Integrity assessments are widely used in workplace selection and development, but linguistic complexity may affect respondents’ understanding and compromise assessment fairness in multilingual contexts.
This study examined preliminary scale-score correspondence between the original Giotto integrity assessment and a simplified-language version developed to support respondents with varying English-language proficiency.
In South Africa’s linguistically diverse labour market, assessment language may introduce construct-irrelevant variance. Simplified wording may improve accessibility while preserving the intended measurement purpose.
A quantitative repeated-measures design was used. A sample of 165 working adults completed both the original and simplified-language versions of the Giotto instrument. Because the instrument uses an ipsative forced-choice format, conventional factor analysis and internal consistency estimation were not emphasised. Descriptive statistics, correlations between corresponding scale scores and supplementary exploratory Analysis of Variances (ANOVAs) were used to examine scale-score correspondence and demographic response patterns.
Statistically significant moderate-to-good correlations were found between corresponding dimensions across the two versions, suggesting meaningful scale-score correspondence. Some demographic group differences varied across versions and require further investigation.
Simplified-language assessment formats may support fairer and more accessible workplace assessment practices, particularly in multilingual contexts.
The study contributes preliminary South African evidence on language accessibility in integrity assessment while cautioning that full measurement equivalence was not established.
Integrity is conceptualised as a multidimensional workplace disposition in which personal consistency, moral orientation and professional alignment shape ethically relevant behaviour (Barnard et al.,
This relationship between integrity and job performance should not be surprising, given that counterproductive work behaviour (CWB) is related to other performance behaviours such as organisational citizenship behaviours (Sackett et al.,
In this study, workplace integrity is defined as a multidimensional disposition reflecting consistency between personal values, ethical principles and organisationally appropriate behaviour, expressed through honesty, fairness, responsibility, dependability, self-regulation and commitment (Barnard et al.,
Based on this framework, assessing integrity requires attention to both internal dispositions and their likely behavioural expression in organisational contexts. Instruments such as the Giotto assessment do not measure moral philosophy directly; rather, they assess personality-related tendencies associated with ethical conduct, dependability, responsibility, fairness and resistance to counterproductive behaviour. The relevance of the Giotto dimensions, therefore, lies in their role as behavioural indicators of the broader integrity construct. Integrity assessment in workplace contexts is concerned with the measurement of personality-related tendencies that are relevant to ethical conduct, dependability, responsibility and CWB. The Giotto instrument is a personality-based integrity assessment designed for selection and development contexts and consists of seven dimensions derived from the Prudentius model of personality: (1) Prudence, (2) Fortitude, (3) Temperance, (4) Justice, (5) Faith, (6) Charity and (7) Hope. In the present study, these dimensions are interpreted as indicators of workplace integrity-related dispositions. The focus of the study is not on behaviour change theory, but on whether a simplified language version produces scale scores that correspond with those obtained from the original version.
The Giotto instrument is a work-based personality questionnaire designed by John Rust, from the United Kingdom, intended for selection and development in an ever-changing work environment. It provides a measure of integrity based on seven personality traits. It is a work-based personality questionnaire consisting of 101 items that generates seven traits or dimensions based on the Prudentius model of personality. (Turner & Board,
Limited recent, peer-reviewed research was found on integrity assessment with the Giotto instrument and specifically not for South African organisations.
The seven scales with the underlying traits and score descriptions are presented in
Classical virtues or vices and integrity traits measured with the Giotto instrument.
| Classical virtues or vices | Giotto scale | Integrity traits | Low score | High score |
|---|---|---|---|---|
| Prudence or folly | Prudence | Competence vs. carelessness | Careless Making mistakes or negligence Unreliable Impulsive Risk-taker |
Prudent and cautious Focused on detail Accurate and meticulous |
| Fortitude or inconstancy | Fortitude | Work orientation vs. absenteeism | No long-term job commitment Work only as necessary Lazy Dislikes work pressure Often absent |
Being committed Works hard Cope with work pressure Goal-oriented Ambitious Punctual |
| Temperance or anger | Temperance | Patience vs. hostility | Aggressive Lacks self-control Bad-tempered Emotional Poor interpersonal relationships |
Reconciliation Self-controlled Even-tempered Calm and compromising Well-mannered Inoffensive |
| Justice or injustice | Justice | Fair-mindedness vs. subversion | Irrational judgement Biased and partial Bearing grudges Dissatisfied Unreasonable Suspicious |
Rational judgement Fair-minded Unbiased or impartial Ethical Balanced Satisfied |
| Faith or infidelity | Faith | Loyalty vs. disloyalty | Loyal only to self Self-centred Independent-minded Use company time and property for personal interest |
Loyal to the company Company values and goals Adherence to rules Acceptance of authority Dependable |
| Charity or envy | Charity | Openness vs. disclosure | Dishonest Crafty or scheming Shrewd Covering up mistakes Excessive privacy |
Open and honest Trustworthy Transparent Sincere Straightforward |
| Hope or despair | Hope | Initiative vs. inertia | Resist change Inability to cope with change Passive obstructive Discouraged Disheartened Negative |
Welcomes change Optimistic or positive Innovative Resourceful Adapt to change Enthusiastic |
vs., versus.
The Giotto instrument utilises an ipsative scale format where forced-choice items (least like me to most like me) are matched for social desirability. Examples of the items in Part A of the original version are words like ‘shallow’, ‘threatening’, ‘ethical’ and ‘unselfish’.
The intrinsic non-linearity of the ipsative framework is addressed by using neural network programming techniques. The original Giotto instrument has been standardised on 701 candidates in a variety of employment positions (Pearson Education,
The original Giotto instrument is registered in South Africa and provides local norms, which are updated regularly. The instrument has been standardised in South Africa for all race groups and across a broad range of occupational levels. The Giotto instrument conforms to South African labour and employment equity legislation. The Giotto instrument was classified as a psychological test by the Professional Board of Psychology of the Health Professions Council of South Africa (HPCSA). Professor Mauer, an Emeritus Professor at University of South Africa (UNISA), researched the reliability and predictive validity of the South African version of the test (Mauer,
The prediction validity of the original test was explored (unpublished) using matched samples of members of the public and perpetrators of financial crimes (Mauer,
To determine a reliability coefficient, item-level data are required for internal consistency estimates such as Cronbach’s alpha, while repeated measurements or parallel test forms are required for test–retest or alternate-form reliability. In this study, reliability was assessed through correlations between the original and alternative versions, representing evidence of alternate-form reliability.
Developments in forced-choice psychometrics have increasingly emphasised model-based approaches to ipsative data, particularly Thurstonian Item Response Theory (TIRT). Traditional scoring of multidimensional forced-choice instruments often produces ipsative or partially ipsative scores, which constrain total scores, distort inter-scale correlations, complicate reliability estimation and limit inter-individual comparisons (Brown,
Applications of TIRT have demonstrated its usefulness for examining the factor structure and validity of multidimensional forced-choice personality and character measures, although reviews caution that its success depends on appropriate item-level data, adequate sample sizes, balanced block design, item keying and explicit model-fit evaluation (Dueber et al.,
Occupational psychological assessment is a vital part of recruitment that can prevent poor hiring decisions. Inappropriate recruitment can have significant implications for organisations.
There has been recent interest in ‘dark personality’ traits in the workplace and predicting CWB. This is mostly done through covert integrity assessment instruments. Although these are designed to predict CWB, they have also been reported to predict job performance (Ones et al., 1993; Yang et al.,
The original Giotto instrument was certified by the HPCSA as a psychological test in 2002 and has been used extensively across different industries in South Africa. Giotto (Part A) uses words to describe different ‘personalities’ and in Part B, the candidates need to decide on the extent to which the words (from Part A) apply to themselves. Therefore, candidates must properly understand the meaning of the words presented. Challenges in understanding the meaning attached to the words or constructs and how they manifest in different cultures and/or languages surfaced during the administration of the original Giotto instrument over the last 12+ years. It is important to acknowledge the social history and understanding of words or constructs when personality assessment is done in a language other than the respondent’s home language. English proficiency can be seen as a limitation of personality or integrity assessment (Grobler & De Beer,
An alternative Giotto version was developed in South Africa, where short descriptive sentences were added next to the words in Part A, to address language proficiency challenges. Additionally, a glossary of terms was added to the instrument. The English language used in the original Giotto instrument was simplified and contextualised for South African English-language users. The need to analyse the responses to both versions was identified, and the application value was investigated.
A quantitative cross-sectional research design was employed to examine the preliminary evidence of correspondence between the two versions of the instrument. As no agreement-based visual analyses were conducted, the findings should be interpreted as evidence of scale-score correspondence rather than evidence of agreement, interchangeability or full measurement equivalence.
A sample of 165 working adults completed both the original Giotto instrument and an alternative version containing simplified item wording.
The responses for the Giotto instrument are on an ipsative scale with forced-choice items, which can be problematic for Classical Test Theory analyses. The total scores were therefore investigated. The Giotto instrument presents four words or statements in a block. When completing the Giotto instrument (original and alternative versions), participants were asked to choose the item that describes them the best and the one that describes them the least (the response mode is referred to as MOLE) (Cao & Drasgow,
The Giotto original and alternative versions were administered to 165 candidates to gather data to investigate the psychometric soundness thereof.
Distributional properties were examined using means, standard deviations, skewness and kurtosis values to describe the score distributions for each dimension. Single-stimulus response formats tend to show biases; a general tendency to alternatively use a forced-choice format has been developed in the scientific community (Jackson et al.,
Bartram (
Data is ipsative when the sum of all scores obtained for any individual is constant. When using CTT scoring, regardless of the choices made, item scores in every block always add up to the same number, four for the Giotto instrument. Interpreting ipsative data means that only intra-individual comparisons are possible. In a questionnaire with more than one dimension, it is not possible to achieve high scores on all scales because choosing one statement and thereby one dimension inevitably leads to lower scores on all other dimensions, making the data interdependent on the choices made earlier. Consequently, all ipsative profiles have the same average score regardless of whether the true scores are above or below average. As a result, interindividual comparisons cannot be made (Brown & Maydeu-Olivares,
These findings may be interpreted as preliminary evidence of alternate-form consistency, but they do not by themselves establish construct validity or full measurement equivalence. An exploratory ANOVA analysis was conducted to generate a significance value indicating whether there are differences within the comparison groups (Babbie et al.,
Ethical clearance to conduct this study was obtained from the UNISA College of Economic and Management Sciences_ERC Industrial and Organisational Psychology (Ref. No. CEMS_IOP_13523).
The results are presented in three parts. Firstly, descriptive statistics are reported to summarise the score distributions for the original and simplified-language version. Secondly, correlations between corresponding dimensions are reported as the primary evidence of scale-score correspondence. Thirdly, exploratory ANOVA findings are presented as supplementary information regarding demographic group differences.
Internal consistency indices such as Cronbach’s alpha are commonly reported in psychometric studies; their applicability to ipsative, forced-choice instruments is limited. The interdependence of item responses in ipsative formats violates the assumptions underlying alpha, often resulting in artificially deflated or uninterpretable estimates. In line with methodological recommendations, this study therefore prioritised the examination of alternate-form reliability and convergent validity through correlational analyses between instrument versions, which is considered a more appropriate approach for establishing measurement consistency in forced-choice designs.
The descriptive statistics for the original and alternative versions are presented in
Descriptive statistics of the traits measured with the original and alternative versions of the Giotto instrument.
| Trait (raw scores) | Original |
Alternative |
||||||
|---|---|---|---|---|---|---|---|---|
| Mean | SD | Skewness | Kurtosis | Mean | SD | Skewness | Kurtosis | |
| Charity | 16.68 | 7.30 | −0.45 | 0.67 | 17.12 | 7.45 | −0.39 | −0.11 |
| Faith | 7.54 | 6.92 | 0.11 | −0.70 | 7.31 | 6.55 | −0.12 | −0.03 |
| Fortitude | 19.66 | 8.25 | −0.21 | −0.51 | 19.80 | 7.75 | −0.19 | −0.45 |
| Hope | −11.04 | 6.56 | −0.17 | −0.37 | −11.30 | 6.87 | −0.45 | −0.07 |
| Justice | 11.13 | 8.01 | −0.72 | 0.77 | 10.92 | 7.86 | −0.38 | −0.10 |
| Prudence | 6.78 | 8.45 | −0.38 | −0.08 | 10.05 | 7.35 | −0.88 | 1.33 |
| Temperance | 5.92 | 7.02 | −0.28 | −0.07 | 7.96 | 6.68 | −0.34 | −0.15 |
SD, standard deviation.
The descriptive statistics in
The relationship between the original and simplified-language versions was examined by correlating corresponding scale scores across the seven Giotto dimensions. These analyses were used to evaluate scale-score correspondence and alternate-form consistency between the two versions. The observed correlations ranged from moderate to good, suggesting that the simplified-language version produced scores that were meaningfully associated with those of the original version. However, these correlations should not be interpreted as evidence of full measurement equivalence or as confirmation that the underlying construct structure remained unchanged. Statistically significant moderate-to-good correlations were observed between corresponding dimensions of the original and simplified-language version, such as adequate correlations (0.55 ≤
The ANOVA analyses were exploratory and should not be interpreted as direct evidence of measurement equivalence between the two versions. Rather, they were used to examine whether demographic group differences appeared to vary across versions in ways that may indicate areas for further investigation. The analysis was done on the responses to the original, as well as on the alternative versions of the Giotto instrument. Exploratory ANOVA analyses were conducted to examine whether demographic group differences appeared across the original and simplified-language version. Only statistically significant on a 95% level are reported in
Supplementary analyses of demographic group differences.
| Personality trait | Comparison group | Original version |
Alternative version |
|---|---|---|---|
| Charity | Gender | 7.30 |
4.30 |
| Faith | Gender | 4.54 |
13.83 |
| Fortitude | English | 0.56 | 4.40 |
| Hope | Gender | 15.17 |
9.05 |
| Hope | Gender and English | 3.94 |
5.67 |
| Justice | Education level | 2.99 |
2.34 |
| Justice | Gender and Education level | 2.35 |
1.08 |
| Justice | English | 0.16 | 5.48 |
| Prudence | English | 1.69 | 4.92 |
| Temperance | Gender | 1.65 | 8.13 |
,
From
However, these findings do not directly establish equivalence between the original and simplified-language version. Rather, they suggest that some group-difference patterns varied across versions, particularly for English language groups on Fortitude, Justice and Prudence.
These differences should be interpreted cautiously and may reflect either improved language accessibility, altered item interpretation or sample-specific variation. Further research using designs specifically intended to evaluate differential item functioning or measurement invariance would be required before firm conclusions can be drawn. The exploratory ANOVA findings provide descriptive evidence that some demographic response patterns were similar across versions, while others differed. These results should be interpreted as indicators of areas requiring further investigation rather than evidence of demographic equivalence or fairness.
While the underlying structure of the Giotto instrument has been established in prior research and standardisation studies, the present study focused specifically on whether this structure is preserved under linguistic simplification, as reflected in the strong correspondence between versions. The findings provide preliminary evidence of scale-score correspondence between the original and simplified-language version of the Giotto instrument. The moderate-to-good correlations suggest that respondents’ scores on the simplified-language version were meaningfully related to their scores on the original version across the seven dimensions.
However, these results should not be interpreted as conclusive evidence of construct validity or full measurement equivalence. Construct validity requires a broader body of evidence, including structural, convergent, discriminant and criterion-related evidence, while measurement equivalence would require analyses specifically designed to test whether the two versions measure the same constructs in the same way. The present study, therefore, supports cautious claims of scale-score correspondence and alternate-form consistency at the scale-score level. These findings provide preliminary evidence of correspondence between the two versions. Simplified wording may, however, improve test fairness and accessibility.
The strong correlations provide evidence of alternate-form consistency and scale-score correspondence between the original and simplified-language version, although these findings should be interpreted cautiously, given the possibility of response consistency effects.
Consequently, the present findings should be interpreted as preliminary evidence of scale-score correspondence between the two versions, rather than conclusive evidence of measurement invariance or full construct validation. This finding is consistent with previous research indicating that carefully designed linguistic adaptations can improve respondents’ comprehension without compromising measurement equivalence. In multilingual organisational environments, the ability to adapt assessment language while retaining psychometric integrity is particularly important.
The findings suggest scale-score correspondence between the versions, although they do not establish full measurement equivalence. In many workplace contexts, employees may vary in their levels of English language proficiency, which can influence how assessment items are interpreted. Complex phrasing or unfamiliar terminology may introduce measurement error if respondents misunderstand item content. Providing alternative wording that enhances clarity may therefore improve the accessibility of assessments and contribute to fairer testing practices.
The exploratory group-difference analyses provide additional descriptive context, but they should not be interpreted as direct evidence of measurement equivalence. The fact that some demographic differences appeared in one version but not the other suggests that linguistic simplification may influence how certain groups respond to the instrument. This is an important finding, but it requires further investigation using item-level methods such as differential item functioning analysis or forced-choice IRT modelling.
From an organisational perspective, integrity assessments are widely used in personnel selection and talent management processes to identify individuals who are likely to demonstrate responsible and ethical behaviour in the workplace. The findings of this study suggest that organisations can implement linguistically simplified versions of integrity assessments to support comprehension without undermining the validity of the measurement process. This is particularly relevant in culturally diverse workplaces where linguistic accessibility is an important component of ethical and fair assessment practices.
This study contributes to the limited body of South African research on integrity testing and highlights the importance of linguistic accessibility in psychological assessment. Despite the widespread use of psychometric instruments in organisational settings, relatively limited research has examined how linguistic factors may influence assessment outcomes. By empirically evaluating an alternative version of the Giotto instrument, the study provides evidence that language adaptations can be implemented responsibly when supported by psychometric analysis.
Several limitations should be acknowledged. A limitation concerns the administration of both versions of the instrument to the same participants. Although this design enabled direct comparison between the original and simplified-language version, it may have introduced order effects, learning effects, fatigue or response consistency bias. Participants may have recognised similar item content across versions and attempted to respond consistently, which could have artificially inflated the correlations between the two forms. Future studies should address this limitation by using counterbalanced administration, random assignment to version order, or a time-lagged design to reduce carry-over effects. The sample size was relatively modest and may limit the generalisability of the findings. In addition, the study relied on a cross-sectional design and did not examine predictive validity in relation to workplace outcomes. Future research could extend the current findings by using larger and more diverse samples, as well as by examining the predictive validity of the alternative instrument in relation to organisational performance indicators and CWBs. The study did not employ formal measurement invariance testing (e.g. multi-group confirmatory factor analysis), which is commonly used to establish structural equivalence. However, such approaches are not well-suited to ipsative forced-choice data, and the present study therefore relied on alternative, methodologically appropriate indicators of equivalence.
Despite these limitations, the results provide encouraging evidence that linguistically adapted assessment instruments can maintain their psychometric integrity while improving accessibility for respondents. The findings underscore the importance of considering language proficiency in the design and administration of psychological assessments in multicultural organisational environments. A limitation of the study is that measurement equivalence was inferred primarily from correlations between corresponding scale scores. Although these correlations provide evidence of association and alternate-form consistency, they do not establish full measurement equivalence. Correlations in the observed range indicate that the two versions are meaningfully related, but not interchangeable. Future research should therefore examine structural equivalence using methods appropriate for ipsative forced-choice data, such as Thurstonian IRT or other modelling approaches designed for multidimensional forced-choice instruments.
By positioning integrity as a multidimensional workplace construct, the study clarifies the theoretical basis for evaluating whether simplified language changes affect the scale-score correspondence of an integrity assessment instrument. This study contributes to the field of industrial and organisational psychology by extending current knowledge on the measurement of integrity-related personality traits in multilingual workplace contexts. Although integrity assessments are widely used in personnel selection and talent management processes, relatively little empirical research has examined how linguistic complexity in assessment instruments may influence respondents’ comprehension and the validity of measurement outcomes. By evaluating the preliminary evidence of correspondence of an alternative language version of the Giotto integrity assessment, this study provides empirical evidence that simplified wording can be introduced without substantially altering the measurement of the underlying constructs.
From a methodological perspective, the study contributes to the growing body of research on the adaptation of psychological assessment instruments in culturally and linguistically diverse environments. The findings demonstrate that carefully developed language adaptations can maintain the psychometric properties of an assessment while improving accessibility for respondents with varying levels of language proficiency. This is particularly relevant in the South African labour market, where linguistic diversity presents ongoing challenges for fair and valid psychological assessment practices.
Practically, the study offers important implications for industrial psychologists, psychometrists and human resource practitioners who rely on integrity assessments in organisational settings.
The results suggest that organisations can implement linguistically simplified assessment instruments to enhance comprehension and inclusivity without compromising measurement equivalence. In doing so, the study highlights the importance of considering language accessibility as a critical component of ethical and fair psychological testing in multicultural workplaces.
The present study provides preliminary evidence of scale-score correspondence between the original and simplified language versions of the Giotto integrity assessment. The moderate-to-good correlations between corresponding dimensions suggest that the two versions produce related scale scores, supporting cautious use of the simplified-language version as a language-accessible alternative. However, the findings do not establish full measurement equivalence, structural invariance or unchanged construct validity. Further research using larger samples, counterbalanced designs and analytical approaches suitable for ipsative forced-choice data is required before stronger claims regarding equivalence can be made. The findings support preliminary scale-score correspondence, but not full psychometric equivalence.
The present study provides preliminary evidence of scale-score correspondence between the original and simplified-language versions of the Giotto integrity assessment. Correlations between corresponding dimensions were moderate to good, suggesting that the simplified version produces scores that are meaningfully related to those of the original version. However, these findings do not establish full measurement equivalence, structural invariance or interchangeability. The results should therefore be interpreted as initial evidence of scale-score correspondence and alternate-form consistency. Further research using larger samples, counterbalanced administration, item-level forced-choice modelling and criterion-related validation is required before stronger equivalence claims can be made.
The author, Sonja Grobler, declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
Sonja Grobler: Conceptualisation, Formal analysis, Methodology, Project administration, Resources, Software, Validation, Visualisation, Writing – original draft, Writing – review & editing. The author confirms that this work is entirely their own, has reviewed the article, approved the final version for submission and publication, and takes full responsibility for the integrity of its findings.
The data that support the findings of this study are not openly available and are available from the corresponding author, Sonja Grobler, upon reasonable request.
The views and opinions expressed in this article are those of the authors and are the product of professional research. It does not necessarily reflect the official policy or position of any affiliated institution, funder, agency, or that of the publisher. The authors are responsible for this article’s results, findings, and content.