Statistical analysis
Descriptive statistics were used to define the study population. For
some analysis missing values were replaced by individual scale means,
when at least 50% of the scale items were filled in, to guarantee representative
results. Kendall’s tau_b and Pearson correlations were used
when appropriate, according to the continuous or nominal nature of
the variables. Correlations were used to explore possible relationships
among self-management behaviours, judgment skills, health literacy,
education attainment and asthma control. A composite score of judgment
skills was computed for every participant based on the final
eleven scenarios (i.e. eight scenarios were deleted during preliminary
analysis, due to their low performance, including poor discrimination)
ranging theoretically from 11 to 44, with higher values representing
better judgment skills. A Low and a High judgment group were formed
by splitting the score along the median (Low: 28-36, High 37-44). This
partition is supported by the data since 38% of people in the Low judgment
group selected the most or second most inadequate option four
times on average, while 47% of participants in the High judgment
group selected the same options only once on average. This indicates,
indeed, that the second group had better judgment skills on responding
to the described situation on the scenarios.
Furthermore, several independent t-tests were carried out to explore