Home Measures of relating and interrelating
Measures of relating and interrelating

In this section I will focus upon only those measures that are based upon relating theory (see section on Relating Theory ) and that have been developed since 1990. These measures have evolved from earlier measures (see Birtchnell, 1988, Birtchnell, Falkowski & Steffert (1992), though once relating theory had fallen into place, they replaced the earlier measures, and always contained eight scales that were based upon the octants of the interpersonal octagon (see section on Relating Theory ).


The Person's Relating to Others Questionnaire (PROQ)

The PROQ (as distinct from the PROQ2) was the earliest relating measure and was described in Birtchnell, Falkowski & Steffert (1992). It was later described in Chapter 9 of both How Humans Relate (Birtchnell, 1993/6) and Relating in Psychotherapy (Birtchnell, 1999/2002). It is a 96-item, self-administered questionnaire, with twelve items contributing to each of the eight scales, which correspond to each octant of the interpersonal octagon (see section on Relating Theory ). Of the twelve items, two refer to positive relating, and are not normally scored, and ten refer to negative relating. The positive items are included to give respondents something good to say about themselves. There are four options for each item providing a score range of 0-3. Thus for each octant scale, the score range is 0-30, and the total score, combining the scores for each scale has a maximum of 240.



In 1995, a revised version of the PROQ was produced, called the PROQ2. Its aims were to improve the clarity and factorial structure and to reduce the correlation between scales. The wording of the response options was changed to "Nearly always true" "Quite often true" "Sometimes true" and "Rarely true." Since the introduction of the PROQ2, the original PROQ has dropped out of use. The Birtchnell, Falkowski & Steffert (1992) paper is the only one on the original PROQ; and at present, the only paper on the PROQ2 is the one by Birtchnell & Shine (2000). The 1992 paper contains the items of the PROQ. The items of the PROQ2 have not been published, but copies of the PROQ2 are available from me.

The scales of the PROQ and the PROQ2, named after the octants of the interpersonal octagon, are called upper neutral (UN), upper close (UC), neutral close (NC), lower close (LC), lower neutral (LN), lower distant (LD), neutral distant (ND) and upper distant (UD).

Both the PROQ and the PROQ2 are scored by computer, the computer print-out comprising both a list of octant scores and a graphic representation of scores in the form of shaded areas of octants (see Birtchnell, 1997, 1999 and 2001).


The PROQ2 can be downloaded in the Questionnaires section


An example of a non-patient's PROQ2 scores



An example of a psychotherapy patient's PROQ2 scores



An adolescent version of the PROQ2

A slightly modified version of the PROQ2 was used in a study of boarders and non-boarders at an English public school. (See section on Research Applications )


Psychometric data for the PROQ2

Some psychometric data on the PROQ were published in Birtchnell, Falkowski and Steffert (1992) and Birtchnell, 1993/96 and 1999; and some psychometric data on the PROQ2 were published in Birtchnell (1999) and (on 107 male prisoners) in the Birtchnell & Shine (2000) paper. More extensive psychometric data on the PROQ2, (on samples of 276 non-patients and 432 psychotherapy patients) are provided in Birtchnell & Evans (2003). In the non-patients, the alpha coefficients were in the .70s for three scales (ND, UD and UN) and in the .80s for the remaining five scales. In the patients they were in the .80s for all eight scales. In keeping with the circular ordering of the octants, there were high, positive correlations between some neighbouring scales, and moderate, negative correlations between some opposite ones. There was also a degree of bipolarity between LD and UD. A principal components analysis supported the factorial structure for UN, UC, NC, LN, LC and ND in the non-patients and UN, UC, NC, LN and ND in the patients. For the non-patients, women had higher mean scores on UC and men had higher mean scores on ND. For the patients, women additionally had higher mean scores on NC, LC and LD. Compared with the non-patients, the patients had higher mean scores on all but the UC and UD scales. The mean total scores were 98.5% (sd 26.9) for non-patients and 132.7 (sd 23.9) for patients.


Greek translation of the PROQ2

A Greek translation of the PROQ2, called the PROQ2-GR, was administered to a Greek population sample of 457. The psychometric findings were compared with those of the English population sample of Birtchnell & Evans (2003). The Greeks had higher mean scores on five out of the eight scales. The Greek version of the PROQ2 showed a greater degree of bipolarity than the English version. (Kalaitzaki & Nestoros, 2003).

The PROQ2a and PROQ3

These two measures are still being developed, and do not have the same extensive data as the PROQ2. They are attempts to produce a shortened version of the PROQ2. Both are half the length of the PROQ2, i.e. they have 48 items, six for each octant. Five of these six items are negative and one is positive. The PROQ2a is made up entirely of items from the PROQ2, whereas the PROQ3 includes some new items. The PROQ2a items comprise those with the highest loadings on the eight factors that emerged from a principal components analysis of the PROQ2 items and with the lowest commonalities. In the PROQ3, all of the UC items, and some of the LD items have been replaced. The point of this is to render the UC scale more pathological and to reduce the high correlation between the LD scale and the LN scale.

In 2002, Sean Hammond, at the Department of Applied Psychology at University College Cork, Cork, Ireland, administered the PROQ3, together with the IPIP, a measure of the big five personality factors, to a population sample of 403. The alpha internal reliability scores were .72, .78, .69, .75, .72, .59, .75 and .65 for the UN, UC, NC, LC, LN, LD, ND and UD scales respectively. The allocation of items to the eight scales was confirmed by a confirmatory factor analysis. Two of the scales (UN and UD) did not correlate with any of the big five and two of the big five scales (conscientiousness and open-mindedness) did not correlate with any of the PROQ3 scales. These data have not yet been submitted for publication.


The PROQ2a and PROQ3 can be downloaded in the Questionnaires section


OPROQ3 Questionnaire

This version of the PROQ3 is intended for people to rate others as for instance for members of a psychotherapy group, hence the items are in the third person. The 'O' stands for observer.

The Person's Relating Interview (PRI)

This is a structured interview that covers the same eight measures as the PROQ2, and like the PROQ2, it is a measure of the person's general relating tendencies. It was developed in collaboration with Dr Roberta Leoni, a visiting researcher from the University of Padua, Italy. In contrast to the questionnaire measures, its items are presented one octant at a time, and the interviewer explains before each set of items the general theme to be covered. The octants are presented in the following sequence: neutral close, neutral distant, the three upper octants and the three lower octants. This sequence makes more sense to the interviewee than to work around the eight position of the octagon. It enables the interviewer to examine first the person's tendencies to get close to or distant from others, then to examine classes of upperness and finally to examine classes of lowerness. Each item carries a score of 0-2, depending upon the degree of certainty with which the interviewee answers the question.

An interview has the advantage over a questionnaire that it enables the interviewer to ensure that the interviewee fully understands what each of the items means, and is able to satisfy her/himself that the interviewee actually does have the characteristic enquired about. Although the questions are precisely scripted, both the interviewer and the interviewee are able to ask clarifying questions and even to seek examples. For each octant, the questions are grouped into five sets of five; thus there are twenty five questions per octant, creating a total of 200 questions. This may seem a daunting number of questions, but, with a cooperative subject, they can be completed within 45 minutes.

Each of the five sets deals with a different category of relating. The categories are grouped under the headings of SECURE, EXTREME, DESPERATE, INSECURE AND AVOIDANT, creating the acronym SEDIA (which is Italian for chair). The five SECURE items represent a more determined attempt than with the PROQ2 to measure positive relating. By way of these items, the interviewer is able to attempt to determine whether the interviewee is competent in the relating of each particular octant. This can be done much more effectively in an interview setting than by way of a questionnaire.

The five EXTREME items represent a departure from the standard relating categories. In a sense it is a concession to the thinking that lies behind the interpersonal circle of the interpersonal psychologists (Leary, 1957; Kiesler, 1996) and the concept of a "preferred" relating style. A person who relates extremely has a marked inclination to relate in a particular way, be it either positively or negatively. The category was included because it appears that sometimes, people who get high negative score on a particular octant also get high positive scores on it.

The remaining three sets of five items refer to the three standard negative relating categories, as defined in the section on Relating Theory (see section on Relating Theory ). However, whereas for the PROQ2, the ten negative items for any particular octant may have been derived from any of the three negative categories, in the PRI there are five items for each of these three categories. Thus it has a separate measure for each of the three classes of negative relating.

The DESPERATE, INSECURE and AVOIDANT scores for each octant can be summated to produce a total negative score, the maximum of which (30) is the same as the maximum negative score for each octant of the PROQ. This permits a degree of comparison between the scores of the two instruments. Because the items of the PRI are listed in an orderly sequence, they can easily be scored by hand.

Observations of Relating Behaviour (ORB)

The ORB is a checklist completed by an observer, concerning the relating behaviour of another. It also was developed in collaboration with Dr Roberta Leoni. As with all the other instruments that are described here, its scales are based upon the octants of the interpersonal octagon. Structurally, it resembles the PRI, in that it adopts the same five SEDIA categories; but unlike the PRI, there is only one item for each category. As with the PRI, each item carries a score of 0-2, depending upon whether the observer considers the characteristic to be not present, slightly present or markedly present. For any octant, there is simply a 0-2 score for SECURE, a 0-2 score for EXTREME, and a 0-2 score for each of the three negative categories of DESPERATE, INSECURE and AVOIDANT; but the D, I and A scores can be added to give a total negative score of 0-6.

There are two versions of the ORB, one in which the five categories for each octant are grouped in one place and one in which all items for all categories and all octants are distributed randomly. The second version is to avoid any halo effect of the observer's judgement on one category influencing her/his judgement on another. In the first version the scoring is easier because all items for each octagon are in the same place. In the second, a roster is required to identify the octants and categories the items belong to.

Where the PROQ and to a lesser extent, the PRI rely upon the person's subjective judgement about how s/he considers that s/he relates to others, the ORB is the objective judgement made by another person. Both the PROQ and the PRI rest upon the assumption that the person is (1) telling the truth and/or (2) is sufficiently aware of her/his relating tendencies to be able to provide an accurate description of them. However, one has to acknowledge that, with the ORB, the independent observer can (1) not be telling the truth and/or (2) be a poor judge of the person's relating behaviour.



The Couple's Relating to Each Other Questionnaires (CREOQ)

The interrelating between two people can be measured by a set of questionnaires called the Couple's Relating to Each Other Questionnaires (CREOQ). These were first developed for the measurement of the interrelating between two people in a couple relationship. They can however be modified to measure the interrelating between any two specified people. The CREOQ is made up of a set of four questionnaires called the MS, MP, WS and WP. The MS measures how the man considers he relates to the woman, the MP measures how the man considers the woman relates to him, the WS measures how the woman considers she relates to the man and the WP measures how the woman considers the man relates to her. As with the PROQ, each of the four questionnaires has 96 items, 12 for each of the eight octants. Again, ten of these concern negative relating and two concern positive relating. As with the PROQ, the questionnaires are scored by computer, the computer print-out including both the straight numerical scores and the scores represented by shaded areas of octants. The item allocations for the scales are published in Touliatos, Perlmutter & Holden (2000).


There is an optional, additional questionnaire that is commonly used with the CREOQ called the US. It is the same for each partner. The letters US do not stand for anything; they simply mean us, as in the two of us. The US has 20 true/false items. It measures how each partner considers the two partners get on together. Each item scores zero or one. Half the items require a true response for a score and half require a false response for a score. Scores contribute to a measure of relationship difficulty; so a maximally poor relationship gets a score of 20. In an unpublished study (Birtchnell & Spicer) the mean US scores for 32 couples reporting a good relationship were 1.4 (sd 1.8) for men and 1.7 (sd 2.1) for women. The mean scores for 92 couples seeking couple therapy were 8.8 (sd 5.3) for men and 10.5 (sd 5.6) for women. The t values were 7.62 for men and 8.66 for women, both p <0.001) A sample of over 100 couples reporting a good relationship was recently collected by Gordon (unpublished).


The CREOQ questionnaires can be downloaded in the Questionnaires section




An example of CREOQ scores for a couple not seeking therapy




An example of CREOQ scores for a couple seeking therapy



Unpublished data on the CREOQ

In an unpublished study by Birtchnell & Spicer, statistical data were presented for 32 couples with apparently good marriages and 92 couples seeking couple therapy. Because there were four questionnaires and because each questionnaire had eight scales this amounted to a large amount of data. There were for example 64 different alpha coefficients. The mean alpha coefficient was 0.77 (sd 0.09) for the 32 scales of the good marriage sample and 0.71 (sd 0.11) for the 32 scales of the couple therapy sample. There was a marked tendency for women to be more critical than men, to some extent of themselves, but to a larger extent of their partners. In the good marriage sample, both men and women were much more critical of themselves than of their partners. In the couple therapy sample, individuals criticised themselves more on some scales and their partners more on others. For both the self-ratings and the partner-ratings, the mean couple therapy scores were much higher than the mean good marriage scores; there being one exception: the good marriage women rated themselves significantly higher on lower neutral. Thus being a meek helpless woman was good for the relationship.


Lately, additional CREOQ data have become available from questionnaires accumulated by Deidre Gordon as part of her PhD study for the University of Kent. By adding Deidre's data to existing and recently acquired data two new samples have been created comprising a population sample of 130 couples and a couple therapy sample of 157 couples. With the help of Stijn Voortman, a student from Nijmegen University in the Netherlands, a new set of psychometric and factor analytic data have been created and submitted for publication (Birtchnell, Voortman & Gordon, submitted). Stijn Voortman and Cor DeJong (also at Nijmegen University) translated the US and the CREOQ into Dutch and gave this, together with a version of the Interpersonal Check List (ICL) that had been modified by adding two additional scales, bringing the total number of scales up to ten, to a sample of 89 couples. The members of the couples were asked to respond to the ICL items (1) as they applied to themselves and (2) as they applied to their partner. The Dutch psychometric data were compared with those from the English population sample. Correlations between the CREOQ and the modified ICL were calculated. (Birtchnell, DeJong, Voortman & Gordon, submitted).



On the basis of an exploratory factor analysis, a shorter version of the CREOQ, with only 48 items compared with the original 96, for each of the four scales, MS, MP, WS and WP, was created. As with the CREOQ, the items for the MS, and the WS and for the MP and the WP, are identical apart from gender words. It is made up of those items with the heaviest loadings on the extracted factors, though excludes items loaded on more than one factor. A number of new items were introduced in order to more clearly differentiate between certain neighbouring scales (e.g. UN and UD). This new version is called the CREOQ3, so as to be comparable with the PROQ3 (though there never was a CREOQ2).


The CREOQ3 can be downloaded in the Questionnaires section

Applications of the CREOQ scales

The four measures of MS, MP, WS and WP represent the starting point for any measure of interrelating. They can be modified to measure the interrelating between any specified pair of people, such as, for example, a parent and child. If the tense is modified, they can be used retrospectively, to measure say the interrelating of a couple at an earlier stage of the relationship, or the interrelating of a parent and child during childhood. Vaughan & Fowler (in press) have adapted them for the measurement of the interrelating of a voice hearer and the voice.


Measures that are appropriate for families of three or four

A set of measures has been developed, called the Father, Mother, Child Questionnaires (FCMQ) to measure the interrelating between an adult and her/his two parents. Further questionnaires have been added to create the FCCMQ, for comparing the interrelating of two adult siblings and the same two parents (see Birtchnell, 2001).



Birtchnell, J. (1988a) Depression and family relationships. British Journal of Psychiatry, 153, 758-769.

Birtchnell, J. (1988b) The assessment of the marital relationship by questionnaire. Sexual and Marital Therapy, 3, 57‑70.

Birtchnell, J. (1993/96) How Humans Relate: A New Interpersonal Theory. Hardback, Westport, Con.: Praeger; paperback, Hove, Sussex: Psychology Press.

Birtchnell, J. (1997) Attachment in an interpersonal context. British Journal of Medical Psychology, 70, 269-275.

Birtchnell, J. (1999/2002) Relating in Psychotherapy. Hardback, Westport, Con.: Praeger; paperback London: Brunner-Routledge.

Birtchnell, J. (2001) Relating therapy with individuals, couples and families. Journal of Family Therapy, 23, 63-84.

Birtchnell, J., DeJong, C., Voortman, S. & Gordon, D. (2006) The Dutch version of the Couple's Relating to Each Other Questionnaires (CREOQ-D) Birtchnell, J. & Evans, C. (2003) The Person's Relating to Others Questionnaire (PROQ2) Personality and Individual Differences 36, 125-140. Birtchnell, J., Falkowski, J. & Steffert, B. (1992) The negative relating of depressed patients: A new approach. Journal of Affective Disorders, 24, 165-176.

Birtchnell, J. & Shine, J. (2000) Personality disorders and the interpersonal octagon. British Journal of Medical Psychology, 73, 433-448.

Birtchnell, J. & Spicer, C. (unpublished) A new interpersonal system for describing and measuring the relating of marital partners.

Birtchnell, J., Voortman, S., DeJong, C. and Gordon, D. Measuring interrelating within couples:  The Couple's Relating to Each Other Questionnaires (CREOQ). Psychology and Psychotherapy: Theory, Research and Practice, 79, 339-364.

Kalaitzaki, A.E. & Nestoros, J.N. (2003) The Greek version of the Person's Relating to Others Questionnaire (PROQ2-GR): Psychometric properties and factor structure. Psychology and Psychotherapy, Theory, Research and Practice 76, 301-314. Kiesler, D.J. (1996) Contemporary Interpersonal Theory and Research. New York: Wiley.

Leary, T. (1957) Interpersonal Diagnosis of personality. New York: Ronald Press.

Touliatos, J., Perlmutter, B. & Holden, G.W. (2000) Handbook of Family Measurement Techniques-NCFR, Second Edition. Sage Publications, Inc.: Thousand Oaks, Cal.