Confirmatory factor analysis of the cognitive failures questionnaire: evidence for dimensionality and construct validity

https://doi.org/10.1016/j.paid.2003.09.005Get rights and content

Abstract

The Cognitive Failures Questionnaire (CFQ) is a measure of everyday task failures that individuals are normally capable of completing. Prior studies have attempted to gauge the dimensionality of the measure using exploratory techniques, but not a confirmatory approach. Over two studies, the dimensionality of the CFQ was examined using structural equation modeling by comparing all previously reported factor models. Additionally, construct validity was obtained in study 2 by beginning to delineate a nomological network. It was found that a four factor solution consistently provided better fit and was more theoretically related to original conceptualization of the construct. Results of the nomological network also provided support for the four factors of the CFQ by demonstrating convergence with similar constructs (e.g., neuroticism) and divergence with dissimilar constructs (e.g., attention). Implications, limitations, and future research ideas are discussed.

Introduction

Freud (1941) was perhaps the first investigator to examine and speculate on possible interpretations of action slips as highlighted in his book The Psychopathology of Everyday Life. This seminal work, along with the work of Heckhausen and Beckmann (1990), Reason (1979) and Norman (1981), has influenced many error and action slip researchers to search for the mechanism(s) that might influence or cause such slips. One particular construct that has recently received a great deal of attention is that of cognitive failure (Broadbent, Cooper, Fitzgerald, & Parkes, 1982; Larson & Merritt, 1991; Robertson, Manly, Andrade, Baddeley, & Yiend, 1997; Wallace, Kass, & Stanny, 2002).

Cognitive failure is defined as a person’s failure to complete a task that he/she is normally capable of completing (Wallace et al., 2002). Several common examples of cognitive failure have been presented over the years such as putting flour in one’s coffee instead of sugar or throwing away an item one meant to keep and retaining the undesired item. While such errors are comical at best, other more serious errors can also occur under relatively simple and familiar conditions (e.g., turning the wrong way on a one-way street) and can be manifested in a variety of mannerisms (e.g., stress, negative affect). The definition of the construct implies that a person possesses the ability to complete the task, but rather something else interferes with the completion of the task. This ‘something else’ is precisely what researchers have been trying to identify since Broadbent et al. (1982) coined the term cognitive failure.

The cognitive failures questionnaire (CFQ: Broadbent et al., 1982) was developed to provide an everyday, all-encompassing measure of common errors and was derived from three categories: memory slips (e.g., absentmindedness), attention slips (e.g., fail to notice something relevant), and psychomotor slips (e.g., action slips). Several studies have attempted to ascertain the dimensionality of the construct with each investigation yielding differential results (e.g., 1 factor, 2 factors, 3 factors). Empirical investigations into the cognitive failure/performance relationship have utilized total scores derived from the CFQ as well as various factor scores that have been presented from a variety of researchers using varied factor analytic methods (to be described below). As one might suspect, results have greatly varied as a function of which factor structure one might employ. Thus, the purpose of the present study is to (1) assess the factor structure of the CFQ using confirmatory factor analysis to test all the proposed solutions that have been derived using exploratory factor analytic methods and (2) re-test the factor model identified in study 1 and examine the construct validity of this solution with other personality and individual difference measures.

Section snippets

Study 1: confirmatory factor analyses

Originally, Broadbent et al. (1982) stated that the CFQ only assesses a general liability to failure. He arrived at this conclusion based on factor analyses he conducted and stated that “apart from the obvious general factor the results (of the factor analyses) are highly variable from one group to another” (p. 5). The results of his factor analyses are not reported and he concludes the scale should only be used to assess a single general factor: cognitive failure. He further supports his claim

Participants

A total of 709 participants were recruited to participate in the study. Such a large sample helps increase the likelihood of identifying a stable model (Guadagnoli & Velicer, 1988). Participants were recruited from two large universities located in the Southeastern part of the United States. No differences in cognitive failure were found between university students and thus the samples were combined. Of the sample 385 were male and the sample had a mean age of 20.3 (SD=3.4). Additionally, the

Results

The first model tested was the single factor model proposed by Broadbent et al. (1982). This model failed to meet the criteria of Hu and Bentler. Specifically, this model yielded a χ2275=1208.89 with a CFI=0.92, RMSEA=0.07, and a SRMR=0.038. Table 1 contains additional fit indices for this and all other models tested. The second model tested was the two factor solution presented by Larson et al. (1997). This model yielded a χ2274=987.12 with a CFI=0.94, RMSEA=0.06, and a SRMR=0.033. This model

Discussion

The results of study 1 supported the Wallace et al. (2002) 4 factor solution as the best factor structure of all others that have been examined and proposed to date. This factor structure is also very similar to the original conceptual development of the scale in that the factors are theoretically related to the three categories Broadbent drew from (i.e., memory, perception, and psychomotor).

An additional note should be made regarding the high correlations between the factors proposed. While

Study 2

Factor scores might allow researchers to gain a much more comprehensive assessment and understanding of cognitive processes and behaviors such as cognitive failures. Study 2 was designed to gain additional evidence for the factor structure proposed by Wallace et al. (2002). In the article by Wallace and colleagues, construct validity evidence was obtained for the four factor solution. For example, it was found that the factors correlated with facets of boredom, attention deficit, and

Five-factor model of personality

Conscientiousness is a dimension of personality that is typically described as hard-working, organized, disciplined, and volitional. The two personality constructs of cognitive failure and conscientiousness seem at odds with each other and it was believed that a strong negative relationship exists between the two based on the finding of Wallace and Vodanovich (in press). Additionally, it was believed no differential prediction exists between the CFQ factors and conscientiousness as

Confirmatory factor analysis: replication of study 1

In addition to the nomological network that will be examined in study 2, the factor structure of Wallace et al. (2002) will again be tested to determine if the fit of the model still meets the criteria established by Hu and Bentler (1999). However, model comparisons will not be made as they were in study 1 except for the comparison of the Pollina solution as it was a very close second to the Wallace solution.

Participants and procedure

The sample consisted of 386 undergraduate students from a large university in the Southeastern United States. Data collection was part of a larger project and all data presented herein are unique to the current paper. Participation was voluntary and all data were collected anonymously in a single session. The average age of the sample was 19.8 (SD=1.92) and consisted of 183 males with 63% of the sample being Caucasian, 25% Asian-American, 8% African-American, 2% Hispanic, and 2% other or

Confirmatory factor analysis

To assess the fit of the factor structure of the CFQ proposed by Wallace et al. (2002) and supported in study 1, the four factor model was assessed using the sample for study 2. This analysis yielded a χ2246=576.39 with a CFI=0.90, RMSEA=0.05, and a SRMR=0.04. While the CFI is lower in this sample as opposed to the sample in study 1, the fit indices meet the criteria of Hu and Bentler (i.e., RMSEA < 0.06 and SRMR < 0.08) and the minimal acceptance level for CFI for any model tested with Structural

Discussion

As stated previously, ample support was gained for the nomological network that was proposed for study 2. Specifically, it was found that cognitive failures did not relate to ability, but did relate to various facets of personality such as conscientiousness and neuroticism as well as a host of off-task processes conceptually related to cognitive failure such as thought occurrence and action-state orientation. While there were a number of significant correlations among the CFQ factors and other

General discussion

The Wallace et al. structure appears to be the single best structure that is most representative of cognitive failure and the types of cognitive failures one might experience in everyday life. Additionally, the structure is more parsimonious and possesses more utility in distinguishing types of cognitive failure than of any other structure tested. It also outperformed all other proposed solutions in the confirmatory tests employed over two distinct samples. Study 2 provided construct validity

Acknowledgements

This paper was presented at the American Psychological Society’s annual conference, May 2003 Atlanta, GA. I wish to thank Erin Page and Gilnd Chen for their insightful comments and assistance in the development of this article.

References (28)

  • N. Meiran et al.

    Individual differences in self-reported cognitive failures: the attention hypothesis revisited

    Personality and Individual Differences

    (1994)
  • I.H. Robertson et al.

    ‘Oops!’: performance correlates of everyday attentional failures in traumatic brain injured and normal subjects

    Neuropsychologia

    (1997)
  • D.E. Broadbent et al.

    The cognitive failures questionnaire (CFQ) and its correlates

    British Journal of Clinical Psychology

    (1982)
  • B.M. Byrne

    Structural equation modeling with LISREl, PRELIS, and SIMPLIS: basic concepts, applications, and programming

    (1998)
  • J.M. Diefendorff et al.

    Action-state orientation: construct validity of a revised measure and its relationship to work-related variables

    Journal of Applied Psychology

    (2000)
  • Freud, S. (1941). Zur Psychopathologie des Alltagslebens [The psychopathology of everyday life]. Gessammelte Werke, Bd....
  • L.R. Goldberg

    The development of markers for the big-five factor structure

    Psychological Assessment

    (1992)
  • E. Guadagnoli et al.

    Relation of sample size to the stability of component factors

    Psychological Bulletin

    (1988)
  • S.M. Gully et al.

    The impact of error training and individual differences on training outcomes: an attribute-treatment interaction perspective

    Journal of Applied Psychology

    (2002)
  • H. Heckhausen et al.

    Intentional action and action slips

    Psychological Review

    (1990)
  • L.M. Hough et al.

    Personality traits, taxonomies, and applications in organizations

  • L. Hu et al.

    Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives

    Structural Equation Modeling

    (1999)
  • K. Jöreskog et al.

    LISREL 8: structural equation modeling with the SIMPLIS command language

    (1993)
  • R. Kanfer et al.

    Motivation and cognitive abilities: an integrative/aptitude-treatment interaction approach to skill acquisition

    Journal of Applied Psychology

    (1989)
  • Cited by (0)

    View full text