The survey results corroborated these performance gains in
ComLittee. The workload required to complete the task (measured via NASA-TLX) was significantly reduced in
ComLittee (for
ComLittee, M=14.1, SD=5.56; for baseline M=17.0, SD=6.01, Wilcoxon
W=13.5,
p=0.01). Users also responded that
ComLittee better supported (a) author discovery: ‘helped me find relevant authors’, M=6.1 (
ComLittee) vs. M=3.9 (baseline), Wilcoxon
W=2.5,
p=0.002 ; ‘helped me make sense of author’s research’, M=4.8 (
ComLittee) vs. M=3.5 (baseline), Wilcoxon
W=0.0,
p=0.008 ; ‘made me curious about author’s research’, M=6.1 (
ComLittee) vs. M=4.3 (baseline), Wilcoxon
W=3.5,
p=0.001 ; ‘explanations of relevant authors became more helpful the more I used the system’, M=4.9 (
ComLittee) vs. M=3.9 (baseline), Wilcoxon
W=11.0,
p=0.02 and (b) paper discovery: ‘helped me find relevant papers’, M=6.1 (
ComLittee) vs. M=3.9 (baseline), Wilcoxon
W=6.0,
p=0.002 ; ‘made me curious about the papers I found’, M=6.2 (
ComLittee) vs. M=4.8 (baseline), Wilcoxon
W=3.5,
p=0.01 ; ‘explanations of relevant papers became more helpful the more I used the system’, M=4.6 (
ComLittee) vs. M=3.6 (baseline), Wilcoxon
W=21.5,
p=0.05 (see Table
3 for details). Consistent with the perception of helpfulness, users favored
ComLittee in terms of the overall technology compatibility with their existing scholarly discovery workflows (for
ComLittee, M=22.6, SD=2.99; for baseline, M=19.4, SD=5.08, Wilcoxon
W=24.0,
p=0.02) and the plausibility of future adoption (for
ComLittee, M=6.2, SD=0.75; for baseline M=4.9, SD=1.50, Wilcoxon
W=4.0,
p=0.003, see Table
2 for details).