Psychometric tests are a systematic procedure for scoring and evaluating samples according to standards (Urbina, 2004). There are certain issues taken into consideration when creating a new psychometric measure, such as maintaining consistency of results by using identical materials (Howitt and Cramer, 2011). It is also essential to avoid response bias, this is achieved through not using leading questions which suggests expected answers therefore limiting the effect it has on the variability of answers (Howitt and Cramer, 2011).
Hogan and Cannon (2003) defined reliability as the consistency of a measure regardless of what it’s measuring, which may vary, however validity is ...view middle of the document...
In addition, the scale consisted of negatively and positively worded items that help deal with response sets and an overall of 41 questions.
An ethical form (see Appendix B) was submitted to the university, who gave consent to conduct the research. In regards to the British Psychological Society’s ethical guidelines, all participants were given an informed consent. Participants were reminded they had right to withdraw and all information they provided was confidential. They were told the questionnaire was based on the British benefit system and once completed they were made aware that the research was measuring reliability of the scale and were reminded again if they want to withdraw their data. Therefore no deception was involved.
An item analysis was carried out to identify weak items.
Table 1 shows that the initial Cronbach’s α = .668, this indicates a low level on internal consistency for the scale used when measured against Kline’s (1986) threshold figure for acceptability of 0.7.
Table 2, presents the value of Cronbach’s alpha if a particular item was deleted from the scale. It can be seen that removing an item with a high value would result in a higher streamlined Cronbach’s alpha. Therefore removal of certain items, such as benefit8 and benefit3, would lead to an improvement and an increase in Cronbach’s alpha and consistency therefore resulting to better reliability.
Table 3 shows that when 16 items were eliminated, the streamlined Cronbach’s alpha was .890. This suggests that against the threshold of 0.7, the benefit scale has a high...