Parent Mind-Mindedness, Sensitivity, and Infant Affect: Implications for Attachment with Mothers and Fathers

Abstract

Previous research examining links between parenting and attachment has focused on behavioral aspects of parenting such as sensitivity. However, by assessing how parents reflect on infants’ mental states (mind-mindedness) we gain a broader understanding of parenting and how it impacts attachment. Mothers, fathers, and their infants (N = 135) participated in the Still Face Paradigm (SFP) at 3-, 5-, and 7- months of age, and the Strange Situation with mothers at 12 months and fathers at 14 months. Parent sensitivity and infant affect were coded from the SFP and all videos were transcribed and later coded for parents’ use of appropriate and non-attuned mind-mindedness toward their infants. Attachment with each parent was coded from the Strange Situation. Mixed effects models examined trajectories of parents’ mind-mindedness in relation to parent sensitivity and infant affect across attachment groups. Significant differences between parent gender and attachment category were detected. Specifically, parents who were less sensitive were also less mind-minded toward insecure-avoidant infants; parents used more non-attuned mind-mindedness when infants had higher negative affect. Findings suggest that, in addition to parent sensitivity, parents’ use of appropriate and non-attuned mind-mindedness during a parent-infant interaction provides insight into the developing attachment relationship for mothers and fathers.

1. Parent Mind-Mindedness, Sensitivity, and Infant Affect: Implications for Attachment with Mothers and Fathers

Attachment theory suggests that infants develop a bond with caregivers that helps organize their interactions and behaviors, both with the caregiver and the surrounding environment (Ainsworth & Wittig, 1969; Bowlby, 1969). Important components of parent-infant social exchanges early in life impact the developing attachment relationship, including parent sensitivity, infant affect, and parental mind-mindedness (Meins & Fernyhough, 2010). Recently, there have been calls to include fathers in research on parenting and child development with a focus on direct statistical comparisons with mothers (Cabrera, Volling, & Barr, 2018) and mind-mindedness in particular (Shai & Meins, 2018). Fathers play a unique role within the family system; fathers’ behaviors are not simply a reflection or parallel of mothers’ behaviors, but can have different determinants and effects on child development (Braungart-Rieker, Garwood, Powers, & Wang, 2001; Cabrera, Tamis-LeMonda, Bradley, Hofferth, & Lamb, 2000; Cummings & Davies, 2002; Grossmann et al., 2002; Tamis-LeMonda, Shannon, Cabrera, & Lamb, 2004). Thus, we examine how mothers’ and fathers ’ face-to-face interactions with their infants during the Still-Face Paradigm (Tronick, Als, Adamson, Wise, & Brazelton, 1978) in early infancy impact later infant-parent attachment.

1.1. Infant-Parent Attachment

Early attachment classifications are defined by a set of behaviors the infant exhibits during separations and reunions with caregivers (Ainsworth, Blehar, Waters, & Wall, 1978). The attachment bond in early infancy serves as an organizational framework for relationships later in life (Hazan & Shaver, 1994). Research on determinants of infant-parent attachment traditionally focuses on parenting behaviors such as responsiveness and sensitivity. A sensitive caregiver is one who interprets her infant’s needs and accommodates her own behavior to respond to those needs accordingly (Ainsworth et al., 1978). Infants who have experienced a more sensitive and responsive caregiving environment are more likely to develop a secure attachment (Braungart-Rieker et al., 2014; de Wolff & van IJzendoorn, 1997; Mesman, van IJzendoorn, & Bakermans-Kranenburg, 2009). In contrast, infants who have experienced a more rejecting, less sensitive, or inconsistent caregiving environment are more likely to develop an insecure attachment with their caregiver (Bretherton & Munholland, 2008), though patterns differ for insecure-avoidant and insecure-resistant infants. Insecure-avoidant infants do not use caregivers as a secure base of exploration or comfort and tend to avoid contact with caregivers. Insecure-resistant infants exhibit both contact and avoidance behaviors with caregivers and also have a harder time exploring their surroundings, even in the presence of a caregiver.

1.2. Parent Sensitivity and Mind-Mindedness

Sensitivity has been a core focus of attachment research, yet there are relatively modest relations between observed mother (r = .24) and father (r = .17) sensitivity and infant-parent attachment (see (de Wolff & van IJzendoorn, 1997; van IJzendoorn & de Wolff, 1997) for meta-analyses). Researchers have made an effort to update and expand upon Bowlby’s initial model relating sensitivity to attachment (e.g., Bretherton, 1999; Meins, Fernyhough, Fradley, & Tuckey, 2001) and account for these smaller effect sizes. Parents’ ability to correctly interpret their infants’ signals may be related to the parents’ ability to understand their infants’ emotional, cognitive and physical needs. Some alternative characterizations of this type of behavior include reflective functioning (Fonagy, Steele, Steele, Moran, & Higgitt, 1991), insightfulness (Koren-Karie, Oppenheim, Dolev, Sher, & Etzion-Carasso, 2002; Oppenheim & Koren-Karie, 2002; Oppenheim, Koren-Karie, Dolev, & Yirmiya, 2012), and mind-mindedness (Meins, 2013; Meins et al., 2012, 2001). Each of these theoretical constructs uses slightly different methods to understand how parents think about their infants’ behavior. For example, reflective functioning is coded from interviews with parents about a variety of potential scenarios but does not necessarily focus on their own child. Insightfulness assessments asks parents to talk through videos they see of themselves with their babies and then code those responses as insightful (or not). Mind-mindedness reflects parents’ ability to use appropriate mind-related comments to convey information about their intentions and their infants’ internal mental state (Beeghly, Bretherton, & Mervis, 1986; Meins et al., 2001). These three constructs essentially measure similar modes of parenting using different methods. For the purposes of this work, we focus on mind-mindedness.

Mind-mindedness was proposed as a reinterpretation of parental sensitivity, particularly in relation to infant-parent attachment (Meins, 2013; Meins et al., 2001). Parents who use more appropriate mind-minded comments are more attuned to their child’s needs, which is necessary to correctly interpret and respond to the infant’s behavioral cues (Beeghly et al., 1986). On the other hand, parents may also use non-attuned mind-minded comments, reflecting inappropriate or incorrect interpretations of their infants’ mental state. Of note, parents’ higher use of appropriate comments does not necessarily mean they do not also use non-attuned comments; these are not mutually exclusive categories. In terms of attachment, for parents of insecure-avoidant infants, there may be a disconnect between the parents’ correct interpretation or thought about the infant’s needs (mind-mindedness) and the parents behavioral response to the infant (sensitivity) (Bernier & Dozier, 2003).

Mind-mindedness can be assessed during parent-child face-to-face interactions, when parents often “talk for” the infant. During such interactions, parents’ use of appropriate mind-related comments reflect their recognition of and attunement with their infant’s signals (McMahon & Bernier, 2017; Meins et al., 2012).

1.3. Mind-mindedness and Infant-Parent Attachment

Though perhaps not as robust a predictor as sensitivity (McMahon & Bernier, 2017), there is evidence that mind-mindedness relates to infant-parent attachment (Laranjo, Bernier, & Meins, 2008; Lundy, 2003). Mothers’ mind-related language such as “You’re so happy” or “Are you interested in that?” during a 20-minute free play at six months predicted attachment security with infants at 12 months more strongly than did a behavioral measure of maternal sensitivity (Meins et al., 2001). Similarly, infants later classified as insecure-avoidant and insecure-resistant had mothers who were less appropriate to their needs and used less mind-related language than mothers of secure infants (Meins et al., 2012). Mothers who use more non-attuned mind related comments, or attempt to interpret their infants mental state but do so incorrectly, have infants later classified insecure but the same was not found for infant-father attachment (Arnott & Meins, 2007).

To our knowledge, few studies have examined how fathers’ mind-mindedness relates to infant attachment relationships (McMahon & Bernier, 2017; Zeegers, Colonnesi, Stams, & Meins, 2017), with none including both father behavioral sensitivity and mind-mindedness. Lundy (2003) showed that parents (mothers and fathers) who used more comments reflecting on an infant’s thoughts and needs at six months (i.e., “you want Daddy’s glasses”) had higher security of attachment rated by a parent reported Q-sort seven months later. Additionally, mind-related language related to attachment security with mothers and fathers differently (Lundy, 2003). Both mothers and fathers who showed greater use of language relating to infants’ thoughts (e.g., “are you concentrating on something?”) had higher attachment security scores; fathers’, but not mothers’, increased use of emotionally driven mind-related language (e.g., “You’re really fascinated by that,” “You’re bored with that already”) related to higher attachment security scores with fathers. Arnott and Meins (2007) found that parents’ non-attuned mind-mindedness (where the parent misinterpreted the infant’s mental state) was related to insecure attachment with mothers but not fathers. Thus, processes differ for infant-mother and infant-father dyads. Notably, the sample sizes for these studies were small; Lundy (2003) studied 16 families, and Arnott and Meins (2007) examined only 18 mother-infant and 15 father-infant dyads.

Mind-mindedness is comprised of both appropriate and non-attuned mind related comments, though fewer papers report on attachment related differences between the two constructs. A recent report indicates that children who are insecurely attached in the early school years have mothers who use more non-attuned comments reflecting on the child’s mental state (Meins, Bureau, & Fernyhough, 2018). Non-attuned mind related comments reflect parents’ misrepresentations of their infant or child’s internal mental state. It is possible that the parent is incorrectly reading her infant’s needs or the parent is imposing her own perceptions onto the infant. Parents higher in intrusiveness and lower insensitivity have infants later classified insecure (Braungart-Rieker et al., 2014); it follows that parents who use more non-attuned mind related comments may also have infants rated insecure.

1.4. Infant Affect

Notably, neither Arnott and Meins (2007) nor Lundy (2003) included measures of infant affect, yet it is possible that mothers’ and fathers’ socialization of infant affect differentially impacts the developing attachment relationship (Cummings, Davies, & Campbell, 2002; Planalp & Braungart-Rieker, 2013; Volling & Belsky, 1991); mother and father expectations and responses toward infants may differ depending on infant emotion. In a sample of preschool aged children, fathers were more responsive to girls’ sad/anxious emotions and boys’ happy/angry emotions (Chaplin, Cole, & Zahn-Waxier, 2005).

Mothers’ and fathers’ mind-mindedness in response to infant distress may differentially reflect their willingness to accept and acknowledge their infants’ negative emotions. In studies involving infants (Braungart-Rieker et al., 2001, 2014; Mangelsdorf, McHale, Diener, Goldstein, & Lehn, 2000), adolescents (Laible, 2007), or adults (Collins, 1996), attachment style is directly related to or moderated by affect and/or affect regulation. Infants who are less positive (Braungart-Rieker et al., 2001) or show more distress (Leerkes & Wong, 2012) are more likely to be classified insecure. It may be more difficult for parents to interpret their infant’s signals and respond appropriately with infants who exhibit more distress or negativity. Infant affect may differentially impact infant-mother and infant-father relationships Thus, we add to a growing body of research on mind-mindedness by examining both mothers and fathers and incorporating infant affective responses during the Still-Face Paradigm.

1.5. Rationale and Hypotheses

Gaining insight into how and when parents recognize and respond to infant signals may help us better understand how secure bonds form and aid in the development of parenting interventions targeted at recognizing such thought processes. In this work, we examine relations between parent mind-mindedness, sensitivity, and infant affect in early infancy across three attachment groups (secure, insecure-avoidant, insecure-resistant). We anticipate that patterns of relations will differ for infants depending on attachment classification with mothers and fathers.

In addition, parenting behaviors and infant affect influence not only infant attachment security (secure vs. insecure), but they can also alter the ways with which infants exhibit insecurity – with avoidant or resistant behaviors. In addition, mothers’ sensitivity during distressing tasks is a more robust determinant of attachment security than sensitivity during a free-play (Leerkes, 2011). The Still-Face Paradigm (SFP) provides an interesting context in which to study this process. Unlike free play situations, which tend to only elicit positive affect, the SFP is designed to elicit both positive and negative affect. Thus, we can observe parents’ interpretation of the range of young infants’ affective behaviors and assess the extent to which attachment group differences (secure, insecure-avoidant and insecure-resistant) are manifest by variation in mind-mindedness, parent sensitivity, and infant affect. Our hypotheses are as follows:

1. Previous research (Arnott & Meins, 2007; Lundy, 2003; Meins et al., 2012, 2001) has looked at mind-mindedness at six months and attachment at one year but not necessarily changes over time. Previous work with this sample (Planalp et al., 2013) shows slight increases in parent sensitivity across time. Since sensitivity is a related construct to mind-mindedness, we expect that appropriate mind-mindedness will also increase across early infancy (3 to 7 months), and non-attuned mind-mindedness will decrease.

2. We expect parents of secure infants to be more sensitive and more appropriately mind-minded toward infant signals than parents of insecure infants, though we do not have a priori expectations of how mind-mindedness may manifest differentially with infants rated insecure-avoidant versus insecure-resistant.

3. We expect that parents who use more non-attuned mind-related comments will have infants later classified as insecure, though differences between avoidant and resistant attachment are not specified (Meins et al., 2018).

4. Similar to previous research, we expect infants higher in negative affect adopt an insecure-resistant attachment with parents, though this may differ depending on parent gender (Braungart-Rieker et al., 2001, 2014).

Studies that have included infant-father dyads were limited in sample size: Lundy (2003) examined 24 pairs, and Arnott and Meins (2007) were limited to 15 pairs of infant-father dyads. Thus, we examine parent differences in mind-mindedness during the time in which infants are beginning to develop their internal working models of attachment with caregivers across multiple time points (at 3-, 5-, and 7-months). We include a measure of parent sensitivity and infant positive and negative affect during the SFP to identify unique influences on mind-mindedness and infant-parent attachment. To our knowledge, this is the first study to include fathers, infant affect, and measures of mind-mindedness across time.

Data are drawn from a longitudinal study with multiple measures of sensitivity, affect and mind-mindedness across time for mother and father-infant dyads; attachment was measured at only one time point. In addition, we score sensitivity and mind-mindedness across two portions of an interactional episode, adding a third layer of dyadic complexity to our models. Analyses use Mixed Effects Multilevel Models which account for the nested nature of longitudinal, family-wide, dyadic data. These models also allow the inclusion of continuous and fixed variables (i.e., attachment and parent with whom the infant is interacting are fixed; sensitivity, mind-mindedness and infant affect are continuous). Finally, we are able to statistically infer differences between mothers’ and fathers’ behaviors in the same family. Such differences are often assumed but not formally tested.

2. Method

2.1. Participants

Families (N = 135) attended six laboratory visits when infants were 3-, 5-, 7-, 12-, 14-, and 20-months old (+/− 14 days). Data from the first five waves were used in the present study. We recruited participants at local hospitals, birth classes, and informational booths at community events. All parents completed an informed consent and study procedures were approved by the home University’s Institutional Review Board. In the full sample, 52.6% (n = 71) infants were girls, and parents were mostly Caucasian (90.4% mothers, 87.4% fathers) and middle class (average income $45,000 - $59,999). Most parents in the sample (59.3% mothers, 53.7% fathers) had completed some college, with an additional 20% of mothers and 2.1% of fathers having earned a post-graduate degree. Parent age at the initial visit ranged from 17 - 44 for mothers (M = 29.34, SD = 5.32) and 18 - 44 for fathers (M = 3.79, SD = 5.62).

2.2. Procedures

2.2.1. 3-, 5-, and 7-month visits.

During the 3, 5, and 7-month laboratory visits, parents individually participated in the Still-Face Paradigm (SFP; Tronick et al., 1978) with their infants. Parent order was counterbalanced to control for order effects. Once the SFP was complete with one parent, the infant could return to a neutral state and the second parent completed the same procedure. Parents were given instructions both verbally and on an index card they had with them in the laboratory. The SFP was divided into three 90 second episodes. The parent was first instructed to sit down facing their infant, who was sitting in a booster seat on a table, and interact normally. During this “play” episode, parents could talk to, sing with, or touch the infant. Next, at the sound of a bell parents were instructed to stop interacting and display a flat or blank face during the “still-face” episode. For the last “reunion” episode, parents were instructed to resume interacting with their infant as they did during the “play” episode. The current study uses only behaviors coded during the play and reunion episodes of the SFP when parents and infants interacted. Of note, the same coders did not score parent sensitivity and mind-mindedness as coding took place over several years; we believe this enhances the independence of the two measures.

2.2.2. 12- and 14-month visits.

Attachment security was measured with the Strange Situation at 12 months with mothers and 14 months with fathers (Ainsworth & Wittig, 1969). Because parent order effects are not apparent when Strange Situation laboratory visits have at least a 4-week separation in between, parent order was not counterbalanced (Belsky, Rovine, & Taylor, 1984). The Strange Situation involves eight 3-minute episodes designed to elicit attachment behaviors in infants through a series of separations and reunions with a caregiver. From these segments, proximity and contact seeking and avoidance, contact maintenance, and resistance to comforting are coded. Based on behaviors in each category, infants can be classified as insecure-avoidant (A), secure (B), insecure-resistant (C) (Ainsworth et al., 1978), or disorganized (D; Main & Solomon, 1986).

2.3. Measures

2.3.1. Parents’ Mind-mindedness during the SFP.

Similar to previous work, parents’ mind-mindedness was scored during the parent-child interaction episodes of the SFP (McMahon & Newey, 2018). Coding took place in three steps: first, all videos were transcribed for parents’ speech with their infants; second, we coded parents’ use of mind related comments from the transcriptions; third, we re-watched the videos to determine whether parents’ language was appropriate or not, given the context of the verbal exchange.

Parent language was segmented by transcribing individual exchanges consisting of a single word, phrase, or set of words between pauses. The total number or phrases, or utterances, was recorded for the combined pay and reunion episodes of the SFP. Notably, the total number of utterances was, upon visual examination, similar to those found in McMahon and Newey (2018). At 3 months of age, mothers used, on average 96 utterances (range 38-153); at 5 and 7 months of age, mothers used 95 and 92 utterances, respectively (ranges 51-151 and 31-157). Fathers used an average of 87 utterances at 3 months (range 2-148), and 86 and 87 utterances at 5 and 7 months, respectively (ranges 25-152 and 32-143).

Coders used the transcripts and videos to code parents’ mind-mindedness relating to infant mental states. Of note, the infant’s mental state could refer to emotionally valenced (positive and negative) or physiological states. For example, if the parent reflected on the infant’s mental state by using words referring to positive affect, affection, or pleasure such as “funny, love, interested, or happy”; if the parent reflected on the infant’s mental state by using words referring to negative affect, dislike, or displeasure, such as “angry or yucky”; finally, if the parent used language reflecting on their infant’s physiological state of arousal, hunger, fatigue, or illness, such as “tired or hungry.”¹ Some exchanges did not fall into any of these three categories. For example, “look at you sitting there” does not reflect on the infant’s mental state so was not scored. “You’re so happy sitting there” does relate to the infant’s mental state and was scored. Coders were trained by a gold standard coder until κ = .80. The gold standard coder double scored 20% of the transcriptions, on which reliability was analyzed. Interclass correlations (ICCs) of parents’ mind related language at each time point were calculated for mothers (M = .92, range = .75 – 1.00 across coders) and fathers (M = .94, range = .69 - 1.00 across coders).

Based on a mind-mindedness scoring system used during direct observation of parent-infant interactions (Meins & Fernyhough, 2010), appropriate and non-attuned mind-mindedness coding was completed on all video segments in which parents uses mind-related language. Of note, not all parents used such language: at 3 months of age, 10 (7%) mothers and 26 (19%) fathers did not use mind-related language, at 5 months of age, 17 (13%) mothers and 24 (18%) fathers did not use mind-related language, and at 7 months of age, 21 (17%) mothers and 33 (26%) fathers did not use mind-related language. Coders watched the videos of the SFP in conjunction with reading the transcripts and scored the accuracy of the parents’ comments. When parents’ mind-related comments were well-timed, consistent and relevant to the infants’ observed mental state, they were considered to be more appropriate to the infants’ needs. The extent to which a parent’s responses were poorly timed, inconsistent, not relevant to the infant’s state or ambiguous indicated non-attuned mind-mindedness. Again, coders were trained by a gold standard coder, who also double coded 20% of videos to calculate reliabilities. ICCs for appropriate and non-attuned mind-minded comments ranged from .94-,99 (M = .98) and .76-,99 (M = .88), respectively, for mothers and from .93-1.00 (M = .98) and .93-1.00 (M = .98), respectively, for fathers. Because not all parents used mind-related language, scores were not normally distributed; all variables were log-transformed to account for non-normality in the data across parent, age, and SFP episode. After transformation, appropriate mind-mindedness approximated a normal curve, and non-attuned mind-mindedness was more normally distributed than the raw scores, though still slightly kurtotic.

2.3.2. Parent Sensitivity.

Parent sensitivity was coded during the parent-child interaction episodes of the SFP at 3-, 5-, and 7-months of age. Based on Mary Ainsworth’s sensitivity scoring (Ainsworth et al., 1978), sensitivity was defined as the parent responding appropriately to the infant’s state and the ability to make the proper modifications to their own behavior as the infant’s state changes. Examples of sensitivity include but are not limited to: physically putting body or face at child’s level, following the infant’s, responding to the infant’s need to be stimulated or to stop stimulation, and accurately identifying an infant’s source of distress and adjusting behavior to soothe accordingly. Intrusiveness was defined as the parent displaying aggressiveness by following his or her own agenda instead of responding to the infant. Examples of intrusiveness include but are not limited to: forcing own agenda on infant, incongruent parent/infant affect, continuing to respond in same ineffective manner when other responses are available, teasing or overstimulating the infant beyond her enjoyment.

Sensitivity and intrusiveness were coded separately for mothers and fathers and for each episode (play & reunion) using five-point Likert scales (1 = none, 2 = low, 3 = some, 4 = mostly, 5 = high) every 10 seconds. Inter-rater reliability was assessed on approximately 25% of the infants using gold standard coders. Intraclass correlations (ICCs) were obtained during the play and reunion episodes across all three ages for mothers (sensitivity: M = .94, range = .88 - .96; intrusiveness: M = .93, range = .88-.96) and for fathers (sensitivity: M = .92, range = .90-.95; intrusiveness: M = .91, range = .84-.95). Sensitivity and intrusiveness (reverse scored so that high scores indicate low intrusiveness) were highly correlated at each age and for both parents (3-months: n = 132, r = .77, p<001; 5-months: n = 126, r = .82, p<001; 7-months: n = 122, r = .73 , p<.001 for mothers; 3-months: n = 132, r = .69, p<.001; 5-months: n = 124, r = .80, p<.001; 7-months: n = 120, r = .70, p<.001 for fathers). Thus, we averaged sensitivity and intrusiveness to create composite variables for parent sensitivity for mothers and fathers with higher scores indicating higher levels of sensitivity and lower levels of intrusiveness.

2.3.4. Infant Affect.

Infant affect was rated on a second-by-second basis during the play and reunion episodes of the SFP (Braungart-Rieker, Garwood, Powers, & Notaro, 1998). Coding involved a 7-point Likert scale, ranging from −3 (e.g. extreme negative affect) to 3 (e.g. extreme positive affect). Examples of negative affect include screaming, crying, and frowning, and examples of positive affect include laughing and smiling. Infant affect was coded separately for each mother and father SFP, and coders were not allowed to code both parents within one age. Gold standard coders recoded a random sample of approximately 25% of tapes to check reliability; ICCs ranged from .75 to .95 (M = .85) for all waves. We created two separate affect scores that indicated the proportion of intervals positive affect (scores of 1, 2, or 3) and proportion showing negative affect (absolute value of −1, −2, −3). Thus, higher scores indicate more positive and higher negative affect in infants.

2.3.5. Infant-Parent Attachment.

Videos of the Strange Situation with infant-mother dyads at 12 months and infant-father dyads at 14 months were coded by a two-person coding team led by Dr. Elizabeth Carlson at the University of Minnesota. Inter-rater reliability was assessed on 16% of the infant-mother and 17% of the infant-father videos, yielding a 90% agreement with a Cohen’s kappa = .84 for infant-mother dyads, and an 80% agreement with a Cohen’s kappa = .71 for infant-father dyads. Infants were classified secure (B), insecure-avoidant (A), insecure-resistant (C), or disorganized (D) with a secondary B-A-C classification. Infant classifications of D/A, D/B and D/C were included in the B, A, C groupings to increase sample sizes of each group (Burgess, Marshall, Rubin, & Fox, 2003). Two infant-father classifications were D/A/C and thus excluded from the analyses. This resulted in a final sample size of 124 infant-mother and 115 infant-father dyads. See Table 1 for sample sizes and attachment classification distributions for mothers and fathers.

Table 1.

Distribution of infants classified as secure, insecure-avoidant, insecure-resistant, and disorganized with mothers and fathers

Classification	Mothers (n = 124)		Fathers (n = 115)
	N	%	N	%
Insecure-Avoidant (A)	12	9.84	11	9.65
Primary classification	5	4.10	8	7.18
Disorganized/Avoidant (D/A)	7	5.74	3	2.63
Secure (B)	96	77.42	92	80.00
Primary classification	90	72.58	81	70.43
Disorganized/Secure (D/B)	6	4.84	11	9.57
Insecure-Ambivalent (C)	16	12.90	12	10.53
Primary classification	10	8.06	7	6.14
Disorganized/Ambivalent (D/C)	6	4.84	5	4.39

3. Results

3.1. Descriptive Statistics

Attrition analyses showed that of the original 135 families, 130 returned for the 5-month visit, 125 returned for the 7-month visit, 124 mothers and infants returned for the 12-month visit, and 117 fathers and infants returned for the 14-month visit. Statistical comparisons showed that, compared to the full sample (n = 135), the sample used in analyses (n = 124 for mothers, n = 117 for fathers) was older, had higher education levels, and had higher incomes. In addition, several demographic variables related to parent sensitivity, mind-mindedness and infant affect, though not in consistent patterns (see Table 3). Therefore, a risk variable was calculated based on these demographic factors and included as a covariate in further analyses. All scores were z-transformed then averaged to create a composite risk score in which higher values indicated a higher risk (parents were younger, less educated, and reported a lower income).

Table 3.

Demographic relations with variables of interest

3-months	Mother Age (r)	Father Age (r)	Mother Education (r)	Father Education (r)	Family Income with mothers (r)	Family Income with fathers (r)
1. Parent Sensitivity	.06	−.04	.24**	.12	.15	.09
2. Infant Positive Affect	.12	.00	.06	.01	.14	.23*
3. Infant Negativity	.00	.17	−.03	−.15	.02	−.09
4. Appropriate Mind-mindedness	.09	−.07	.03	.00	.09	.25**
5. Non-attuned Mind-mindedness	−.15	−.12	−.20*	−.24**	−.05	−.02
5-months
1. Parent Sensitivity	.25**	.26*	.15	.15	.13	.11
2. Infant Positive Affect	.13	.04	.10	−.05	.10	.02
3. Infant Negativity	.02	−.06	.07	−.02	.08	.09
4. Appropriate Mind-mindedness	−.07	.16	−.03	.04	−.02	.12
5. Non-attuned Mind-mindedness	−.02	−.07	−.05	−.04	.04	.01
7-months
1. Parent Sensitivity	.12	.15	.07	.22*	.18	.06
2. Infant Positive Affect	.07	−.06	−.03	−.02	.11	.11
3. Infant Negativity	−.04	−.17*	.02	−.10	−.08	−.03
4. Appropriate Mind-mindedness	.09	.01	.14	.01	.15	.01
5. Non-attuned Mind-mindedness	−.03	−.09	.02	−.08	−.01	.04

Note:

^***p < .001,

^**p < .01,

^*p < .05; p-values are two-tailed; values are averaged across the play and reunion episodes of the SFP. Relations for mother demographic variables only within mother behaviors; relations for father demographic variables only within father behaviors; family income for mother and father both

Means and standard deviations of parents’ levels of sensitivity and use of mind-related comments are in Table 2. In addition, correlations between parent sensitivity and mind-mindedness within each age for each parent are presented in Table 3. Simple relations were somewhat inconsistent, and parent sensitivity was related to mind-mindedness in different ways for mothers and fathers. Specifically, at 3 months, mother sensitivity was related to more appropriate mind-mindedness and at 5 months to higher non-attuned mind mindedness. Father sensitivity at 3 months and mother sensitivity at 7 months was related to less non-attunement. For both parents, use of appropriate and non-attuned comments were related across age.

Table 2.

Means and Standard Deviations for parent sensitivity, infant affect and mind-mindedness by parent and age

	Mother			Father
	3-months	5-months	7-months	3-months	5-months	7-months
Sample Size	n=132	n=127	n=122	n=132	n=124	n=121
Parent Sensitivity	4.21 (.58)	4.21 (.53)	4.38 (.46)	4.08 (.59)	4.10 (.58)	4.17 (.51)
Infant Positive Affect	.11 (.11)	.20 (.19)	.20 (.17)	.09 (.10)	.14 (.15)	.17 (.17)
Infant Negative Affect	.11 (.24)	.05 (.11)	.10 (.19)	.11 (.23)	.07 (.17)	.07 (.14)
Number of Appropriate Mind-Minded Comments	2.49 (2.10)	1.94 (1.70)	1.31 (1.41)	1.86 (2.17)	1.76 (2.05)	1.08 (1.15)
Number of Non-attuned Mind-Minded Comments	.53 (.94)	.23 (.57)	.22 (.47)	.42 (.98)	.27 (.84)	.36 (1.13)
Total Number of Utterances	48.01 (9.93)	47.33 (10.22)	45.80 (11.09)	43.39 (15.25)	42.76 (12.93)	43.66 (12.88)

Note: Values are averaged across the play and reunion episodes of the SFP.

In our data, there were no infant gender differences relating to parents’ mind-mindedness or sensitivity toward their infant (of 36 correlations examined, only 1 approached significance, such that mothers were slightly more sensitive toward male infants during reunion at 7 months, p = .05). There were also no differences in attachment grouping across gender (χ²(3) = 6.66, p = .08 for infant-mother attachment and χ²(3) = 1.87, p = .60 for infant-father attachment). Therefore, we did not include infant gender as a covariate in analyses.

3.2. Mixed Effects Models Comparing Attachment Groups for Mothers and Fathers

We used SAS PROC MIXED to fit mixed models (Singer & Willett, 2003) relating trajectories of parent sensitivity and infant affect to parent mind-mindedness across attachment groups. Figure 1 presents a conceptual model depicting the mixed effects models conducted. Mixed effects models can be used to study change over time with repeated-measures data and allows for the inclusion of both continuous (sensitivity and infant affect) and categorical (attachment classification) covariates (Boyle & Willms, 2001). All analyses used maximum likelihood estimation to account for missing data (Enders, 2010). In addition, we used data from mothers, fathers, and infants within the same family in one overall model (one each for appropriate and non-attuned behaviors) (Kenny, Kashy, Cook, & Simpson, 2006; Planalp, Du, Braungart-Rieker, & Wang, 2016). To aid in the interpretation of results, we used models which controlled for parent order during the SFP at each age (mother first or father first) and SFP episode (play or reunion). Notably, these variables were not significant determinants of parent-mind-mindedness.

Figure 1.

Conceptual model of mixed effects model predicting parent mind-mindedness

Because attachment was assessed at only one age per infant-parent dyad, it was treated as a fixed effect. We used dummy coded variables to be able to statistically infer differences between attachment groups; comparing secure (B) infants to avoidant (A) and resistant (C) infants within each parent, but not across parents. That is, when talking about mother behaviors, we only examined mother-infant attachment classification and did not examine patterns with infant-father attachment (and vice versa with infant-father data).

Attachment group differences in parents’ mind-mindedness, in relation to infant age, parents’ sensitivity, and infant affect are in Table 5. Estimates reflect the differential effect of each variable (age, parent sensitivity, infant positive and negative affect) on mind-mindedness by attachment category (B, A, C). In addition, parent was coded such that we could statistically infer differences between mothers’ and fathers’ behaviors within a family.

Table 5.

Standard Estimates for Sensitivity, Infant Affect, and Parent Mind-mindedness by Attachment Grouping

		Infant-mother attachment
		Secure		Insecure-Avoidant		Insecure-Resistant
		Est (s.e.)		Est (s.e.)		Est (s.e.)
Appropriate Mind-mindedness (average)		1.26***	(.09)	1.57***	(.18)	1.27***	(.16)
Effects for:	Age	−.17***	(.03)	−.27***	(.07)	−.17*	(.07)
	Sensitivity	.05	(.05) a	.27*	(.11) b	.04	(.10)
	Infant Positivity	.11	(.15)	.35	(.38)	−.25	(.39)
	Infant Negativity	.52***	(.13)	.48	(.35)	.35	(.28)
Non-attuned Mind-mindedness (average)		.31***	(.05)	.39**	(.11)	.35***	(.09)
Effects for:	Age	−.06**	(.02)	−.07	(.04)	−.06	(.04)
	Sensitivity	.04	(.03)a	.09	(.06)	.15*	(.06)b
	Infant Positivity	−.11	(.09)	−.13	(.22)	.02	(.23)
	Infant Negativity	.38***	(.08)a	.38	(.20)	.70***	(.17)b
		Infant-father attachment
		Secure		Insecure-Avoidant		Insecure-Resistant
		Est (s.e.)		Est (s.e.)		Est (s.e.)
Appropriate Mind-mindedness (average)		.99***	(.09)	1.29***	(.18)	1.00***	(.16)
Effects for:	Age	−.10**	(.03)	−.19*	(.08)	−.10	(.07)
	Sensitivity	.10	(.05) a	.32**	(.11) b	.08	(.10)
	Infant Positivity	−.27	(.19)	−.02	(.38)	−.64	(.37)
	Infant Negativity	.42***	(.14)	.39	(.33)	.25	(.30)
Non-attuned Mind-mindedness (average)		.17**	(.05)	.24*	(.11)	.20*	(.09)
Effects for:	Age	.01	(.02)	−.01	(.05)	.01	(.04)
	Sensitivity	.02	(.03)	.06	(.06)	.12*	(.06)
	Infant Positivity	−.28*	(.11)	−.30	(.22)	−.14	(.22)
	Infant Negativity	.24***	(.08)a	.25	(.19)	.57**	(.18)b

Note: All analyses control for demographic risk, parent order, SFP episode, and total number of utterances (verbosity). Within a column, * reflects significance of effects of age, sensitivity, and infant affect on average levels of parent mind-mindedness within each attachment group, with *p<.05, **p<.01, ***p<.001. Mother vs. father differences are bolded. Superscripts (^ab) represent differences across attachment group. Differences between mothers and fathers are explained in text.

3.2.1. Differences in mother and father mind-mindedness.

As indicated by the bolded estimates for mothers’ and fathers’ average mind-mindedness variables, mothers used significantly more appropriate and non-attuned mind-mindedness than fathers. Both parents decreased in appropriate mind-mindedness at similar rates (no mother vs. father differences), but mothers decreased in non-attunement whereas fathers used the same number of non-attuned phrases across time, regardless of attachment classification (all mother vs. father comparisons significant at p<.01 for average levels and the effect of age). There were no significant mother vs. father differences for trajectories of parent sensitivity or infant affect.

3.2.2. Parent Sensitivity and Infant Affect.

Positive relations between parent sensitivity and mind-mindedness indicate that for infants later classified as insecure-avoidant, parents were less appropriate when they were also rated lower in sensitivity (Est. = .27 for mothers and .32 for fathers). This was significantly different than secure infants, as indicated by subscripts in Table 4. For infants later classified insecure-resistant, parents used less non-attuned mind-mindedness when parents were less sensitive (Est. = .15 for mothers and .12 for fathers).

Table 4.

Mother and father within age correlations relating parent sensitivity, infant affect, and mind-mindedness

3-months	1.	2.	3.	4.	5.
1. Parent Sensitivity	.08	.31***	−.24**	.16	−.23*
2. Infant Positive Affect	.10	.21*	−.29***	−.17	−.13
3. Infant Negativity	−.03	−.33***	.17	.07	.19*
4. Appropriate Mind-mindedness	.24**	.21*	.14	.23*	.28***
5. Non-attuned Mind-mindedness	.10	−.11	.30***	.27**	.06
5-months	1.	2.	3.	4.	5.
1. Parent Sensitivity	.16	.18*	−.10	.18	.06
2. Infant Positive Affect	.26**	.30***	−.17	.00	−.09
3. Infant Negativity	.02	−.13	.05	.20*	.52***
4. Appropriate Mind-mindedness	−.01	−.04	.18*	.06	.21*
5. Non-attuned Mind-mindedness	.23*	.02	.32***	.18*	.06
7-months	1.	2.	3.	4.	5.
1. Parent Sensitivity	.05	.20*	−.23*	.08	−.04
2. Infant Positive Affect	.17	.48***	−.21*	−.22*	−.08
3. Infant Negativity	−.41***	−.26**	.26**	.20*	.10
4. Appropriate Mind-mindedness	−.12	−.05	.24*	.03	.14
5. Non-attuned Mind-mindedness	−.26**	−.10	.25**	.28**	.12

Note:

^***p < .001,

^**p < .01,

^*p < .05; p-values are two-tailed; mother correlations are below the diagonal and father correlations are above the diagonal; correlations average across SFP episode; mother/father correlations within each variable are bolded and on the diagonal.

In terms of infant affect, fathers of infants later rated secure used less non-attuned mind-mindedness when infants were more positive (Est. = −.28). Parents used more appropriate and non-attuned mind-mindedness when infants were rated more negative, with varying levels of statistical significance depending on attachment group (see Table 5). Attachment group comparisons revealed stronger differences between secure and insecure-resistant infants: parents were more appropriate to secure than insecure-resistant infants rated higher in negativity (Est. = .38 vs. .70 for mothers, p<.05 for secure vs. resistant infants; Est. = .24 and .57 for fathers, p<.05 for secure vs. resistant infants). Full results are in Table 5.

4. Discussion

Early infancy, between two to seven months when infants are forming their internal working models of expectations for caregiver relationships and felt security, is a particularly salient time in which to study potential processes relating to later attachment (Bowlby, 1969). By relating a parent’s mind-mindedness during this time with infant-parent attachment, we can examine if a parents’ ability to mentalize their infants’ individual needs is indicative of later attachment. Results from the current study suggest that parents differ in their use of mind-related comments with infants, and that these are differentially related to parent sensitivity, infant affect, and later attachment classification.

4.1. Parent Differences in Mind-mindedness

Similar to previous work (Arnott & Meins, 2007; Lundy, 2003), mothers were both more appropriate and less non-attuned to their infants’ needs than were fathers, even after controlling for verbosity. Mothers’ may be more verbally responsive toward their infant than fathers, or mothers’ interactions with their infants may be more reliant on mother-initiated action, while fathers wait to respond to infant bids. In addition, mothers tend to be more involved with their young infants overall, yet fathers engage in more play behaviors (Jones & Mosher, 2013; Parke, 2000). At this young age, fathers may not be as accustomed to interacting with their 3-7-month olds without books, toys, or other props, either during this laboratory-based assessment or potentially in the home. This may make the SFP a more challenging situation for fathers than mothers. In addition, previous work with these data (Planalp et al., 2013) indicates that mothers are slightly more sensitive than fathers toward their infants during the SFP context. Thus, differences in mind-mindedness may be context specific, or may be unique to the parent. Future research could measure sensitivity and mind-mindedness during separate paradigms to disentangle these complex effects.

Further, mothers, but not fathers, decreased in mind-mindedness across early infancy (3 to 7 months). This is perhaps unexpected because previous work with this sample indicates slight increases in sensitivity, a related but distinct construct. Mothers may be becoming more behaviorally sensitivity, but may “talk” for their infants less as the infant is more able to do so for themselves. By 7 months, infants are using more self-directed communication than they are at 3 months (Tronick, 1989). In addition, infants at three and seven months of age would presumably react differently to the emotional unavailability of a parent. For example, at three months, infants turn their gaze toward the parent during the still-face more so than at seven months; whereas at seven months infants use more distraction to regulate their arousal during the still-face (Planalp & Braungart-Rieker, 2015). In response to infants’ gaze directed behavior, parents may not feel the need to respond as infants get better at regulating their affective response themselves. Fathers, however, did not change in non-attuned mind-mindedness across early infancy. Fathers also tend to increase their level of involvement with infants as the infant ages (Planalp & Braungart-Rieker, 2016). It may be that as infants get older, fathers play more with their infants which, in turn, enables fathers to learn more about their infants’ signals and ultimately make more attributions about them. Future studies might examine parenting roles and experiences to more carefully examine the developing infant-mother and infant-father attachment relationship.

4.2. Parent Sensitivity

Increased parent sensitivity is associated with secure infant-parent attachment with differing effect sizes for mother and father dyads (de Wolff & van IJzendoorn, 1997; Zeegers et al., 2017). Further, in general parents who are more sensitive are also more appropriate to their infants’ needs (Laranjo et al., 2008; Meins et al., 2012). Our results add to this complicated pattern of mother and father differences by finding that sensitivity and mind-mindedness relate to insecure attachment in varying patterns.

4.2.1. Insecure-Avoidant Attachment and Parent Sensitivity.

For infants later classified as insecure-avoidant, parents used more appropriate mind-mindedness when they were also rated higher in sensitivity. This may seem counterintuitive, but it is possible that a parent of an insecure-avoidant infant might accurately interpret the baby’s distress but not act appropriately to that signal. Alternatively, because parents of insecure infants are generally less sensitive, reverse of this relation indicates that parents were actually less appropriately mind-minded when they were less sensitive. Parents were not necessarily less non-attuned, however, when they were rated higher in sensitivity. This suggests that parents of insecure-avoidant infants who are less sensitive may not be showing much mind-mindedness, in general. That is, they appear to be reflecting less about their infants’ signals. Parents of insecure-avoidant infants are more dismissive or have a difficult time relating to the infant (Bretherton & Munholland, 2008). It is difficult to discern whether infants who become avoidant simply show fewer signals necessitating parent response or whether parents’ limit their comments. Moreover, it is not entirely clear how sensitivity and attunement are linked for parents of insecure avoidant infants. It is possible that lower sensitivity is the result of being less reflective; alternatively, it is possible that lower attunement is a result of engaging in fewer sensitive behaviors. Nevertheless, results suggest that both mind-mindedness and sensitivity are areas that should be targeted for improvement in parent-infant relationships.

4.3. Infant Affect

Overall, parents used more mind-mindedness (appropriate and non-attuned) when infants were rated more negative. These infants may make bids for their parents’ attention more, or parents may “talk” for infants who express their feelings more strongly. Parents appear to respond by talking more for the infant as a way to solve the issue. Although the comments are not always in-tune with infants’ signals, it shows that parents are at least making attempts to determine what infants are “telling” them.

4.3.1. Insecure-Resistant Attachment and Infant Affect.

We found that when infants were later classified secure, parents used more mind-mindedness overall; yet when infant were classified insecure-resistant, parents used more non-attuned mind-mindedness. Interestingly, for infants rated higher in negativity, when parents were more appropriate this was more strongly related to a secure attachment; when parents use more non-attuned mindedness, this was more strongly related to insecure-resistant attachment as might be expected. Insecure-resistant infants may simply be more distressed or express more negative affect (Braungart-Rieker et al., 2001; Leerkes & Wong, 2012) or fussy. Mothers who are not able to correctly read their infants’ signals perhaps are less adept at soothing a fussy baby, resulting in an insecure-resistant attachment. Alternatively, that parents may be incorrectly interpreting their infants’ needs, resulting in an insecure-resistant attachment. Such parents exhibit inconsistent parenting (Cassidy & Berlin, 1994). Though these parents exhibit more sensitivity during a laboratory task, they are not as adept at interpreting their infants’ needs, leading to an insecure-resistant attachment. Nonetheless, it is possible that for infants who are more negative, parent attunement creates a buffer between infant negativity and infant-parent attachment.

The only other paper to use the SFP as an experimental condition with which to measure mind-mindedness did not find that infant negativity was related to attunement, though the authors admit this was unexpected (McMahon & Newey, 2018). Our differing results could be due to the fact that we split our sample by attachment group, resulting in smaller group sizes, but also potentially different patterns of infant negativity (Borelli et al., 2010; Braungart-Rieker et al., 2001). In addition, we used the SFP at three ages beginning at 3 months and ending at 7 months of age. McMahon and Newey only looked at 7 month old infant-mother dyads. The younger age studied and fact that we found decreases in mind-mindedness over time may account for different results. Further work using the SFP to measure mind-mindedness could clarify such discrepancies.

4.4. Father-Infant Dyads

Some findings were specific to the father-infant dyad. Fathers of infants later rated secure showed less non-attuned mind-mindedness when infants were more positive. Infants use smiling and positivity as a form of social communication by which to inform caregivers of feeling states, needs, or desires (Ramsey & Gentzler, 2015; Sroufe & Waters, 1976). It is possible that infants rated higher in positivity communicate their needs to parents more efficiently. Alternatively, fathers engage with temperamentally easy infants more so than difficult infants (Brown, McBride, Bost, & Shin, 2011; Volling & Belsky, 1991), so it is also possible that fathers are more accustomed to interacting with their more positive children at home, which also plays out during this laboratory task.

In general, there were attachment group differences in parents’ mind-mindedness with their children during early infancy. These differences may reflect that mothers’ and fathers’ roles in the home differ; the contexts in which parents and infants interact may elicit different patterns of reflection from mothers versus fathers. Mothers may more typically engage with the infant directly to soothe infant distress and fathers may more typically act as social playmates with their infants (Parke, 2010). As infant affect was included in our analyses, these differences cannot be attributed to differential infant affective responses toward mothers and fathers.

Though Arnott and Meins (2007) did not find significant associations between non-attuned comments and infant-father attachment, our findings indicate that complex relations between mind-mindedness, sensitivity and infant affect impact the developing attachment relationship. Future research would benefit from a thorough examination of infant cues in relation to sensitivity and parent language use to identify the multiple influences at work in the developing attachment relationship.

4.5. Limitations and Conclusions

The current study adds to the literature by illustrating the longitudinal shift in parent mind-mindedness in relation to the developing attachment relationship, yet we were limited by several factors. First, results are generalizable to a mostly middle class, educated, Caucasian population. Parents’ ability to be mind-minded toward their infants may differ in populations with a lower socioeconomic or educational status. In addition, we did not examine disorganized attachment because the number of infants in our avoidant and resistant categories was fairly small. To maximize group size, we used infants’ secondary classifications of A or C and not their primary D classification. Parents’ of disorganized infants can be inconsistent or even frightening toward their infant, thus it is possible that relations exist between parent mind-mindedness and disorganized attachment that were not detected in the current sample due to our small group sizes. In addition, maternal stress (Demers, Bernier, Tarabulsy, & Provost, 2010; McMahon & Meins, 2012) and psychological factors such as depression (Lundy, 2003) relate to maternal mind-mindedness and marital conflict relates to paternal mind-mindedness (Lundy, 2003). The current study did not assess these psychosocial factors but future work would benefit from a comprehensive examination of the developing attachment relationship, as it is clear that factors other than behavioral sensitivity are related.

Previous research on mind-mindedness calculates proportion scores for attunement to account for parent verbosity and total length of parent-infant interaction (McMahon & Meins, 2012; McMahon & Newey, 2018; Meins et al., 2018) We control for the total number of utterances parents use with their infants during each episode of the SFP to account for parent verbosity, but still use a count or number of appropriate and non-attuned phrases in analyses. It is possible that even one instance of non-attunement may adversely impact the parent-infant relationship. If a parent is very verbose but uses only one non-attuned phrase, this come through in data analyses as a very low proportion score. If a parent uses very few verbal phrases and one is non-attuned, this comes through as a high proportion. Thus, though we control for verbosity, we did not calculate proportion scores.

Further, we scored parent and infant behaviors from the same paradigm (the Still-Face) in infancy and related to later attachment. We believe that this is both a limitation and strength of our study. Previous reports relating mind-mindedness to sensitivity and attachment find different patterns of relation depending on measurement (McMahon & Bernier, 2017). While it is true that the SFP provides only a limited window through which to examine the parent-infant relationship, the SFP is an effective tool to measure parent-infant relationships using an ecologically sound task (Mesman et al., 2009; Tronick et al., 1978) and has been used in mind-mindedness research as well (McMahon & Newey, 2018). Though we coded many behaviors from the same SFP, we used strictly observational measures of infant-parent interactions so our findings were not confounded by the shared method variance common of multiple parent reports. Coding took place over several years and coding systems were independent of one another; for example, the same people who scored parent sensitivity did not also score mind-related comments, nor did coders score the same family within a given age. Therefore, we believe the information derived from one SFP across three time points perhaps best reflects typical parent-infant interactions and we have confidence that our findings reflect independent processes for infant-mother and infant-father sensitivity and mind-mindedness with their infants.

Our analyses used complex multi-level, multi-group models examining mind-mindedness, sensitivity, and infant affect for mother and father-infant dyads across time; we did not test interactions among all variables nor relations with parents’ valenced responses to infant positive and negative affect. In this same sample (Braungart-Rieker et al., 2014), analyses of infant affect and parent sensitivity as predictors of infant mother and infant-father attachment indicate that sensitivity interacts with infant affect to predict infant-parent attachment. Though of potential interest, our small sample size and already complex longitudinal models preclude us from fully understanding these processes in relation to mind-mindedness. Future work should attempt to extricate infant affect and parent sensitivity to better understand the dynamics of sensitivity and mind-mindedness during times of infant distress across infancy.

Fathers are increasingly becoming more active in childcare (i.e., feeding, bathing), which provides contexts from which fathers’ can learn to correctly attune to their infants’ signals; assessing mind-mindedness in both mothers and fathers may be particularly important in future research. Though previous work has found relations between parent mind-mindedness and infant attachment, we add to existing literature by examining differential relations of attunement with both mothers and fathers in longitudinal analyses. Incorporating mind-mindedness into the study of attachment could provide a powerful way to understand the development of attachment at multiple levels of analysis, in addition to behavioral sensitivity and infant affective processes. Parents who read their infants’ signals appropriately (mind-mindedness) and respond accordingly (behavioral sensitivity) have infants later classified secure. Sensitivity interventions are aimed at changing a parents’ behavior so that it is more responsive to toward infants (Bakermans-Kranenburg, Van Ijzendoorn, & Juffer, 2003; Broberg, 2000; van IJzendoorn, Juffer, & Duyvesteyn, 1995). The added component of mind-mindedness may help researchers and clinicians better understand what parents think about infant behaviors, how parents interpret infant behaviors, and give tangible, potentially verbal cues, parents can use to respond to infant needs most appropriately. Incorporating mind-mindedness into sensitivity interventions may provide a unique opportunity to train parents on how best to recognize and respond to their infants’ needs, thereby enhancing attachment security.

Highlights.

• We examined the longitudinal shift in parent mind-mindedness in relation to the developing attachment relationship in N = 135 families.

• Significant differences between parent gender and attachment category were detected, such that parents who were more sensitive were also more mind-minded toward insecure-avoidant infants; parents used more non-attuned mind-mindedness when infants had higher negative affect.

• Findings suggest that, in addition to parent sensitivity, parents’ use of appropriate and non-attuned mind-mindedness during a parent-infant interaction provides insight into the developing attachment relationship for mothers and fathers.