Table 1.

Comparison of results from web-based and lab-based measurements. To obtain the reference values, the median was used as an estimate of μ when individual data points or median were reported in the reference study, and sample mean values were used if those were reported instead. The sample standard error of the mean (SEM) along with the number of participants was used to estimate σ, unless standard deviations were directly reported in the reference study. The effect size of the difference (i.e., Lab - Web values) was estimated using the procedure described in Hedges (1982). The values obtained using our infrastructure were indistinguishable except for MRT-in-babble where the performance scores on the web-based measure were higher (better).

Task	Web data (μ ± σ)	Reference source	Reference values (μ ± σ)	Lab - Web difference (Hedges’ g ±σ)
Gap detection	6.1 ± 9 ms	Patro et al. (2021)	5.9±4.15 ms	−0.02 ± 0.19
F0 discrimination	0.5 ± 0.79%	Madsen et al. (2017)	0.51 ± 0.8%	0.01 ± 0.16
ITD detection	28 ± 19 μs	Borjigin et al. (2022)	29 ± 11 μs	0.06 ± 0.17
ILD detection	0.8 ± 0.7 dB	Mills (1960)	0.7 ± 0.4 dB	−0.14 ± 0.45
MRT-in-babble (0 dB SNR)	85 ± 15%	Miner and Danhauer (1976)	75 ± 10%	−0.68 ± 0.18