Factors within Multidimensional Perfectionism Scales: Complexity of relationships with Self-Esteem, Narcissism, Self-Control, and Self-Criticism

https://doi.org/10.1016/j.paid.2006.03.014Get rights and content

Abstract

Each subscale of the Multidimensional Perfectionism Scale (Hewitt & Flett, 1991) was factor analyzed in an attempt to determine whether perfectionism is best described as a categorical or as a dimensional construct. Relationships with Self-Esteem, Narcissism, Self-Control, and Self-Criticism confirmed the existence of relatively more adaptive and maladaptive elements within each subscale. Partial correlations looking at the adaptive while controlling for the maladaptive elements (and vice versa) yielded even clearer contrasts between the two forms of perfectionism. Self-Criticism was especially important in defining maladaptive perfectionism, whereas adaptive perfectionism reflected more positive forms of Self-Esteem regulation. These data supported a dimensional interpretation of perfectionism.

Introduction

Fundamental to understanding personality disorders and other forms of psychological dysfunction is the question of whether they should be conceptualized in categorical or in dimensional terms (e.g., Flett and Hewitt, 2002, Widiger and Frances, 2002). With the categorical approach, a difference in kind is presumed to exist between normality and abnormality, and the apparently straightforward challenge for research and practice is to develop procedures for validly assigning individuals to one category or another. With a dimensional approach, a continuous scale of functioning theoretically connects the polar oppositions that define normality and abnormality. Here, the seemingly more complex challenge is to devise methods for locating individuals along the hypothetical dimension and then to use that information for meeting the practical demands of diagnosis. Each approach has advantages and disadvantages (Widiger & Frances, 2002, pp. 24–28). The potentials of a dimensional model nevertheless have been illustrated in use of the Narcissistic Personality Inventory (Raskin & Hall, 1981) to demonstrate how the supposed pathology of narcissism can be related to healthy self-esteem (e.g., Watson, Little, Sawrie, & Biderman, 1992).

Though designed to assess pathological self-functioning, the Narcissistic Personality Inventory nevertheless includes factors that have adaptive, as well as maladaptive implications (Emmons, 1984, Emmons, 1987, Watson et al., 1984). Specifically, Leadership/Authority, Superiority/Arrogance, and Self-Absorption/Self-Admiration factors often tend to predict adjustment, whereas Exploitativeness/Entitlement is a more consistent correlate of maladjustment. Leadership/Authority, Superiority/Arrogance, and Self-Absorption/Self-Admiration become even more indicative of healthier self-functioning once Exploitativeness/Entitlement is partialed out, and Exploitativeness/Entitlement becomes even more obviously maladjusted once its covariance with the other three factors is removed (e.g., Watson et al., 1988, Watson et al., 1987). Relationships of Leadership/Authority, Superiority/Arrogance, and Self-Absorption/Self-Admiration with relative mental health appear to be mediated by their associations with greater self-esteem (Watson, Hickman, & Morris, 1996).

Efforts to explain these results have focused on a dimensional model of narcissism. The argument of the “continuum hypothesis,” for example, is that conscious representations of the self range along a dimension of self-esteem regulation that changes dynamically with on-going interpersonal relationships (Watson et al., 1992, Watson et al., 2002; cf., Raskin, Novacek, & Hogan, 1991). Healthy self-esteem theoretically anchors the adjusted pole of this continuum. For heuristic purposes, the assumption is that an overlap between healthy self-esteem and adaptive narcissism defines the adjacent “location” along the hypothetical dimension, and this is followed by adaptive narcissism, then by an overlap between adaptive and maladaptive narcissism, and finally by a more purely maladaptive form of narcissism. Leadership/Authority, Superiority/Arrogance, and Self-Absorption/Self-Admiration presumably operationalize variance ranging from the overlap between healthy self-esteem and adaptive narcissism to the overlap between adaptive and maladaptive narcissism as measured by Exploitativeness/Entitlement. Partialing out healthy self-esteem, therefore, “pushes” these factors in a more maladaptive direction, whereas controlling for Exploitativeness/Entitlement has the opposite effect. In conformity with this continuum hypothesis, self-reported narcissism and self-esteem vary predictably when individuals imagine themselves in relationships that are nurturing or destructive (Watson, Trumpeter, O’Leary, Morris, & Culhane, in press).

This model conforms to at least some theoretical speculation about narcissism (e.g., Kohut, 1977, Watson et al., 2002). Rothstein (1984), for example, describes narcissism as “a felt quality of perfection  invested in a panoply of self- and/or object representations in a spectrum of integrations” (p. 4) that ranges “from psychotic to normal” (p. 28). Among other things, this interpretation implies that perfectionism might also define a continuum of self-functioning (Watson, Varnell, & Morris, 1999–2000). That possibility has, in fact, appeared in studies using the Self-Oriented, Other-Oriented, and Socially Prescribed Perfectionism measures from the Multidimensional Perfectionism Scale (MPS; Hewitt & Flett, 1991). Self-Oriented Perfectionism (e.g., “I must always be successful at school or work”) records “a striving to attain perfection in one’s endeavors as well as strivings to avoid failures” (Hewitt & Flett, 1991, p. 457). Other-Oriented Perfectionism (e.g., “the people who matter to me should never let me down”) measures “unrealistic standards for significant others” (Hewitt & Flett, 1991, p. 457). Socially Prescribed Perfectionism (e.g., “the people around me expect me to succeed at everything I do”) reflects “a perceived need to attain standards and expectations prescribed by significant others” (Hewitt & Flett, 1991, p. 457).

All three measures predict maladjustment; but Self- and Other-Oriented Perfectionism also display connections with healthier self-functioning (Enns & Cox, 2002), including greater self-esteem (Flett, Hewitt, Blankstein, & O’Brien, 1991, also see Parker, 1997, Rice and Mirzadeh, 2000). Such data tend to confirm the early argument of Hamachek (1978) that perfectionism has adaptive as well as maladaptive implications, a claim that nevertheless can be controversial (e.g., Flett & Hewitt, 2005). That the continuum hypothesis might be relevant to this debate was suggested when partial correlations controlling for Leadership/Authority, Superiority/Arrogance, and Self-Absorption/Self-Admiration eliminated linkages of perfectionism with greater self-esteem (Watson et al., 1999–2000).

In general terms, the present investigation sought to explore the possibility of relating perfectionism to the dimensional model reflected in the continuum hypothesis. Theoretically, an on-going commitment to high standards for the self and for others could be essential to an adaptive pursuit of competence, but could also move progressively along a “continuum” toward maladjusted perfectionism if taken to extremes (cf., Blatt, 1995, Frost et al., 1990). Indeed, Campbell and Di Paula (2002) conducted a finer-grained analysis of perfectionism by looking for factors within the Self-Oriented and Socially Prescribed subscales of the MPS and discovered that one of two Self-Oriented factors predicted adjustment rather than maladjustment. Campbell and Di Paula, however, (1) did not factor analyze the Other-Oriented Perfectionism subscale, (2) did not assess the consequences of partialing out maladaptive perfectionism on data obtained for adaptive perfectionism (and vice versa), and (3) did not include the NPI in their procedures. The central purpose of this study was to build upon this previous Campbell and Di Paula analysis of the MPS by including all three of these elements.

Along with the NPI, this project included administration of Self-Esteem (Rosenberg, 1965), Self-Control (Tangney, Baumeister, & Boone, 2004), and Self-Criticism (Thompson & Zuroff, 2004) Scales. Measurement of Self-Esteem was essential in defining the more adaptive pole of the hypothetical continuum. Self-Control and Self-Criticism have both been identified as relevant to perfectionism (e.g., Blankstein & Dunkley, 2002, pp. 288–289, 293). Self-Control, like Self-Esteem, was presumed to reflect more purely adjusted forms of self-functioning (Tangney et al., 2004). The Self-Criticism Scale contained Internalized and Comparative Self-Criticism subscales. Both measures predict maladjustment, with such outcomes being more consistent and robust for Comparative Self-Criticism (Thompson & Zuroff, 2004).

Support for a dimensional model of perfectionism would occur if one or more MPS subscale included factors that had both positive and negative mental health implications. Relative adjustment would be apparent in positive relationships of perfectionism with Self-Esteem, Self-Control, and adaptive narcissism and in negative correlations with Self-Criticism and maladaptive narcissism. Relative maladjustment would, of course, be evident in an opposite pattern of results. A further, and perhaps more important expectation, was that these relationships for adaptive perfectionism would become even more obvious in partial correlations controlling for maladaptive perfectionism and vice versa. Such data would point toward the existence of intricate interconnections between adaptive and maladaptive perfectionism along a common continuum that was relevant to adjustment.

Section snippets

Participants

Participants were 531 undergraduates from an introductory psychology course. All earned extra credit for their voluntary contributions to the study. The average age of these 181 men, 343 women, and seven students who failed to indicate their gender was 19.3 (SD = 4.0). Thirteen individuals did not report their race, but the remaining sample was 27.9% African American, 63.8% Caucasian, 1.1% Hispanic, 0.8% Middle Eastern, 2.3% Asian, and 1.7% “Other.”

Instruments

The five inventories used in this project were

Results

Scale statistics for and zero-order correlations among all but the perfectionism measures are presented in Table 1. The relatively lower coefficient alphas for the NPI conform to recent concerns about the internal reliabilities of factors within this instrument (e.g., del Rosario & White, 2005) and may suggest a need for caution in interpreting these data. These correlations, nevertheless, confirmed expectations that Self-Esteem, Leadership/Authority, Superiority/Arrogance,

Discussion

Traditionally, psychologists often conceptualized perfectionism in pathological terms, but measures of perfectionism can predict adjustment as well as maladjustment (Enns and Cox, 2002, Slaney et al., 2002). Do correlations with relative mental health reveal the inadequate operationalization of a wholly pathological construct? Or do such data point instead toward continuities between healthier and unhealthier forms of perfectionism? Progress in understanding this literature seems to require

References (32)

  • G.L. Flett et al.

    Perfectionism and maladjustment: An overview of theoretical, definitional, and treatment issues

  • G.L. Flett et al.

    The perils of perfectionism in sports and exercise

    Current Directions in Psychological Science

    (2005)
  • R.O. Frost et al.

    The dimensions of perfectionism

    Cognitive Therapy and Research

    (1990)
  • D.E. Hamachek

    Psychodynamics of normal and neurotic perfectionism

    Psychology

    (1978)
  • P.L. Hewitt et al.

    Perfectionism in the self and social context: Conceptualization, assessment, and association with psychopathology

    Journal of Personality and Social Psychology

    (1991)
  • H. Kohut

    The restoration of the self

    (1977)
  • Cited by (0)

    View full text