One of the most world-famous CEO’s once said that, “most skills can be learned, but it is difficult to train people on their personality” (Branson, 2013). The ability to confidently use personality as a predicting factor in various scenarios, is highly useful.
Personality measurement is a method of using explicit rules to score and interpret characteristic features of interpersonal style, which are representative of empirically supported traits. These assessments are used to predict responses in specific settings (Hogan, Hogan, ; Roberts, 1996). These assessments are used widely in both professional and informal settings, such as clinical, educational, organisational and forensic contexts. There are many types of personality measurement tests, all varying in degrees of reliability, ability to be consistently reproduced, and validity, extent to which the result shows what is believed to show or that a test manipulation is doing what it set out to do (Haslam ; McGarty, 2014).
The essay will focus on two diverse personality tests, the Minnesota Multiphasic Personality Inventory-2 (MMPI-2) and the Rorschach Inkblot Test. Both tests will be evaluatively compared in terms of reliability and validity and their subsequent sub-types, including inter-rater reliability, test-retest reliability, convergent validity and internal validity, using a range of contemporary research, to establish their fortes and downfalls.
The MMPI-2 is a self-report personality inventory, consisting of 567 statements, in the topics of health, social attitudes, relationships and behaviours, which the participant rates as true or false. The MMPI-2 is based on 10 clinical scales assessing 10 categories of abnormal behaviour, so a participant’s answers are scored in comparison to these group norms (Drayton, 2009)(Passer & Smith, 2008).
The MMPI-2 is used in health care settings, legal cases, and in screening for high-risk professions (Funder, 2016). It is also being increasingly used in legal cases, such as a recent study which used the test to differentiate offenders into subgroups with specific cognitive, affective and behavioural characteristics (Dargis & Koenigs, 2018).
An example of a projective personality measurement is the Rorschach Ink Blot test. Originally developed to identify mental illnesses, it was later identified as being able to recognise certain personality traits. It consists of 10 ambiguous inkblots, which the participant must describe to the psychologist. The notion is that the mind will try to impose meaning on the inkblot, and this inkblot description is actually a description of the person and how they project meaning onto their surroundings (Weiner, 1998). The psychologist scores and interprets their responses, to determine the personality characteristics they possess (Thygesen, 2018).
The Rorschach is most commonly used by clinicians to diagnose personality disorders, (James, 2013); as many as 82% of clinical psychologists have used it as tool of assessment (Watkins, Campbell, Nieberding, & Hallmark, 1995). Other than being used by clinicians, it has been used in schools, courtrooms and custody units, with 20% of psychologists in correctional facilities in the UK using it routinely (Raynor & McIvor, 2008). It has come to light recently, that the Rorschach was even used to study Nazi war criminals in the past (Dimsdale, 2015).
During the use of the Rorschach, the role of the psychologist is to interpret the participant’s responses. This is viewed as a very subjective method, therefore important that the inter-rater reliability, how similar measurements are between different ‘raters’ (Haslam ; McGarty, 2014), is high, to ensure reliability of the personality test. Viglione ; Taylor (2003) specficially investigated this reliability issue using the Comprehensive System, a method introduced to increase the rigor of the Rorschach’s scoring system (Garb, 1999). Out of the twenty-four published papers reviewed, including 84 ‘raters’, the majority of studies reported inter-rater reliabilites between 85% and 95%. Supportingly, a more recent study found a mean of 88% and a median of 92% (Viglione, Blume-Marcovici, Miller, Giromini, & Meyer, 2012). Granted that a psychologist trained to use the Rorschach adminsters the test, this level of high inter-rater reliability, means that it can be used with confidence, and a wildly different outcome of personality measurement is unlikely to be achieved. However, this limits the use of the assessment, as only trained professionals can interpret test responses.
Test-retest reliability, arguably the most important measure of reliability, is defined as “a measure of the reproducibility of the scale, that is, the ability to provide consistent scores over time in a stable population” (Aaronson, et al., 2002). The same test is run on two separate occasions, correlated, and a test-retest reliability coefficient score is produced. A range of research has investigated the test-retest reliability of both the Rorschach and the MMPI-2 personality assessments. A comparative meta-analysis investigated this measure of reliability and found average values above chance levels for both, 0.84 for the MMPI-2 and 0.86 for the Rorschach. These high coefficients suggest they are acceptable for use as psychometric personality tests (Parker, Hanson, & Hunsley, 1988). A key strength of this research is the aggregation of data from a variety of studies, as it signifies high statistical power for the test-retest reliability coefficients. Despite this, the age of the study reduces its evaluative power, as more contemporary research may show different findings.
A study by Exner & Weiner (1995) found the coefficient score of the Rorschach ranged from 0.70 to 0.80, when tested in time periods from one week to three years. The test-retest coefficient shows a good reliability score, and the significant dur6ation between tests, demonstrates the stability of the test, and its ability to reproduce similar results over a long time period. However, this research was conducted over twenty years ago, and more current research is needed to support the contemporary use of the Rorschach.
The test-retest reliability of the MMPI-2, is further supported by current research. Correlation coefficients exceeding 0.70 were found in a study investigating personality in Spanish adolescents (Zubeidat, Sierra, Salina, & Rojas-Garcia, 2011), parallel to the results of Carlson & Hofstra, 2001 which found statistically significant coefficients. However, other pieces of research have found correlation coefficients of the MMPI-2 to be wide ranging, such as 0.36 to 0.90 extrapolated from a sample of Mexican adolescents (Perez y Farias, Duran, & Gomez-Maqueo, 2003). This demonstrates that despite the majority of research showing the MMPI-2 to have high test-retest reliability coefficients, the reliability is often dependent on the situation, and non-ideal settings can produce massive variation in scores. A apparent limitation of the above study’s, is their sample used, as only school children were studied. This data can only be generalised to this population, and not the population of all children or older populations, as schooling and school friendship groups can influence a child’s personality development.
Additioanlly, Zubeidat, Sierra, Salina, & Rojas-Garcia (2011) also studied the internal consistency of the MMPI-2. Internal consistency is a measure of reliability, which ensures the various test items measuring the different concepts produce consistent scores (Haslam & McGarty, 2014). Measured using Crohnbach’s alpha, the MMPI-2 achieved similar similar scores pre-test and post-test, further supporting the reliability of the contemporary use of the MMPI-2.
These sub-types of reliability, inter-rater reliability, test-retest reliability and internal consistent, are important that they are achieved in personality assessments. Without good reliability, it proves difficult to be confident that the ‘said’ personality outcomes are an accurate representation of the participant’s ‘actual’ personality determined by the instrument, and not an artefact of the test.
Also used to evaluate personality assessments, validity is arguably equally as important as reliability. A valid test ensures that the findings are an accurate representation of the personality it is assessing. It is important that personality measures are a valid assessment tool, particularly the Rorschach as it has been a talking point in the news (BBC News, 2012). As the Rorschach was originally intended to be used to measure disordered thinking, such as found in schizophrenics, we need to be sure it accurately measures personality as well.
Convergent validity is one measure of validity. It refers to the degree to which scores on the personality test correlate with scores on other personality tests, that have been designed to investigate the same personality construct (Connolly, Kavanagh, & Viswesvaran, 2007). A meta-analysis by Garb, Florio, & Grove, (1998) found the MMPI to have a higher convergent validity coefficient than the Rorschach. The researchers fround that the MMPI explained 23-30% of variance, in comparison to 8-14% of variance explained by the Rorschach. Similarly, a less current meta-analysis found that the MMPI demonstrated a coefficient of 0.46 in comparison to Rorschach’s coefficient of 0.41 (Parker, Hanson, ; Hunsley, 1988). These meta-analyses establish that in terms of convergent validity, the Rorschach is not as valid as the MMPI.
The lack of convergent validity of the Rorschach, raises questions about the use of this personality assessment. It could be recommended that in practices requiring a personality measure, the MMPI should be recommended due to its higher convergent validity. However, as a meta-analysis, there is broad variability among coefficients, for both personality tests, suggesting that some scales may be more valid than others, in the same populations and for the same purposes (Hillier, Rosenthal, Bornstein, Berry, ; Brunell-Neulieb, 1999). Further contemporary research is needed to determine which scales, populations, settings and purposes create the highest convergent validity for both personality measures.
A second measure of validity is internal validity, which refers to how strong the cause and effect relationship is, based solely on the effects of the variable being manipulated, and not confounding variables (Haslam ; McGarty, 2014). Social desirability bias can negatively influence the internal validity of a personality assessment. The MMPI-2 is prone to this self-presentation bias, as it is a self-report instrument, and participants may try to answer questions to show themselves in a positive way. The MMPI-2 is more likely to be affected by this bias, negatively affecting its construct validity, whether a measure successfully measures the concept it is supposed to, and subsequently its internal validity, in contrast to the Rorschach (Meyer, 1996).
The general validity of these measures has also been assessed in applied scenarios. The Rorschach was used to differentiate psychopaths from non-psychopaths. Despite the common view that the Rorschach is a clinically sensitive test for this differentiation, a validity coefficient of 0.062 was established, contradicting this view (Wood, et al., 2010). Perhaps the Rorschach achieves higher validity when used in non-clinical settings. Contrastingly, the MMPI-2 achieved the highest validity indicator of other personality assessments, when used in clinical settings to detect feigned psychological symptoms associated with Depression and PTSD (Lange, Sullivan, ; Scott, 2012). The ability for a measure to detect exaggerated traits is hugely valuable. In clinical settings, the use of the MMPI-2 would prevent patients, with no need for support, wasting the resources of the NHS. Additionally, if future research can produce similar results in an employability context, it would ensure only candidates’ with the ‘perfect personality’ are chosen by employers.
The contemporary research discussed provides a wide perspective on the evaluative strengths and flaws of the the MMPI-2 and the Rorschach. In terms of reliability, the Rorschach is highly scored on its ability to provide consistent scores between ‘raters’, as well as it’s significant test-retest reliability coefficients. Despite it’s lower and wider-ranging test-retest coefficients, the MMPI-2 has been shown to have good internal consistency. Alike the variations in reliability, the two tests differ in validity. Whilst the MMPI-2 achieves higher convergent validity scores, the internal validity of the Rorschach is comparatively higher. It has also been suggested that the validity of the Rorschach, and other similar projective tests, is relevant. When used in a clinical setting, it can simply be used to “break the ice” between the therapist and the patient. This may be a non-psychometric function, but this practical use makes the Rorschach worth, as it encourages self-reflection and is able to start a conversation about the person’s internal world. The research discussed indicates that both personality measures have strengths and weaknesses, and further research is required to determine what conditions are most ideal. Recent longditudinal research is also needed, to establish the long-term predicting properties of personality assessments, as most study’s are now out-of-da