot of good qualities* H1PF 30
You have a lot to be proud of* H1PF 32
You like yourself just the way you are* H1PF33
You feel like you are doing everything just about right* H1PF34
You feel socially accepted* H1PF35
You feel wanted and loved* H1PF36
Conscientiousness When you have a problem to solve, one of the first things you do is get as many facts about the
problem as possible* H1PF18
When you are attempting to find a solution to a problem, you usually try to think of as
many different ways to approach the problem as possible* H1PF19
When making decisions, you generally use a systematic method for judging and
comparing alternatives* H1PF20
After carrying out a solution to a problem, you usually try to analyze what went right and
what went wrong* H1PF21
Note: Child hood temperament was part of Wave 1; Young and Beaujean (2011) demonstrate the construct validity of the
each facet. Cronbach's alpha for Neuroticism, Extraversion and Conscientiousness is 0.86, 0.76, and 0.76, respectively.
40
Table A5: Mental and Physical Health Measures
Physical Health Mental Health
General Health: A general health status measure is constructed
from a question asked in Wave 1: “In general, how is your
health?”. The cohort members could respond with poor, fair, good,
very good, or excellent. Despite the subjectivity of this variable,
research has shown that it is a strong predictor of objective health
measures such as mortality and health care utilization (Miilunpalo
et al., 1997).
41
Table A6: Full estimation results including all control variables (excluding adult socioeconomic status)
OLS Siblings Fixed Effects
Extra Neur Agree Consc Open Extra Neur Agree Consc Open
(1) (2) (3) (4) (5) (1) (2) (3) (4) (5)
Maltreatment (Std) 0.059** 0.091*** -0.041 -0.076*** -0.026 0.053* 0.073** -0.009 -0.050 -0.067** (0.030) (0.029) (0.028)
(0.029) (0.029) (0.030) (0.030) (0.028) (0.031) (0.029)
Age
-0.016 0.001 0.011 0.020*-
0.032*** -0.001 -0.007 0.029** 0.026* -0.030**
(0.012) (0.012) (0.012) (0.012) (0.012) (0.013) (0.014) (0.012) (0.014) (0.013)
Female
0.030 0.334*** 0.467*** 0.131***-
0.228*** 0.061 0.275*** 0.645*** 0.170***
-
0.132***
(0.043) (0.042) (0.041) (0.042) (0.041) (0.046) (0.047) (0.043) (0.047) (0.045)
Family inform. miss.
0.031 -0.071 -0.040 -0.062 0.002 -0.047 -0.126- 0.197***
-
0.220*** -0.035
(0.052) (0.051) (0.050) (0.051) (0.051) (0.075) (0.077) (0.070) (0.077) (0.074)
Log birth weight 0.063 -0.052 -0.104 0.043 -0.101 -0.035 -0.018 0.037 0.104 0.133 (0.086) (0.084) (0.083) (0.084) (0.083) (0.130)
(0.132) (0.120) (0.133) (0.127)
Birth weight missing -0.038 -0.039 0.235** 0.081 0.215** -0.091 -0.188* 0.257** 0.164 0.109 (0.098) (0.095) (0.094) (0.095)
(0.094) (0.109) (0.111) (0.101) (0.112) (0.107)
PVT score Wave 1 -0.019 -0.110*** 0.164*** -0.063*** 0.250*** -0.101*** -0.085*** -0.029 -0.069** 0.139*** (0.024) (0.023)
(0.023) (0.023) (0.023) (0.031) (0.031) (0.029) (0.032) (0.030)
Math grade Wave 1 0.006 -0.040* 0.025 0.040* -0.014 0.028 -0.052** 0.004 0.004 0.013 (0.021) (0.021) (0.020) (0.021) (0.021)
(0.021) (0.022) (0.020) (0.022) (0.021)
Math grade missing -0.020 0.030 -0.007 -0.039 0.021 -0.129 -0.027 -0.034 0.010 -0.095 (0.095) (0.092) (0.091) (0.092) (0.092)
(0.092) (0.094) (0.086) (0.095) (0.091)
General health Wave 1 0.061** -0.024 0.025 0.029 0.063*** 0.042* -0.025 0.022 0.073*** 0.025 (0.025) (0.024) (0.024) (0.024)
(0.024) (0.024) (0.025) (0.023) (0.025) (0.024)
Obesity Wave 1 0.129 -0.018 -0.051 -0.241*** 0.016 0.104 0.001 -0.033 -0.172** -0.130 (0.080) (0.078) (0.076) (0.078) (0.077)
(0.083) (0.084) (0.077) (0.085) (0.081)
Obesity Missing 0.047 -0.099 0.176 0.110 0.015 -0.050 -0.229* -0.071 -0.053 -0.029 (0.134) (0.130) (0.128) (0.130) (0.129) (0.133)
(0.135) (0.123) (0.136) (0.130)
Asthma Wave 4 0.039 0.139** 0.067 -0.039 0.129** 0.030 0.081 -0.039 0.062 0.029 (0.058) (0.057) (0.056) (0.057) (0.056) (0.056)
(0.057) (0.052) (0.057) (0.055)
42
Diabetes Wave 4
0.032 0.415*** 0.090 -0.079 -0.127 0.028 0.547*** 0.345*** -0.188-
0.371***
(0.131) (0.128) (0.126) (0.127) (0.126) (0.125) (0.127) (0.116) (0.128) (0.123)
Diff. hands Wave 1 0.078 -0.048 0.154 -0.314 -0.031 0.004 -0.309 0.024 0.407 0.167 (0.326) (0.317) (0.313) (0.317) (0.314) (0.306)
(0.311) (0.283) (0.314) (0.300)
Difficulty feet Wave 1 0.198 0.277* -0.031 0.051 0.071 0.261 -0.124 -0.175 0.034 -0.073 (0.171) (0.166) (0.164) (0.166) (0.165)
(0.172) (0.175) (0.159) (0.176) (0.169)
Epilepsy Wave 4 0.205 0.272* -0.057 -0.243 -0.115 0.243 0.196 -0.159 -0.293* 0.115 (0.159) (0.154) (0.152) (0.154) (0.153) (0.148)
(0.151) (0.137) (0.152) (0.145)
Blindness Wave 4
0.054 0.349 -0.224 -0.310 0.039 -0.258 0.223-
0.752*** -0.270 -0.357
(0.291) (0.283) (0.279) (0.283) (0.281) (0.272) (0.277) (0.252) (0.279) (0.267)
Diff feet missing -0.659* 0.287 -0.463 0.197 -0.134 -0.065 0.578 -0.415 -0.327 -0.502 (0.393) (0.382) (0.376) (0.382) (0.379) (0.369)
(0.375) (0.341) (0.378) (0.362)
Diff hands missing 0.533 -0.145 0.458 0.201 0.288 0.117 -0.202 0.365 0.351 0.766** (0.379) (0.369) (0.363) (0.369) (0.366) (0.353)
(0.359) (0.327) (0.362) (0.346)
Depressed Wave 1 0.124 0.317*** -0.008 0.106 0.089 0.005 0.157** -0.020 0.168** -0.003 (0.079) (0.077) (0.076) (0.077) (0.077)
(0.072) (0.073) (0.067) (0.074) (0.071)
ADHD Wave 4 0.042 0.108 0.043 -0.276*** 0.103 -0.073 0.170 -0.002 -0.206* 0.060 (0.107) (0.104) (0.103) (0.104) (0.103) (0.108)
(0.110) (0.100) (0.111) (0.106)
Learning disab. Wave 1
-0.047 0.082 -0.114* -0.227***-
0.215*** -0.004 0.084 -0.119*
-
0.243*** -0.176**
(0.072) (0.070) (0.069) (0.070) (0.070) (0.074) (0.076) (0.069) (0.076) (0.073)
Learning disability missing 0.030 -0.035 -0.206 -0.471*** -0.238 0.009 0.008 -0.013 -0.202 -0.164 (0.171) (0.166) (0.163) (0.166)
(0.164) (0.162) (0.165) (0.150) (0.166) (0.159)
-
Conscientiousness Wave 1 -0.027 0.006 -0.030 -0.092*** -0.052** -0.041** 0.011 0.016
0.075*** -0.048**
(0.021) (0.021) (0.021) (0.021) (0.021) (0.021) (0.021) (0.020) (0.022) (0.021)
Neuroticism Wave 1
-
0.140*** 0.092*** -0.009 -0.108***
-
0.073*** -0.104*** 0.131***
-
0.069***
-
0.087*** -0.057**
(0.027) (0.026) (0.026) (0.026) (0.026) (0.027) (0.028) (0.025) (0.028) (0.027)
Extraversion Wave 1 0.106*** -0.073*** 0.061*** 0.028 0.008 0.069*** -0.034 0.044** 0.033 0.001 (0.022) (0.021) (0.021)
(0.021) (0.021) (0.022) (0.022) (0.020) (0.023) (0.022)
Conscientiousness Wave 1 Missing -0.361* -0.081 -0.127 0.214 0.153 0.001 -0.423* -0.112 0.136 0.197 (0.218) (0.212) (0.208)
(0.212) (0.210) (0.227) (0.231) (0.210) (0.233) (0.223)
Extraversion Wave 1 Missing 0.018 0.090** -0.063 0.000 -0.019 0.012 0.031 -0.041 0.011 -0.109**
(0.043) (0.042) (0.042) (0.042) (0.042) (0.044) (0.045) (0.041) (0.045) (0.043)
Education Wave 4
Earnings Wave 4
Constant
0.052 -0.026 -0.475 -0.777* 0.914** -0.062 0.292-
1.259*** -1.208** 0.620
(0.442) (0.430) (0.423) (0.430) (0.426) (0.509) (0.518) (0.471) (0.521) (0.499)
Observations 2,319 2,319 2,319 2,319 2,319 2,319 2,319 2,319 2,319 2,319 R-squared 0.051 0.120 0.102 0.073 0.105 0.036 0.081
0.122 0.052 0.046 Note: Personality traits are standardized to mean 0 and standard deviation 1. Standard errors in parentheses. ***
p<0.01, ** p<0.05, * p<0.1.
Table A7: Full estimation results with indicator variables for maltreatment type (Siblings-Fixed
Effects model)
Extra Neur Agree Consc Open
(1) (2) (3) (4) (5)
Sexual abuse -0.033 0.410** 0.119 0.025 -0.120 (0.203) (0.207) (0.187) (0.209) (0.199)
Being slapped 0.022 0.002 0.003 -0.048 -0.057 (0.087) (0.088) (0.080) (0.089) (0.085)
Left alone 0.118 0.068 0.007 -0.009 0.041 (0.073) (0.074) (0.067) (0.075) (0.072)
Neglect of needs
-0.120 0.257** -0.148 -0.242**-
0.267***
(0.104) (0.106) (0.096) (0.107) (0.102)
Age
-0.008 -0.004 0.027** 0.025*-
0.034***
(0.013) (0.014) (0.012) (0.014) (0.013)
Female
0.045 0.285*** 0.589*** 0.146***-
0.158***
(0.047) (0.048) (0.043) (0.048) (0.046)
Family inform. miss. -
-0.030 -0.135*- 0.186*** 0.216*** -0.023
(0.075) (0.077) (0.069) (0.077) (0.074)
Log birth weight -0.068 0.002 0.042 0.105 0.118 (0.130) (0.132) (0.119) (0.133) (0.127)
Birth weight missing -0.085 -0.173 0.247** 0.156 0.101 (0.109) (0.111) (0.100) (0.112) (0.107)
0.121*** -0.070** -0.056*
PVT score Wave 1 - -
0.084*** 0.115***
(0.031) (0.032) (0.029) (0.032) (0.031)
Math grade Wave 1 0.013 -0.042* -0.012 -0.003 -0.000 (0.022) (0.022) (0.020) (0.022) (0.021)
Math grade missing -0.099 -0.058 0.014 0.032 -0.057 (0.093) (0.094) (0.085) (0.095) (0.091)
General health Wave 1 0.034 -0.020 0.013 0.069*** 0.017 (0.024) (0.025) (0.023) (0.025) (0.024)
Obesity Wave 1 0.102 0.004 -0.059 -0.182** -0.139* (0.083) (0.084) (0.076) (0.085) (0.081)
Obesity Missing -0.044 -0.234* -0.054 -0.045 -0.021 (0.132) (0.135) (0.122) (0.136) (0.130)
Asthma Wave 4 0.030 0.081 -0.040 0.061 0.028 (0.056) (0.057) (0.052) (0.057) (0.055)
Diabetes Wave 4
0.086 0.504*** 0.404*** -0.161-
0.319***
(0.125) (0.128) (0.116) (0.129) (0.123)
Diff. hands Wave 1 0.065 -0.345 0.059 0.426 0.219 (0.305) (0.311) (0.281) (0.313) (0.299)
Difficulty feet Wave 1 0.269 -0.122 -0.165 0.020 -0.084 (0.171) (0.175) (0.158) (0.176) (0.168)
Epilepsy Wave 4 0.243* 0.189 -0.124 -0.275* 0.130 (0.148) (0.150) (0.136) (0.152) (0.145)
Blindness Wave 4
-0.252 0.261-
0.730*** -0.276 -0.353
(0.273) (0.278) (0.251) (0.280) (0.267)
Diff feet missing -0.048 0.572 -0.364 -0.321 -0.478 (0.367) (0.374) (0.339) (0.377) (0.361)
45
Diff hands missing 0.056 -0.178 0.250 0.323 0.710** (0.353) (0.359) (0.325) (0.362) (0.346)
Depressed Wave 1 0.014 0.155** -0.026 0.164** -0.006 (0.072) (0.073) (0.066) (0.074) (0.071)
ADHD Wave 4 -0.056 0.149 -0.022 -0.218* 0.049 (0.108) (0.110) (0.100) (0.111) (0.106)
Learning disab. Wave 1
0.028 0.055 -0.076-
0.220*** -0.135*
(0.075) (0.076) (0.069) (0.077) (0.073)
Learning disability missing 0.045 -0.040 0.039 -0.175 -0.121 (0.162) (0.165) (0.149) (0.166) (0.159)
-
Conscientiousness Wave 1 -0.043** 0.011 0.013 0.078*** -0.049**
(0.021) (0.021) (0.019) (0.022) (0.021)
Observations 2,319 2,319 2,319 2,319 2,319 R-squared 0.044 0.086 0.136 0.056 0.055 Note: Personality
traits are standardized to mean 0 and standard deviation 1. Standard errors in parentheses. *** p<0.01, **
p<0.05, * p<0.1.