Page 68 - Pengurusan Prestasi Nilai Teras POISE UM

5.2 Reliability, Respondent and Item Separation

The reading of the respondent reliability value (0.97) in Table 5.6
suggests respondents’ consistency in answering the instrument is at a
high level (Bond and Fox, 2015). These findings show that respondents
were very consistent in answering the instrument despite being tested
several times. Meanwhile, the separation value of respondents (3.59) is
excellent and this shows the respondents could be separated into four
different levels. Such respondent separation value suggests that the
instrument is sensitive enough to distinguish between individuals with
different levels of performance (Engelhard and Wind, 2019).

Meanwhile, the item reliability index (0.81) also shows that the level
of item consistency is at a good level despite being tested in different
groups (Table 5.5). The item separation value (2.09) indicates that the
NTP instrument was able to categorise items into two groups, namely
the easy group and the moderate group, with no difficult items being
used. Both respondent and item separation values show good values
because they have readings above 2.0.

68 Table 5.6: Reliability, Respondent Separation and Item Values for
the Construct of Pilot Study Instrument

Reliability Value Separation Value
Respondent 0.97 3.59
Item 0.81 2.09

5.3 Discussions and Summary

Based on the findings of the pilot study conducted, 15 items were
retained. A total of 21 items that did not meet the stipulated analysis
requirements were dropped. An overall summary of the items is
presented in Table 5.7.