From recent observations of z≳7 Lyα emitters (LAEs), a variety of Lyα luminosity functions (LFs) have been derived with contradictory results, either evolution or non-evolution from , the epoch after reionization. This could be because most of the z≳7 LFs comprise photometric candidates, and these might include some contaminations. Using the Subaru telescope, we have conducted the Faint Object Camera and Spectrograph (FOCAS) narrow-band NB980 (λc∼ 9800 Å, FWHM ∼100 Å) imaging and spectroscopy survey of z= 7–7.1 LAEs in order to compare its ‘contamination-free’ result with the z≳7 photometric Lyα LFs that have previously been derived. We have imaged the Subaru Deep Field and the sky around the cluster MS 1520.1+3002. We have found one LAE candidate, but the spectroscopy did not reveal Lyα, although it was deep enough to detect this. We have calculated the expected number of LAEs in our survey, using five z= 7 and three z= 7.7 Lyα LFs from recent surveys. Seven of these are consistent with null detection (– LAEs) within errors, including Poisson statistics and cosmic variance. However, the average values (0.7–1.1 LAEs) predicted from one z= 7 and two z= 7.7 LFs among these seven indicate nearly a single detection. The remaining z= 7 LF predicts LAEs. As for z= 7, the discrepancy is likely to come from the different criteria used for LAE selection. For z= 7.7, there are two possibilities, as follows. (i) If z= 7.7 LAEs are somehow brighter in Lyα luminosity than lower-redshift LAEs, then a z= 7.7 LF is observed to be similar to or higher than lower-redshift LFs, even if attenuated by neutral hydrogen. (ii) All/most of the z= 7.7 candidates are not LAEs, which supports the decline of the LF from z∼ 6 to 7.7 and reionization at z∼ 6–7.7.