Music Learning by older adult novices

These are the data for the PLoS paper of Chmiel, A. et al, 2025, concerning music learning by novice older people. These .csv data files have been exported in R from its .RDS format into .csv. After such exports, Excel does not format ctimprep(counts) data correctly i.e. i-r (e.g. 1-2), instead converting the second, unless 0, to a 3 letter month (Jan or Feb). However, when the .csv is read back into R, without further change, the correct format is regenerated. If a user has any problems with this, ctimprep can easily be reassembled from coexistent separate imprct and repct data, or extracted from the Excel .csv representation. This data also appears correctly in many text and script readers. The five (5) files are listed here (and _di indicates they are anonymised (de-identified): replicreslong_di.csv replication data (4184 obs of 12 variables) improvreslong_di.csv improvisation data for analysing performance items 4,5 ( 7902 obs of 25 variables) improv1reslong_di.csv improvisation data for analysising improvs given as Performance item 1 in improvisation blocks (1980 obs of 24 variables) kdbfluency_di.csv data on self-asssessed fluency in keyboard usage (391 obs of 5 variables) MDT_data_di.csv Melody detection task data (371 obs 10 variables) Like most, our data contain a few elements with missing entries (such as NAs), which in general are disregarded in R analyses. For example there is one incomplete row in the kbdfluency data (with two NA values). There are also a few incidences of 3 personal ids that duplicate others, apparently as a result of data acquisition errors. In certain modelling approaches, these would be included as participants, but as the Group and Session information is correct, there would be only a slight impact on the number of participants in the group effects. We viewed their retention as preferable to assuming that our understanding of their causes was absolutely correct, and so merging the corresponding id pairs. However, a user may conflate the matching pairs into a single pid if they see benefit: lilij and fahif belong together; as do jojav and tikit; and holar and gokaj.The data files are suffixed by _di to indicate that they are anonymised (de-identified). Note that in order to develop models like those we present, it is essential to checkthe factor vs numeric status of all the outcome and predictor variables being used. In translation into and out of R these may be lost. Replication and Improvisation data are in the 'long' format, but subsets andother formats can easily be generated if required for modelling (e.g. in R, with dplyr, or by the use of tibbles).