Running Head: MIDPOINT RATINGS 1 Midpoint Ratings on Personal Constructs: Constriction or the Middle Way?

Midpoint ratings of elements on personal constructs have been given various interpretations, and the placement of the ideal self at the midpoint of a construct has been viewed as violating Kelly‘s (1955/1991) Dichotomy and Choice Corollaries. This paper reports analyses, using a mixed models approach, of repertory grids completed by 80 clients referred to a clinical psychology service at up to five assessment points both preand post-therapy. A larger than expected number of ideal self ratings at the midpoint was found, and consideration of the content of several of the constructs concerned elucidated why it might have been that neither pole was preferred. Midpoint ratings of the ideal self increased over the course of therapy, and were not associated with depression or anxiety. Midpoint ratings of the ideal self and self now were also related to various structural measures of construing. Theoretical and clinical implications of the findings are discussed. Running Head: MIDPOINT RATINGS 3 Midpoint Ratings on Personal Constructs: Constriction or the Middle Way? A basic premise of personal construct psychology, enshrined in its Dichotomy Corollary, is that constructs are bipolar. The importance of this assumption is not least because of its implications for the personal construct view of choice and change. While there is some research evidence supportive of the bipolarity of construing (Bell, 2000; Bonarius, 1965; Millis & Neimeyer, 1991), other studies have challenged the view of constructs as strictly dichotomous (Epting, Suchman, & Nickeman, 1971; Mair, 1967; Riemann, 1990). Amongst the alternative suggestions are that constructs are unipolar (Bonarius, 1984; Riemann, 1990); that the contrast pole of a construct may be selected from a number of possible alternatives, and that as a result constructs may be ―bent‖ (Yorke, 1983); and that some constructs may be more bipolar than others (Walker, Ramsey, & Bell, 1988). The bipolarity assumption was reflected in Kelly‘s original dichotomous method of repertory grid administration, in which the respondent was required to allocate elements (aspects of the person‘s world) to one or other pole of a series of constructs. However, the rating methods that are now much more commonly used allow investigation of the placement of elements not only at particular construct poles but also at other points on construct dimensions. One such point, which will be a primary focus of this paper, is the midpoint. Midpoint ratings of elements on constructs have been given various interpretations by personal construct theorists. For example, it has been suggested that such ratings can imply that the individual is unable to apply the constructs to the elements concerned (Fransella, Bell, & Bannister, 2004). A high number of midpoint ratings for an Running Head: MIDPOINT RATINGS 4 element would therefore imply that the element is relatively meaningless for the individual and, being outside the ―range of convenience‖ (Kelly, 1955) of his or her construct system, is likely to be anxiety-provoking. Some support for this position is provided by the finding in a student sample that high numbers of midpoint ratings in a repertory grid, particularly when applied to the self, were associated with high scores on a measure of meaninglessness from an alienation questionnaire, although with low scores on a measure of normlessness (Winter, Patient, & Sundin, 2009). Midpoint ratings have also been regarded as indicating constriction, the drawing in of the outer boundaries of the perceptual field to deal with apparent incompatibilities, or conflicts, in construing (Kelly, 1955). Essentially, it is argued that conflictual elements may tend to be rated at the midpoint of constructs because the failure to assign them to either pole of the constructs concerned essentially excludes them from the individual‘s field of vision. Some support for this view has been provided by associations that have been demonstrated between the use of midpoint ratings and suicidal tendencies, which themselves have been related to constriction (Kelly, 1961). For example, following a study by Landfield (1976) in which high use of ―not applicable‖ or ―?‖ ratings on a repertory grid were found to characterize suicidal clients, Dzamonja-Ignjatovic (1996) demonstrated that such individuals tended to give a high number of midpoint ratings to the future self (in contrast to ratings of death, fewer of which were at the midpoint in clients who had attempted suicide). Winter et al. (2007), studying clients who had deliberately harmed themselves, found that high numbers of midpoint ratings of the future self were associated with hopelessness and suicidal ideation, and that these, as well as the overall number of midpoint ratings in a repertory grid, reduced significantly over Running Head: MIDPOINT RATINGS 5 the course of personal construct psychotherapy. However, conflicting results were obtained by Neimeyer, Heath, and Strauss (1985), who found no relationship between midpoint ratings and suicidal ideation, and by Hughes and Neimeyer (1993), who showed suicide risk to be predicted by a low number of midpoint ratings and therefore, in their view, a low level of subjective uncertainty. A more recent study by Hanieh and Walker (2007), using an innovative measure of constriction derived from ―psychophotography,‖ found more constricted scores on this measure to be associated with high numbers of midpoint ratings of the self and future self in a repertory grid, and that depressed people displayed higher levels of constriction on all of these measures than did a control group. Our particular concern will be with the placement of the ideal self element on constructs in a repertory grid. Kelly‘s Dichotomy and Choice Corollaries not only imply that constructs are bipolar but that individuals should have a preferred pole of each construct, in that this pole ―seems to provide the best basis for anticipating the ensuing events‖ (Kelly, 1955, p. 64) at that time. The preferred pole is generally assumed to be indicated in a repertory grid by the placement of the ideal self element on the construct concerned. It might be expected, therefore, that the ideal self would tend to be allocated to one or other pole on each construct rather than to intermediate points on construct dimensions, including the midpoint. Our primary aim has been to investigate this assumption. Further aims were to explore the correlates of midpoint allocation of the ideal self, and to consider whether these also applied to allocation of the actual self to the midpoint. Running Head: MIDPOINT RATINGS 6