Use of the average item response, as opposed to the summated
score, ensured that missing responses would not lead to a systematic negative bias
of the scale scores. A scale score was computed only for those respondents who had
valid responses for at least 75% of the items in the scale. If someone answered fewer
than 75% of a scale’s items, the respondent’s scale score was set to be missing and
omitted from that particular scale. The reliability of the obtained scale scores was
estimated using Cronbach alpha (a).