Gender Differences in Young Children’s Temperament Traits: Comparisons Across Observational and Parent-Report Methods
Thomas M. Olino,1 C. Emily Durbin,2 Daniel N. Klein,3
Elizabeth P. Hayden,4 and Margaret W. Dyson3 1University of Pittsburgh School of Medicine 2Michigan State University 3Stony Brook University 4University of Western Ontario
Evidence supporting the continuity between child temperament and adult personality traits is accumulating. One important indicator of continuity is the presence of reliable gender differences in traits across the lifespan. A substantial literature demonstrates gender differences on certain adult personality traits and recent meta-analytic work on child samples suggests similar gender differences for some broad and narrow domains of temperament. However, most existing studies of children rely only on parent-report measures. The present study investigated gender differences in temperament traits assessed by laboratory observation, maternal-report, and paternal-report measures. Across three independent samples, behavioral obser- vations, maternal-report, and paternal-report measures of temperament were collected on 463 boys and 402 girls. Across all three methods, girls demonstrated higher positive affect and fear and lower activity level than boys. For laboratory measures, girls demonstrated higher levels of sociability and lower levels of overall negative emotionality (NE), sadness, anger and impulsivity than boys. However, girls demonstrated higher levels of overall NE and sadness than boys when measured by maternal reports. Finally, girls demonstrated lower levels of sociability based on paternal reports. Results are discussed in relation to past meta-analytic work and developmental implications of the findings.
Contemporary models of temperament emphasize continuity of individual differences in emotion, motivation, and social behavior across the lifespan (Caspi & Shiner, 2006). Continu- ity may refer to similarity in the manifestation and structure of traits in children and adults or the extent to which traits have parallel correlates across the lifespan. There are well- replicated self-reported gender differences between men and women on higher- and lower-order dimensions of personality (Costa, Terracciano, & McCrae, 2001; Feingold, 1994; Lynn & Martin, 1997), raising questions of when these difference emerge during development and whether the magnitude of these differences changes over the course of development.
Meta-analyses of adult samples have explored gender differ- ences on both higher- and lower-order personality traits. Fein- gold (1994) reported on gender differences on personality traits derived from multiple questionnaire measures, including those from the Five-Factor Model tradition (FFM; McCrae & Costa, 1987). Women demonstrated significantly higher levels of the higher-order trait of Extraversion, but no significant gender differences were found for Neuroticism, Agreeableness, Con- scientiousness, or Openness to New Experience. For lower-
order facets of the FFM, women were higher on anxiety (a facet of Neuroticism), gregariousness (a facet of Extraversion), and tendermindedness (a facet of Agreeableness). By contrast, men were higher on the assertiveness and self-esteem facets of Extraversion. In a more recent meta-analysis, Lynn and Martin (1997) found that for higher-order domains assessed by the Eysenck Personality Inventory (EPQ; Eysenck & Eysenck, 1975), women had higher levels of Neuroticism than men, while men scored higher than women on Extraversion and Psychoti- cism. Costa et al. (2001) examined studies using the Revised NEO Personality Inventory (NEO-PI-R; Costa & McCrae, 1992) and found that women scored higher than men on Neu- roticism, Agreeableness, and components of Extraversion (warmth, gregariousness, positive emotions), and Openness
This work was supported by NIMH K01-MH092603 (TMO), RO1-MH069942 (DNK), and GCRC Grant M01-RR10710 to Stony Brook University from the National Center for Research Resources.
Direct correspondence to Thomas M. Olino, University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, 3811 O’Hara St, Pittsburgh, PA, 15213. Email: [email protected]
Journal of Personality 81:2, April 2013 © 2012 Wiley Periodicals, Inc. DOI: 10.1111/jopy.12000
(openness to feelings). In contrast, men scored higher on the assertiveness and excitement seeking facets of Extraversion, and two facets of Openness (fantasy and openness to ideas).
In sum, these studies converge to suggest that the best replicated gender differences in personality traits in adulthood concern the two higher-order traits tapping emotionality. Com- pared to men, women describe themselves as higher on Neuroticism (particularly the facet of Anxiety). Findings for Extraversion varied by facet. Affiliative components of Extra- version, such as warmth and sociability, were higher in women than men (Costa et al., 2001). In contrast, measures of Asser- tiveness, as well as the EPQ Extraversion scale that includes much agentic content, were higher for men than women (Costa et al., 2001). Thus, in addition to differences in higher-order personality dimensions, lower-order aspects of certain traits may reveal additional effects, consistent with claims that lower-order constructs, rather than higher-order, provide greater power to detect individual differences (e.g., Costa et al., 2001). Finally, self-reports of the remaining traits, Agreeableness, Conscientiousness, and Openness, do not appear to differ across adult men and women.
Results of studies of adolescent samples largely parallel those of adults. In a birth cohort sample, Roberts, Caspi, and Moffitt (2001) found that 18-year-old females reported higher levels of constraint, the harm avoidance and stress reaction components of Neuroticism, and the affiliative aspects of Extraversion (positive emotionality, well-being) than males. By contrast, males were higher than females on aggression (an aspect of low Agreeableness), and agentic elements of Extraversion (achievement and social potency). Longitudinal developmental studies of adolescents complement these cross- sectional findings. In a meta-analysis exploring developmental change in self-reported personality traits, Roberts, Walton, and Viechtbauer (2006) found that gender did not moderate mean- level change through adolescence for any personality dimen- sion explored, suggesting that most of the gender differences observed in adulthood are apparent in adolescence, implying that differences are early-emerging.
Thus, a better understanding of gender differences in personality/temperament traits in childhood is critical for understanding the developmental context in which gender dif- ferences are reliably observed in adults and adolescents first appear. Importantly, whereas the literatures on adults and ado- lescents largely employ self-report to assess these dimensions, the vast majority of studies of children use parent reports and a minority of studies use observational measures. Hence, our understanding of the emergence of gender differences in traits must also consider the influence of assessment method on the measurement of temperament.
In a meta-analysis on gender differences in child tempera- ment, Else-Quest, Hyde, Goldsmith, and Van Hulle (2006) examined studies of children aged 3 months to 13 years. As our goal is to describe the evidence for the developmental conti- nuity of personality, we summarize their results by emphasiz- ing constructs related to the three higher-order dimensions that
are identified in most models of adult personality (i.e., the “Big Three” of Extraversion/Positive Emotionality, Neuroticism/ Negative Emotionality, and Constraint; Tellegen & Waller, 2008), as well as of child temperament (as exemplified by Rothbart’s psychobiological model; Rothbart, Ahadi, Hershey, & Fisher, 2001), including the corresponding traits of Surgency/Positive Emotionality, Negative Affectivity, and Effortful Control. Thus, we focus on results from those studies that examined parent report measures of child temperament reflecting Rothbart’s model and used the corresponding instru- ments, which captured a large minority of effect sizes in the meta-analysis.
Else-Quest and colleagues (2006) found no significant gender difference on Negative Affectivity (d = -.06). However, boys exhibited higher levels of broadband Surgency/PE than girls (d = .55) and girls demonstrated higher levels of Effortful Control (d = -1.01). For narrowband dimensions of Surgency, boys displayed higher levels of activity (d = .23), high- intensity pleasure (d = .30), and impulsivity (d = .18) than girls; no differences were found on approach (d = -.04), shyness (d = -.03), or smiling (d = .01). For narrowband dimensions of Negative Affectivity, boys displayed lower levels of fear (d = -.12) than girls. No significant differences were found for anger (d = .04), discomfort (d = -.17), distress to limits (d = .01), sadness (d = -.10), or soothability (d = .05). Among the Effortful Control narrowband dimensions, girls displayed higher levels of attentional focusing (d = -.16), attentional shifting (d = -.16), inhibitory control (d = -.41), low-intensity pleasure (d = -.29), and perceptual sensitivity (d = -.38) than boys.
The Else-Quest et al. (2006) meta-analysis suggests some continuity for some Big Three personality dimension gender differences in older adolescent, adult, and youth samples. For example, although boys and girls do not differ on overall NE as do adult men and women, girls are more fearful than boys. Other findings were less consistent with those from adult samples, but were similar to those found in adolescents. Specifically, girls had higher levels of several facets of Effortful Control, a trait corresponding most closely to Conscientiousness/Constraint. Although adult men and women do not differ on Conscientious- ness, adolescent females are higher on Constraint (Roberts et al., 2001). Some gender differences in youngsters varied according to the developmental period in which they were assessed. However, the effects were small for each age group and no comparisons within these developmental periods were statistically significant.
Thus, it appears that some gender differences in tempera- ment traits in adulthood are foreshadowed in childhood, but others do not reliably emerge until adolescence. However, dif- ferences between child and adolescent/adult samples could reflect methodological factors, rather than developmental effects. Most studies of gender differences in child tempera- ment have relied on parent report measures of temperament. Far less is known about gender differences in child tempera- ment when assessed by methods other than parent report. Else-
Olino, Durbin, Klein, et al.120
Quest et al. (2006) could not conduct moderator analyses to examine whether gender differences varied according to mea- surement approach as only eight studies relied on observa- tional methods. Although useful for providing preliminary evidence, these studies are not definitive, as all but two had relatively modest sample sizes. Of the two with larger samples, both examined only a limited number of temperament dimen- sions, and one (Arcus & Kagan, 1995) did not provide enough information to compute effect sizes. The other large study reported that boys displayed higher activity level and greater difficult temperament, but lower distress to limitations and fear than girls across ages 3 to 5 (Zahn-Waxler, Schmitz, Fulker, Robinson, & Emde, 1996).
As few studies have used observational methods, it is dif- ficult to determine whether gender differences are limited to a single methodology (i.e., parent reports), a particular reporter (i.e., mothers, whose reports are typically the principal source of parent questionnaires), or if they generalize to other assess- ment methods. Without such information, it is unclear whether gender differences on particular traits first appear in adoles- cence because of critical developmental transitions during that period, the effects of maturation or other factors that shape personality, or simply because self-report methods used in older samples are more sensitive to gender differences than parent-report measures typically used with child samples.
It is difficult to compare parent-report and observational studies of gender differences in child temperament traits due the well established fact that only modest associations of trait scores are found across methods (e.g., Gartstein & Marmion, 2008; Stifter, Willoughby, & Towe-Goodman, 2008). However, both assessment approaches have merits and are associated with important outcomes (Dougherty, Klein, Durbin, Hayden, & Olino, 2010; Hayden, Klein, & Durbin, 2005). Limited convergence across assessment methods suggests at least two possibilities for how gender differences may manifest. First, although parent report and observational methods exhibit only modest agreement on rank-ordering of trait levels of children, they may produce similar findings for mean-level differences between males and females. Alternatively, gender differences may vary (i.e., be moderated) by assessment method, a result that was not formally tested in Else-Quest et al. (2006). No published studies have evaluated these possibilities within the same data set. Thus, it is important to directly test whether patterns of gender differences are similar across assessment methods in order to make substantive interpretations about the presence and magnitude of gender differences in temperament traits in youngsters.
Here, we examined whether gender differences were evident in a broad range of temperament traits assessed using both laboratory-based observational measures and maternal- and paternal-reports. We used data collected from three community-based samples of preschool- and early elementary school-aged children. This developmental period is particularly important for identifying gender differences in temperamental traits, as temperament begins to stabilize around age 3 (e.g.,
Caspi & Shiner, 2006), suggesting that this may be the earliest age at which male-female differences can be reliably detected.
As previous findings suggest that nonsignificant differences on higher-order traits may mask significant differences on subordinate traits and findings for higher-order traits may not generalize to all lower-order dimensions, we included both broadband (i.e., higher-order) and narrowband (i.e., lower- order) temperament dimensions to discern the structural level at which gender differences on traits were most prominent. To directly compare gender differences assessed using multiple methods, we identified eight narrowband scales from Roth- bart’s Child Behavior Questionnaire (CBQ; Rothbart et al., 2001) that included similar content to behaviors coded from structured laboratory assessments of temperament. We selected three narrowband scales to represent a broad concep- tualization of Positive Emotionality (PE; positive affect [PA], appetitive motivation, and sociability) and three to represent an overall Negative Emotionality (NE) construct (sadness, anger, and fear). We also examined impulsivity and activity level as two additional narrowband traits, but did not consider them as part of a higher-order trait, as their location within personality dimensions is controversial (Buss, Block, & Block, 1980; Eysenck, 1978). We selected these scales to maximize concep- tual overlap with the Laboratory Temperament Assessment Battery (Lab-TAB; Goldsmith, Reilly, Lemery, Longley, & Prescott, 1995), which was developed using the same general theoretical framework as Rothbart’s model.
We expected to replicate results from Else-Quest et al.’s (2006) meta-analysis for maternal reports, such that girls would have higher levels of fear and boys would have higher levels of activity and impulsivity. Although less information is available on paternal reports, we anticipated that results for fathers’ reports of child temperament would be generally similar to those for mothers’. Finally, we had few strong pre- dictions for gender differences on laboratory observational methods, for two reasons: (1) the available evidence regarding gender differences on child traits assessed via observation is inconsistent, and (2) as noted above, parent-report and obser- vational measures typically demonstrate modest convergence (e.g., Seifer, Sameroff, Barrett, & Krafchuk, 1994), making it questionable to assume that results obtained via one method would be found for the other. However, given that sociocultural expectations regarding gender differences may have a stronger influence on parent-reports than on objective coding of child behavior in response to laboratory tasks (Seifer, 2003), we expected larger effect sizes for gender differences on parent- report than laboratory measures.
Methods Data came from three studies of child temperament: the Stony Brook Temperament Study (SBTS), the Child Personality Development Project (CPDP), and the Northwestern Family Temperament Study (NFTS), yielding 865 child participants with one child included per family.
Gender Differences in Temperament 121
The SBTS sample consisted of 559 three-year-old children and their parents from a suburban community in Long Island, New York. Participants were recruited through a commercial mailing list. Children who lived with at least one English- speaking biological parent and were free of significant medical conditions or developmental disabilities were included (Olino, Klein, Dyson, Rose, & Durbin, 2010). The CPDP sample consisted of 100 three-year-old children from Long Island, New York (Durbin, Klein, Hayden, Buckley, & Moerk, 2005). Children were recruited from a commercial mailing list (51.9%) and ads in local newspapers and preschools (48.1%). Participants obtained through the two methods did not differ on any of the child temperament variables used in this study. Participants in the NFTS sample (N = 206) were recruited from the greater Chicago area for a study of child temperament and were between the ages of 36 and 83 months. Participants were recruited through a commercial mailing list (38.1%), Internet, print, and radio ads (21.4%), referrals from commu- nity agencies (26.2%), and other approaches (e.g., word of mouth, 14.2%). Demographic characteristics for each study and comparisons between study samples are displayed in Table 1. Child gender distribution and percentage of children whose biological parents were currently married did not sig- nificantly differ between the samples. However, child age, race, Peabody Picture Vocabulary Test (PPVT; Dunn & Dunn, 1997) scores, maternal age, paternal age, and percentage of employed mothers differed between samples. Some differ- ences reached statistical significance, but substantive implica- tions were minimal. For example, differences in PPVT scores were at most a quarter of a standard deviation. The differences in child age and race across studies, particularly the NFTS relative to the CPDP and SBTS, bolster generalizability of the results.
Child Assessment Procedures Parent Report Measures. For all studies, mothers and fathers completed the Child Behavior Questionnaire (CBQ;
Rothbart et al., 2001). The CBQ is a widely used 195-item caregiver report measure of temperament for 3- to 7-year-old children. For the scales included in the present study, they have good internal consistency (mean a = .76, range from .69 to .92 in the original publication), modest-to-strong inter-parental consistency (mean r = .48, range from .23 to .79), and moderate-to-strong test–retest reliability (mean r = .67, range from .55 to .79 for maternal reports and mean r = .65, from .58 to .76 for paternal reports) across a 2-year period (Rothbart et al., 2001). The CBQ derived scales are associated with con- current home observations of temperament (Buckley, Klein, Durbin, Hayden, & Moerk, 2002) and prospectively associated with emotional and behavioral problem outcomes (Dougherty et al., 2010; Eisenberg, et al., 2003), thus showing both con- vergent and predictive validity.
In two of the samples (CPDP, SBTS), the parent who accom- panied the child to the laboratory assessment (usually the mother) completed the CBQ during the lab visit, and question- naires were sent home to the other parent to be completed and returned through postal mail. In the NFTS, both mothers and fathers completed the CBQ at home and returned them via postal mail. To maximize conceptual similarity of traits assessed by observational and parent-report measures, we focused on the Smiling/Laughter, Approach Anticipation, Shyness, Fear, Sadness, Anger, Impulsivity, and Activity Level scales from the CBQ. Internal consistency estimates are similar to those reported in Rothbart et al. (2001) and are presented for each scale for each individual study in Table 2. Higher-order PE was computed as the average of standardized values of Smiling/ Laughter, Approach Anticipation, and Shyness (reverse scored). Higher-order NE was computed as the average of standardized values of Fear, Sadness, and Anger. As shown in Table 1, the percentage of mothers who completed the CBQ differed across the studies. Mothers from the CPDP had the highest completion percentage and mothers from the NFTS had the lowest. No differences were found in the percentage of fathers completing the CBQ between studies. Due to missing items on the CBQ, the actual Ns for each scale varied modestly.
Table 1 Demographic Characteristics of the Three Samples
SBTS CPDP NFTS F/c2
Child Age (months) 42.24 (3.14)a 43.20 (3.60)a 56.38 (12.02)b 365.46*** Child Sex, Male‡ 302 (54.0) 53 (53.0) 108 (51.9) .28 Child Race, Caucasian‡ 487 (87.1)a 89 (94.7)b 158 (77.4)c 18.37*** PPVT 102.82 (14.00)a 103.47 (13.87)a 106.62 (15.07)b 6.73** Maternal Age (years) 35.99 (4.44)a 33.81 (4.10)b 36.96 (4.89)c 16.41*** Paternal Age (years) 38.27 (5.39)a 36.73 (5.53)b 38.72 (6.62)a 4.19* Maternal Employment‡ 286 (51.2)a 60 (56.6)a,b 132 (64.0)b 9.02* Parent Marital Status‡ 524 (93.7) 97 (97.0) 191 (92.7) 1.34 Maternal CBQ Completion‡ 514 (91.9)a 99 (99.0)b 162 (77.6)c 42.63*** Paternal CBQ Completion‡ 399 (71.4) 82 (82.0) 156 (75.7) 5.24
Note. Table entries are M (SD). Variables labeled as ‡ display n (%). SBTS = Stony Brook Temperament Study; CPDP = Child Personality Development Project; NFTS = Northwestern Family Temperament Study; PPVT = Peabody Picture Vocabulary Test; CBQ = Child Behavior Questionnaire. Different subscripts reflect significant differences at p < .05. *p < .05. **p< .01. ***p < .001.
Olino, Durbin, Klein, et al.122
Laboratory Assessment of Temperament. The laboratory batteries lasted approximately two hours, when children partici- pated in standardized laboratory episodes with a female experi- menter. Most episodes were from the Lab-TAB (Goldsmith et al., 1995); one (Exploring New Objects) was adapted from an original Lab-TAB episode, and two (Making a T-shirt and Dress Up) were developed by one of us (CED). Episodes were designed to elicit individual differences in temperament traits related to emotionality, behavioral engagement, and social behavior. The child took breaks between episodes to return to a baseline state before entering a new situation. Each task was videotaped through a one-way mirror and later coded. Although episodes are primed to elicit specific dimensions of tempera- ment (indicated in parentheses below), all dimensions were rated in all episodes to provide indices of temperament across multiple contexts. The episodes are described below in the order that they were presented to the children in the SBTS and CPDP (numbers in brackets reflect the episode order for the NFTS).
Risk Room (fear; administered in SBTS and CPDP only). The episode allows children to explore a set of novel, ambigu- ous stimuli (e.g., a Halloween mask, a black box).
Tower of Patience (inhibitory control; interest; SBTS and CPDP only). The child and experimenter alternated turns in building a tower together. The experimenter took increasing delays before placing her block on the tower during each of her turns.
Making a T-shirt (PA; NFTS only ). The child decorated a T-shirt using fabric markers; he or she took the decorated T-shirt home as a gift.
Arc of Toys (PA; interest; anger; SBTS and CPDP only). The child played with toys for a five-minute period. The experi- menter then asked the child to clean up the toys.
Disappointing Toy (sadness, anger; NFTS only ). The experimenter showed the child a picture of an unappealing toy and pictures of two appealing and asked the child which she or
he would prefer. The experimenter left the room and returned with the nonpreferred toy. After 1 minute, an assistant entered with the preferred toy, and the child and experimenter played together for 3.5 min.
Stranger Approach (fear; SBTS, CPDP, and NFTS ). The child was left alone briefly in the assessment room while the experimenter left to look for toys. A male research accomplice entered the room and spoke to the child while walking closer.
Make That Car Go (PA, interest; SBTS and CPDP only). The child and experimenter raced remote controlled cars.
Dress Up (PA; NFTS only ). The child and experimenter played with costumes. The experimenter took a photograph of the child in his or her costume.
Transparent Box (persistence, interest, anger, sadness; SBTS, CPDP, and NFTS ). The experimenter locked an attractive toy in a transparent box. The child was then left alone with a set of keys to attempt to open the box. After a few minutes, the experimenter returned to the child and told them that she had left the wrong set of keys. The child was then encouraged to use the new keys to open the box and play with the toy.
Exploring New Objects (fear; SBTS, CPDP, and NFTS ). The child explored a set of novel and ambiguous stimuli (e.g., a mechanical spider, toy mice inside a pet carrier).
Pop-up Snakes (PA, interest; SBTS, CPDP, NFTS ). The child and experimenter surprised the child’s mother with a can of potato chips that actually contained coiled snakes.
Perfect Circles (anger, sadness, persistence; SBTS, CPDP, and NFTS ). The experimenter repeatedly asked the child to draw a circle. Each attempt was mildly criticized. After about two minutes, the experimenter praised the child for his or her efforts.
Popping Bubbles (PA, interest; SBTS, CPDP, and NFTS ). The child and experimenter played with a bubble- shooting toy.
Table 2 Temperament Dimension Characteristics
SBTS CPDP NFTS
Laboratory Mat. Pat. Laboratory Mat. Pat. Laboratory Mat. Pat.
ICC a a a ICC a a a ICC a a a
PE PA .92 .87 .73 .76 .94 .90 .82 .81 .90 .92 .71 .78 Engagement .75 .68 .71 .71 .72 .56 .76 .65 .65 .69 .68 .62 Sociability .83 .82 .92 .90 .93 .81 .92 .93 .93 .86 .92 .91
NE Fear .64 .63 .73 .65 .66 .59 .74 .63 .66 .68 .76 .72 Sadness .79 .81 .64 .66 .82 .67 .56 .60 .79 .74 .74 .61 Anger .73 .68 .79 .76 .84 .75 .82 .83 .81 .74 .79 .81
Impulsivity .74 .69 .76 .65 — — .70 .72 .70 .77 .77 .79 Activity .84 .73 .76 .70 .75 .83 .71 .73 .94 .82 .79 .77
Note. SBTS = Stony Brook Temperament Study; CPDP = Child Personality Development Project (CPDP); NFTS = Northwestern Family Temperament Study; ICC = intra- class correlation; a = Cronbach’s Alpha; PE = positive emotionality; PA = positive affectivity; NE = negative emotionality. ICCs are presented for directly observed behaviors (but not derived scores [i.e., PE and NE]). ICCs are based on 35 cases in the SBTS, 15 in the CPDP, and 27 in the NFTS. PA was indexed by the CBQ using the Smiling/Laughter scale. Sociability was indexed by the CBQ using the Shyness scale (reverse); Engagement was indexed by the CBQ using the Approach Anticipation scale.
Gender Differences in Temperament 123
Snack Delay (inhibitory control; SBTS and CPDP only). The child was instructed to wait for the experimenter to ring a bell before eating a snack. The experimenter systematically increased the delay before ringing the bell.
Painting a Picture (interest; CPDP only). The child played with watercolor pencils and crayons.
Box Empty (anger, sadness; SBTS, CPDP, and NFTS ). The child was given an elaborately wrapped box, under the impression that a toy was inside. After the child discovered that the box was empty, the experimenter returned with several toys for the child to keep.
Laboratory Episode Coding Procedures Affective Codes in SBTS and NFTS. Each display of facial, bodily and vocal positive affect, fear, sadness, and anger in each episode was rated on a three-point scale (low, moderate, high intensity). Weighted sums (low intensity = 1; moder- ate = 2; high = 3) were computed separately within each channel (facial, bodily, vocal) across the episodes, standard- ized, and then summed across the channels to derive total scores for positive affect, fear, sadness, and anger.
Affective Codes in CPDP. Discrete emotions (positive affect, anger, sadness, and fear) were assessed by coding facial, vocal, and bodily indicators during each episode. Each episode was scored (from 0–3 or 0–4) based on the number and inten- sity of affective displays. These indicators were averaged to produce composite variables for each emotion.
Behavioral Codes. Ratings of additional dimensions of child behavior were made via single global ratings (scored 0–3) based on all behaviors relevant to each dimension during that episode. Ratings of engagement were based on how invested and absorbed the child appeared in …
CHAD 496 Forum Post Guide
This guide supplements the information provided in the syllabus. Refer to the syllabus for the main content guidelines.
Purpose of Comments & Critiques Forum Posts
• These forum posts are an opportunity for you to showcase your critical thinking and research method skills as they relate to child development research and practice
• Understanding how child development researchers work is important to understanding where early childhood curriculum and developmental milestones come from and how research informs developmental practice
• Good critical thinking is essential in the field of child development, even if you do not plan to be a researcher or clinician. If you work in the field of child development, education, or other related fields, parents may see you as an expert and may turn to you to help them understand research or internet articles, especially as there is a lot of conflicting and confusing information out there
• While good research can help children’s healthy development, bad or pseudoscientific information can hurt or damage children’s development. Knowing good research and being able to think about research and claims in a critical way can help you point families toward rigorous research and steer them away from pseudoscience
Scoring Rubric Comments and Critiques posts are scored using the following grading scale
• 8 = Follows guidelines with no spelling or grammar errors. Student makes it clear that they not only understand the article, but also understand the broader themes of the article by bringing up nuanced points and connecting these to themes covered in the article to that of class. It is clear the student understands the implications of the research and critically engaged with the reading and through their original analysis and response. Students arguments or points broadly follow a “what-why-why it matters” style. Student’s writing demonstrates a strong academic style and demonstrates many examples of original thinking on the subject.
• 7 = Student follows guidelines with minimal spelling or grammar errors. Student makes it clear that they understood the content and may make some connections to outside reading. Student makes it clear they understand some implications of the work although they may not fully follow through the point to the conclusion of how this could broadly affect children/research/practice. Student’s points or arguments broadly follow a “what-why” style. Student’s writing is mostly academic with some minor shifts in tone and demonstrates some examples of original thinking on the subject.
• 6 = Student follows guidelines although there may be some spelling or grammar errors. Student understands the premise of the article and makes at least one connection to themes discussed in class. Student’s response may rely too heavily on summary and student’s engagement or critical thinking on the article may not be fully clear. There may be portions where the student’s writing seems disjointed. Student understands the overall point of the article but does not connect this to overall class themes. Student’s points or arguments
broadly follow a “what” style. Student’s writing may not be academic in tone and the student’s writing only partly demonstrates their original thinking on the subject
• 5 = Student gives a summary of the article without providing any original engagement with the material. It is unclear what the student was thinking. The summary given meets the expected length of a forum post but is vague and most importantly, lacks a student’s voice and connections to the reading.
• 4 = Student missed the premise of the article completely, wrote something completely tangential and not related to the course, or did not follow guidelines for forum post length.
• 1 to 3 = Student failed to meet the standards of a score of “4”, and/or the submission was penalized for late submission
• 0 = No submission, and/or the submission was penalized for late submission (10+ days late)
Ways to Structure Strong Forum Posts
• Think about what, why, and why it matters when considering a given article’s conclusion or argument
o Example: § What: This article provides a review of several studies that have found that
Measure A does a poor job of predicting academic success in children from households that speak English as a second language.
§ Why: Measure A’s questions rely on American pop-culture knowledge and idioms that children from immigrant families or children who speak English- as-a-second language (ESL) may have less exposure to.
§ Why it Matters: Measure A has been implemented in some school districts to determine which children can enroll in “gifted and talented” courses. This suggests that this measure may unfairly underestimate the abilities of children from immigrant and ESL backgrounds, which can unfairly prevent these students from accessing higher quality educational experiences.
• There are few “wrong answers” but more points will be given for students that show they can:
o Think critically, make connections to other research, general topics in psychology/child development, potential developmental outcomes, or downstream consequences of the topic
• Remember that while you should be making connections to child development as a whole, if you propose changes or implications of research it should be clear how it directly relates to the research topic.
o Stay focused on making connections and exploring implications relevant to the overall concept or aspect of research being discussed.
o Try not to focus too much on what the researchers could have done o Don’t focus on what researchers could studied instead à Stay focused on the topic
being discussed • Feeling lost? Not sure how a research topic could influence other work? Try this:
o Read the article once or twice, jot down some notes on: § The big picture focus (where this fits into research or practice) § What researchers looked at specifically § What they found
§ What this article makes you think of based on your understanding of psychology and child development. What theories or studies do you think of first?
o Now take a break, let these ideas simmer a little bit. Don’t worry about making big picture connections or understanding the implications right away.
o After some time away, return to your notes. It’s likely that some ideas or connections you hadn’t previously thought of have occurred to you
Notes on Specific Limitations for Empirical Research Articles (see syllabus for main content guidelines):
• Many students choose to talk about sample size as a limitation, but you should always be able to explain why this is a limitation for this study in particular. More points will be given to answers that can clearly articulate how the sample size of a study is tied directly to how well we can interpret the results of this particular study
o Weak: This study only had 20 participants, but researchers should have gotten more participants as more data is better in a research study and makes a study more valid.
o Strong: This study’s sample size of 20 participants is a potential limitation as the research study was examining a rare behavior that none of the 20 participants demonstrated. If researchers had expanded their sample size, there may have been more participants recruited who show this behavior and the researchers would have more comparative data could have been assessed.
o Strongest: This study’s sample size of 20 participants is a potential limitation as the research study was examining a rare behavior that none of the 20 participants demonstrated. If researchers had expanded their sample size, there may have been more participants recruited who show this behavior and the researchers would have more comparative data could have been assessed. While we know this behavior is present in child populations, the way this study conducted their findings may inadvertently lead individuals to believe this behavior is rare or does not happen which may limit further study and treatment of this behavior.
• Representativeness is important in psychology and child development research, but remember that no study can ever be 100% representative. A study that is 100% representative would have to recruit every single person from a population of interest which is not possible. Instead, consider who the researchers want to represent, what is the population they are trying to represent with their smaller sample? Students who choose to cover this in their forum post should be aware that more points will be given to responses that tie directly into the population researchers want to look at.
o Weak: This study only looked at early letter recognition in children from age 3-4 which does not represent older children’s literacy knowledge. Researchers should expand their study to look at children up until 5th grade otherwise this study is biased.
o Strong: This study’s research into early letter recognition in 3-4 year old children may have been limited by demographics as only children enrolled in a preschool program were assessed. Given that lower income children may not have the means to attend preschool, these results may better represent middle class or affluent children rather than all preschool-aged children.
o Strongest: This study’s research into early letter recognition in 3-4-year-old children may have been limited by demographics as only children enrolled in a preschool
program were assessed. Given that lower income children may not have the means to attend preschool, these results may better represent middle class or affluent children rather than all preschool-aged children. Should early literacy milestones be based off this research, this may unfairly underestimate the skills of lower income children as the standards would be artificially higher than the actual mean of all preschool-aged children and further disadvantage children from lower income backgrounds
Tips for Forum Posts Regarding Non-Empirical Articles (see syllabus for main content guidelines):
At first non-empirical articles may seem trickier to comment and respond to than empirical articles. Non-empirical articles may include methods you are unfamiliar with, an opinion/anecdotal piece, or may be a review of multiple studies. This means that non-empirical articles can be more difficult to evaluate them and comment on than experimentally-based work. It can be more challenging to determine whether you think a study presented good evidence for a concept or relationship in psychology/child development. Instead, when reviewing and responding to non-empirical articles:
• Consider the theories the research is based on as well as the connections you can make to psychological research and child development.
o What connections can you make to scientific knowledge of children’s development? o Does this work suggest a connection between different approaches to child research
and clinical practice? o Does the research topic expand on a particular theory or approach to studying
development? • Consider the implications for this topic and whether it changes how we understand
development or involve children in research. o Does it suggest that there may be flaws in the way previous research/clinical practice
has been conducted in the past? o What are the broader implications if previous work or interventions are potentially
invalid? How might this have affected individuals or groups of people? • Consider the broad scope of this research topic.
o Does this research connect to larger societal issues? o Does this change whether we consider current and past work in diverse populations? o Does this affect how researchers should conduct future research and clinical
interventions? • Consider the conclusion of the study.
o Based on previous research or theories do the interpretations, opinions, or conclusions by the article’s authors seem well-supported?
o Is there a specific portion of their logic that could be explained by another explanation?