How to distinguish a ‘good’ test from a ‘not so good’ one.

When behavioural tests are used, questions sometimes arise, such as:

  • ‘How accurate is the test’?
  • ‘How do I know it will give me the same results if I do it in 6 months time?”
  • “Is it an accurate portrayal of me?”

All psychometric tests publishers need to ensure that the test does what it claims to do, in different settings, over time. The role of the test author is to maximize the accuracy and consistency of the test, so that the results are dependable and useful. This is done via the key concepts of reliability and validity. Please select a tab below for further information on reliability and validity.

> Reliability

Reliability refers to the stability and consistency of the test results. This is often measured by;

  • Test-retest reliability: the stability of the results over repeated administrations) and

  • Internal consistency: the extent to which the items relating to the same attribute are consistently related to one another.

The more reliable the test, the closer the correlation coefficient will be closer to 1.0 so those measure that have a correlation coefficient over .7 are considered strong.

For Facet5 reliability data, refer to pages 3-10 of the following report.

Facet5 Reliability & Validity
    Description of the reliability and validity process. It contains information on: Reliability, validity, construct validity, concurrent or predictive validity, reference and notes. 
     
> Validity

There are two broad types of validity as follows -

  1. Construct Validity: the extent to which the test actually measures what it claims to measure.
    • Convergent validity: the degree to which test scores are related to scores on other tests measuring the same thing. (refer to page 11)
    • Discriminant validity: the degree to which test scores are unrelated to scores on other tests measuring the same thing. (refer to page 11)
    • Face validity : the degree to which the test ‘looks like’ it will give the results it says it will. (refer to page 10)
    • Content validity: the degree to which it measures the entire domain. (refer to page 10)

  1. Criterion-rated validity: the relationship between test scores and a measure of an independent outcomes e.g. a score on Will and outbound sales performance
    • Predictive validity: the correlation between a test score and an outcome measure that is gathered at some later point e.g. A score on Will and behavioural evidence of Will in a years time. (refer to page 51)
    • Concurrent validity: the correlation between a test score and an outcome measure that is collected at the same time. e.g. A score on will and a measure of Will at this time. (refer to page 51)


For Facet5 validity data, refer to pages 10-51 of the following report.

Facet5 Reliability & Validity

    Description of the reliability and validity process. It contains information on: Reliability, validity, construct validity, concurrent or predictive validity, reference and notes. 
     



Norm Groups

Facet5 has a large number of norm groups for comparative purposes and these are freely available to choose from when using the system. Facet5 also releases new and updated norm groups when sufficient data has been gathered.

For a list of the currently available norms report to the report below.

Facet5 Norm Groups

    A detailed description of the available Norm Groups. It contains:- Guidelines for Norms, Norms available, miscellaneous norms, Industry and job function norms. 
     



Translation Process

Facet5 translators are required to have current knowledge of the HR/OD domain as well as bi-lingual capability. The translation process involves more than one translator for cross-checking and agreement purposes. For more information on the translation process click on the tab below.

> More on the Translation Process

The translation process is an iterative one. While we may use translation services for a ‘first cut’, we do not rely on them since it is critical that the translators capture the meaning of the statements and relate them to the current business environment. For this reason we insist that translators are not only bi-lingual (at least) but that they are currently active in the Human Resources or Organisation Development field.

We provide guidance to the translators. We advise that they should use a multi-stage process involving more than one translator. Items are translated first from English into the target language then the translation is cross checked by another person who translates back in to English. This process is repeated until there is agreement. Sometimes the original author is consulted to clarify specific issues of interpretation as some concepts do not translate easily from one language and culture to another.

Facet5 Translation Engine
Translation of the main reporting functions of Facet5 is more complex laborious than the core questionnaire. The main Facet5 report consists of approximately 30000 words in English, which is a lot of high-grade translation. To assist this task we have created a Facet5 Translation Engine, which shows the English version in one window on screen and then translators can type the target language version directly into another window on the same screen. If there is already a version in the target language that is also shown in a separate window.

In this way translators have content presented to them in a convenient way with the base language version always in front of them. If they are editing an existing language, the current target version is also shown.

Once the text is translated, the Facet5 Translation Engine automatically writes new SQL scripts to be loaded to the system.


Facet5 has always been committed to being a global product, reflecting the way in which business is undertaken on 2000 and beyond. For a comprehensive list of languages and report availability click here. New translations occur on an ongoing basis. To register an interest in a Facet5 product, in a language that is not yet currently supported, please contact us.

Facet5 Translations

    Outlines the translation process, including:- Introduction to translations, process for analysis or translations, norm creation, reference and notes. 
     


Contact us for further information

Technical Reports


The reason for distinguishing a valid and reliable test from the rest of the field is not to do with psychometrics as such, it is to do with solving a business problem. If the test doesn’t measure what it claims to measure and doesn’t do so reliably and consistently, it is unlikely to be useful to the business, because the results simply can’t be counted on. Results show , Facet5 meets these requirements.

Kate Woodley, Psychologist


 

 


Simple Language

Facet5 provides an easy-to-understand framework and a simple common language, readily adopted by teams and organisations, to help solve business related ‘people problems’.

Global Application

Facet5, available in over 30 languages and with corresponding norm groups, is used by organisations on every continent, making it an effective tool for global and local businesses.

For Management

Specifically designed by psychologists for use by management, Facet5 helps people at all levels understand behaviour and more importantly, what can be achieved.

Value Neutral

Facet5 describes, explains and predicts behaviour in non-judgmental ways, so teams and organisations alike can easily understand and utilise the power of personality.