“The Relationship Structures Questionnaire (ECR-RS)” is a self-report styles with several different people, for example mother (or a mother-like figure), father. 40 items The authors report a measure, the Relationship Structures questionnaire of Based on a sample of over 21, individuals studied online, it is shown. The authors report a measure, the Relationship Structures questionnaire of the Based on a sample of over 21, individuals studied online, it is shown that.
Completing the forced-choice paragraph first serves as a counterbalancing effect to minimize order effects when participants rank the degree to which each prototype is self-characterizing. If you wish your results to correspond to the anxiety dimension, the calculation can be reversed [i.
In the latter calculation, higher scores will refer to higher anxiety and more negative models of self. If you wish your results to correspond to the avoidance dimension, the calculation can be reversed [i. In the latter calculation, higher scores will refer to higher avoidance and and more negative models of the other. On a 5-point scale, participants rate the extent to which each statement best describes their characteristic style in close relationships.
Five statements contribute to the secure and dismissing attachment patterns and four statements contribute to the fearful and preoccupied attachment patterns.
Scores for each attachment pattern are derived by taking the mean of the four or five items representing each attachment prototype. Additionally, the three dimensions used by Collins and Read can also be obtained. Alternatively, and perhaps preferably, you can use the questionnaire to derive scales of the underlying two dimensions of anxiety and avoidance see below. Like the RQ, the RSQ can be worded in terms of general orientations to close relationships, orientations to romantic relationships, or orientations to a specific adult, peer relationship.
The RSQ is designed as a continuous measure of adult attachment. If, however, it is absolutely necessary for you to classify participants into attachment patterns, you must use standard scores. First, you would create the four subscales by computing the mean rating of the items for each subscale. Then you would transform those mean ratings into standard scores.
This is a far from ideal use of the RSQ and should be undertaken only as a last resort!
Thus, researchers are able to relate the RSQ to alternate self-report measures of adult attachment. This can be done in at least three ways: We recommend the third approach.
First, attachment ratings on both the RQ and the RSQ need to be converted into standard scores z-scores. For example, the now standardized RQ secure scores are combined with the now standardized RSQ secure scores to form a single, composite measure of secure attachment. Apply the same procedure to the remaining attachment pattern ratings. These composite scores can be used in all subsequent analyses.
For an example of this procedure see: Adult attachment styles, perceived social support and coping strategies. Journal of Social and Personal Relationships, Vol. The case of adult attachment. Attachment processes in adulthood pp.
To address these four concerns, we explored a derivation of the Indeed, sometimes they do not assess the same constructs. For ECR-R questionnaire that was designed to assess attachment- example, one of the most commonly used self-report measures of related anxiety and avoidance in four kinds of relationships: The IPPA modified questionnaire, which we refer to as the Relationships is used to score participants on three dimensions: Commonly used self-report measures of romantic attach- items total.
In Study 1, we report some of the basic psychometric ment, however, are used to scale people on attachment-related properties of the ECR-RS and provide descriptive summaries of anxiety and avoidance. Although the scales used in these different the scales based on a sample of over 21, people who vary in inventories have obvious conceptual and empirical commonalities age, sex, ethnicity, and country of residence. An ideal assessment tool would use relationships.
We demonstrate that a the ECR-RS measures of the same kinds of items and scales to assess attachment security in romantic attachment are associated with basic aspects of relation- different relationships, making comparisons across relationships ship functioning e.
For example, the Experiences in Close Relation- reliable as those based on longer inventories e. If a researcher was than those from the ECR-R. We also examine the association interested in assessing security across multiple relationship con- between the differentiation in working models across relational texts, the number of items needed for the assessment would mul- contexts and psychological functioning.
One challenge for investigators sacrificing too much measurement precision see Lo et al. An ideal measure would allow attachment to be as- In fact, this situation has become more problematic in recent years sessed across multiple relationships without placing too much of a because, despite the fact that there is a growing interest in assess- burden in terms of the number of items rated on research partic- ing attachment in a contextual fashion, there is no common method ipants.
Indeed, the studies that have been published in the A final concern is that contemporary measures of attachment do literature to date each use different methods and instruction sets, not allow within-person variation to be assessed across relational different self-report questionnaires, and, in some cases, different contexts.
As several researchers have observed, the correlation conceptual models of individual differences in attachment. We between security assessed in different domains e. This implies that there is some widely used measure i. We hope that models. The items we website, which contains a variety of web studies and demonstra- selected are listed in Table 1.
To examine the factor structure of the The site can be found via web searches for a variety of key words ECR-RS items, we conducted an exploratory factor analysis using relevant to personality and relationships and receives approxi- principal axis factoring, followed by varimax rotation, separately mately to visitors a day although not all visitors partic- for the 10 attachment items for each of the four relational domains.
For an The purpose of this analysis was to determine whether the item in-depth comparison of web-based samples and more traditional responses had a two-factor structure similar to that commonly samples in psychological research, please see Gosling, Vazire, found with other instruments. In each of the four relational domains, at least two exclusive dating or marital relationships. The primary reason for factors were necessary to represent the data. Moreover, because we were interested in assessing Table 1 shows the factor loadings for the items in each domain, attachment across distinct relationships, we wanted to ensure that using a two-factor solution.
In each domain, there were clear we could compare security across specific relational domains e. The items that were designed to mother, father, romantic partner, and friendships in a way that was assess avoidance tended to have moderate to strong loadings on the meaningful to each respondent. The sample was composed pri- first factor. The items that were designed to assess anxiety tended marily of women The average age of participants was to have strong loadings on the second factor.
As such, we or elsewhere. The majority of participants were White Another result of interest is Materials and procedure. Participants rated an initial pool that some of the items had moderate cross-loadings.
Relationship Structures (ECR-RS) Questionnaire
However, of 10 items designed to assess their attachment orientation across because the specific items that had moderate cross-loadings in one four kinds of intimate relationships i. The same 10 items relationship domains, we did not exclude these items from subse- were used for each domain, yielding 40 items total. For each item, quent analyses.
The items were selected in modified avoidance based on Items 1— 6 and anxiety based on Items 7—9 see form from the ECR-R, a widely used attachment inventory de- Table 1. Items 5 and 6 in Table 1 were reverse keyed.
As signed to assess the two fundamental dimensions underlying at- mentioned previously, we excluded Item 10 because of its rela- tachment patterns: The tively high cross-loadings across domains. We focus on these anxiety dimension represents the extent to which people tend to composites in the analyses that follow.
The avoidance dimen- and intercorrelations among the scores are reported in Table 2. Prototypically secure First, the means are relatively low, suggesting that the average people tend to score low on both dimensions.
This result is similar to what is found using longer pool of items—items based on previously existing inventories and more general measures of attachment see Study 2. I usually discuss my problems and concerns with this person. I talk things over with this person. It helps to turn to this person in times of need. I find it easy to depend on this person.
I prefer not to show this person how I feel deep down. It is possible point of the scale and, accordingly, a much more normal distribu- that the specificity that is added by contextualizing the targets tion of scores.
For example, the lowest alpha reported in the attachment dimensions in different domains are positive, but Table 2 is. For example, people who report avoidant longer measures, such as the ECR Brennan et al.
Overall, people reported the greatest amount of similarity measure of attachment, the ECR-R Fraley et al.
Self-Report Attachment Measures - Psychology Members' Site
This finding is similar to that reported by Klohnen et attachment across multiple relational domains. Data from people in dating or marital attachment style. As can be seen in Table 2, the reliabilities of relationships were collected as part of a larger research study in these composite scores were also high.
The sample was composed primarily of women avoidance. The average age of participants was The majority of domains is relatively reliable. For the most part, men and women differed in their anxiety and Materials and procedure. In general, men tended Adult attachment.
Participants rated the items from the to report less anxiety and more avoidance than women across ECR-RS questionnaire, again focusing separately on relationships domains. This is consistent with findings by Schmitt and his with mother, father, partner, and best friend.
Although we col- colleagues e. For Finally, it is noteworthy that the correlations between anxiety the Anxiety scale, the alpha reliabilities for these four domains and avoidance are relatively high in each relational domain.
For the Avoidance scale, who tend to be more avoidant in their relationships with their the alpha reliabilities for the four domains were. The lations are worth mentioning because previous research using alpha for the global anxiety scores was. We also administered the ECR-R the two attachment dimensions are only weakly correlated, around Fraley et al.
We discuss some standard, global fashion. Equal variances are not assumed for the t tests. We are interested in how you and intercorrelations among the various ECR-RS scores are re- generally experience relationships, not just in what is happening in ported in Table 4.
For the most part, these statistics were similar to your current relationship. Table 5 reports the correlations between anxiety and avoidance scores were. First, on average, the the average avoidance score was 2. Second, the two measures tended to be most secure. The correlation between the two ECR-R dimensions strongly correlated in the romantic domain.
This is noteworthy because it Relationship functioning. The IMS assesses commitment, investment, satisfaction in a rela- We next examined the association between these two measures tionship, and the quality of alternatives.
The alpha reliabilities for of attachment and various relational and intrapersonal outcomes. In the present study, participants were asked avoidance across most relational domains was related to relation- to rate these items with respect to their current relationship. However, the associations were much more Intrapersonal functioning. The 9-item version of the Center salient in the romantic domain. As expected, the ECR-R dimensions also correlated with those based on the full-scale version in previous research Kohout these relational variables in similar ways.
Highly anxious and et al. The alpha of the scores in the present sample was. Extraversion, Agreeableness, partner dimensions correlated just as strongly with the relational Conscientiousness, Neuroticism, and Openness to Experience.
The and intrapersonal outcomes as the ECR-R dimensions. The excep- alpha reliabilities for these scores were. The ECR-RS measures of anxiety and avoidance also correlated attachment dimensions is similar to what other investigators have with measures of depressive symptoms.
People who were more reported using different measures of attachment e. Specifically, attachment-related anxiety tends to ship domains tended to report a higher number of depressive correlate with neuroticism and attachment-related avoidance tends symptoms. This was true for the ECR-R measures of anxiety and to correlate negatively with agreeableness. One noteworthy feature avoidance as well. These cor- the domains in which attachment and personality traits tend to relations are reported in Table 6.
In general, the pattern of asso- correlate. In this article, Openness. The logic is as follows: If a person behaves in Alternatives.
In contrast, if CES-D. We then averaged those nine standard deviations to index Partner. The average amount of differentiation observed ECR-R in our sample was 1. Correla- ality traits see the column labeled SD. This is compatible with what As revealed in Table 7, differentiation is related to attachment Baird et al. In other words, people who exhibited more dif- measures of personality differentiation. They found that people ferentiation tended to be more insecure as a general rule i.
Indeed, Baird et al. This confounding can emerge when item eration. To control related to intra- and interpersonal measures of adjustment but that for this problem, Baird et al. Once we controlled item mean and the square of that mean. This adjusted index for overall levels of security, the association between differentia- adjusted standard deviation represents the amount of differenti- tion in working models and adjustment largely disappeared.
Table 7 shows the relationship between General Discussion the adjusted standard deviation and the various variables of interest see the column labeled Adjusted SD. What is noteworthy about A primary goal of this article was to present a self-report method these data is that there do not appear to be any associations that would address four limitations of commonly used attachment between attachment differentiation and measures of inter- and measures.
Once we controlled for insecurity see Baird et al. This finding implies that, at least with respect to scores across relationship domains could be compared on a com- the variables we assessed, there may not be any unique implica- mon metric. Third, we wanted a measure that was relatively short tions of heterogeneity in working models per se. Someone who has so that attachment could be assessed across relationship domains relatively secure representations of his or her romantic partner but without placing an unnecessary burden on clients or research insecure working models with respect to his or her parents is not participants.
Finally, we desired a measure that could be used to going to have less satisfying relationships or experience more study not only security versus insecurity in different relational depression because of the lack of congruity in attachment repre- domains but also the amount of homogeneity versus differentiation sentations across domains. Because a com- mon set of items is used to assess attachment in different domains, Limitations, Caveats, and Conclusions security across contexts can be contrasted and compared in mean- ingful ways.
Although the ECR-RS has the potential to be useful, it would be In Study 1, we demonstrated that the two-dimensional structure premature to claim that it is without limitations.
One noteworthy familiar to contemporary attachment researchers emerged in each limitation of the measure is that the items are not well balanced of the four relational domains of interest. Moreover, the composite with respect to keying.
As such, like other measures of attachment that have this small number of items. We also found that the correlations be- problem, the ECR-RS anxiety scales have the potential to suffer tween the attachment dimensions across relational domains were from response acquiescence.
This has a number identified, but not resolved, by Fraley et al. For self-report measures of attachment. Most attachment measures one, it indicates that the common trait-like approach to assessing tend to be good at differentiating among people on the insecure attachment, although valuable in its own right, might not allow the end of the spectrum but are relatively poor at differentiating people nuances of attachment across contexts to be assessed adequately.
If on the secure end. As an analogy, most self-report measures researchers are primarily interested in studying attachment in ro- function in the same way that algebra-focused math tests might mantic relationships, it might be advisable for them to assess function.
These tests can differentiate people with poor math skills individual differences in that domain specifically rather than to fairly well, but they cannot differentiate people with basic knowl- assume that a more global or less specific measure captures the edge in calculus from people with advanced training in differential variance of interest.