Let's Talk

### Information on funding

Staffly sp. z o.o. implements a project called: "Conducting R&D work on the technology of temporary work matchmaking based on psychometric and personality factors" as part of the grant in accordance with the project: „303 Innovation Fund”

Project goal: Development of technology that will enable in-depth analysis and profiling of portal users for the purpose of better matching orders in order to increase the efficiency of their work.

Co-financing of the grant project: PLN 800,000 PLN

Project co-financed by the European Union from the European Regional Development Fund and the National Center for Research and Development under the Smart Growth Programme.

Connect

# Accuracy and reliability

## (Cronbach’s alpha)

Cronbach’s alpha is the most common method of analyzing the reliability of psychological tests, which is designed to test the reliability of questionnaires used in psychology.

Estimating the reliability of a scale is based on the correlations between the measurements or items of the scale that make up the scale in relation to the variance of the items. Each measurement consists of a true result for a given concept or phrase, but also an unknown random error. The more a measurement reflects mostly the true result, the more reliable it is. Therefore, the final result is influenced by the number of concepts or phrases – a larger number of scale elements used makes it ultimately more reliable (measures a given phenomenon more accurately).

This is an important reservation from the point of view of the number of features and the number of expressions proposed in the developed psychometric tests. Well, each feature consists of 6 phrases that measure the candidate’s competence in a given area. It’s not much at all. At the same time, the level of final matching to a given job is measured not by one feature, but by 5-7 features, so ultimately it is a scale of 30-42 phrases.

Cronbach’s alpha coefficient was used in the reliability analysis of the developed scales:

$\alpha = \left(k / \left(k-1\right)\right) * \left[1 – \sum \left(s2$i) / s2sum]

where s2i is the variance of k individual items, s2sum is the variance of the sum of all items. The formula shows that if the expressions do not give a true result, but only an error, then the variance of the sum will be the same as the sum of the variances of the individual expressions – then α = 0. In a situation where the error is negligible and the results are true, α tends to the value 1.

In other words, in the interpretation of the result, the closer the value of Cronbach’s alpha is to 1 (for social measurements we assume 0.9), the more reliable the given scale is. However, it is assumed that above 0.6 the potential unknown error is no longer significant.

## Results of the Cronbach’s alpha coefficient

The table below presents the results of the Cronbach’s alpha coefficient for the individual competencies tested.

## Our partners

Join companies using
Staffly psychometric tools

### Official partner of eRecruiter

Staffly is trusted by HR professionals

## Do you want to recruit with the help of psychometrics?

Book a call with our business account manager and discover the benefits of Staffly!
• ### Fintech Summit Awards 2021

Nomination in the HR Tech category
• ### Symbol of Innovation 2022

Nomination in the Innovation category