The role of psychosocial factors on language use among adult Welsh learners - Data

Roedd yr arolwg yn cynnwys 27 cwestiwn ar draws chwe adran: gwybodaeth gefndirol; gallu yn y Gymraeg a’i defnydd; rhesymau dros ddysgu’r Gymraeg a chymhelliant; hyder a phryder wrth siarad; personoliaeth; a barn am siaradwyr Cymraeg. Mesurwyd amlder y defnydd o’r Gymraeg mewn cyd-destunau gwahanol (gyda phlant, gyda phartner, gydag aelodau eraill o’r teulu, gyda ffrindiau, yn y gwaith, ac mewn digwyddiadau cymdeithasol eraill).

Gofynnwyd i’r ymatebwyr raddio eu hyfedredd eu hunain ar gyfer siarad, darllen, ysgrifennu a deall. Addaswyd yr eitemau personoliaeth o’r Big Five Inventory (Rammstedt a John 2007); gofynnwyd am hyder wrth siarad a phryder wrth siarad gan gyfeirio at gyd-destunau a chyd-siaradwyr penodol. Hysbysebwyd yr holiadur drwy restrau postio sy’n gysylltiedig â’r Ganolfan Dysgu Cymraeg Genedlaethol a thrwy grwpiau cyfryngau cymdeithasol i ddysgwyr Cymraeg, gan ddefnyddio dull samplu gwirfoddol.

Roedd yr ymatebion yn ddienw. Mae’r set ddata’n cynnwys 746 o ymatebion.</p><p dir="ltr">Mae’r dudalen hon yn cynnwys y ffeiliau canlynol:</p><p dir="ltr">•Ymatebion llawn yr arolwg, un rhes fesul ymatebydd, gyda 90 o newidynnau sy’n cwmpasu demograffeg, hanes dysgu a darparwr, hunanraddio hyfedredd, amlder y defnydd o’r Gymraeg fesul maes, cymhellion dros ddysgu, profiadau rhyngweithio â siaradwyr rhugl a dysgwyr eraill, hyder a phryder wrth siarad, personoliaeth, ac agweddau tuag at siaradwyr Cymraeg (welsh_learners_osf.csv).

Sylwer na ddadansoddwyd llawer o’r newidynnau hyn yn y llawysgrif gysylltiedig. Mae Age_Group yn cyfuno’r bandiau oedran yn dri chategori, ac mae Proficiency yn gategori sy’n seiliedig ar fodd y pedair sgil hunanraddedig. Defnyddiwyd y sgôr siarad pan fo dau gategori neu ragor yn gyfartal, a’r categorïau ‘advanced’ a ‘proficient’ wedi’u cyfuno.</p><p dir="ltr">•Sgript R sy’n ailgynhyrchu holl dablau, ffigurau ac ystadegau’r llawysgrif gysylltiedig (welsh_learners_analysis.R).</p><p dir="ltr">•Yr holiadur yn Gymraeg (welsh_learners_holiadur.pdf) ac yn Saesneg (welsh_learners_questionnaire.pdf).</p><p dir="ltr">Cynhaliwyd y dadansoddiadau yn R (fersiwn 4.4.1) gan ddefnyddio’r pecyn ordinal (fersiwn 2025.12-29) ar gyfer y modelau a ggplot2 ar gyfer y ffigurau.

Mae’r sgript yn rhedeg o’r dechrau i’r diwedd ar y ffeil ddata a adneuwyd, ac yn ysgrifennu pob tabl fel ffeil CSV a phob ffigur fel PNG 300 dpi.</p><p dir="ltr">Rhoddwyd caniatâd i gynnal yr ymchwil gan Bwyllgor Moeseg Ymchwil Ysgol y Gwyddorau Cymdeithasol ym Mhrifysgol Caerdydd yn dilyn adolygiad annibynnol (Cyfeirnod: SREC/4293).</p><p dir="ltr"><b>English</b></p><p dir="ltr">This dataset supports the manuscript "The role of psychosocial factors on language use among adult Welsh learners", accepted for publication in <i>Applied Linguistics</i>.</p><p dir="ltr">Adult learners are central to the Welsh Government’s <i>Cymraeg 2050</i> strategy, which aims to increase both the number of Welsh speakers and the frequency of daily use of the language.

Relatively little is known, however, about how much Welsh learners use outside formal tuition, or about the social, linguistic and psychological factors that influence that use. This study addresses two questions: what the patterns of Welsh use among learners are outside of formal tuition, and to what extent psychosocial factors affect the frequency of that use.</p><p dir="ltr">The data were collected through an online questionnaire for adult learners of Welsh, available to complete in both Welsh and English.

The survey comprised 27 questions across six sections: background information; ability in Welsh and its use; reasons for learning Welsh and motivation; speaking confidence and anxiety; personality; and opinions of Welsh speakers. The frequency of Welsh use was measured in different contexts (with children, with a partner, with other family members, with friends, at work, and at other social events). Respondents were asked to rate their own proficiency for speaking, reading, writing and understanding.

The personality items were adapted from the Big Five Inventory (Rammstedt and John 2007); speaking confidence and speaking anxiety were asked with reference to specific contexts and interlocutors. The questionnaire was advertised through mailing lists associated with the National Centre for Learning Welsh and through social media groups for Welsh learners, using a volunteer sampling approach. Responses were anonymous.

The dataset comprises 746 responses.</p><p dir="ltr">This page includes the following files:</p><p dir="ltr">•The full survey responses, one row per respondent, with 90 variables covering demographics, learning history and provider, self-rated proficiency, frequency of Welsh use by domain, motivations for learning, interactional experiences with fluent speakers and other learners, speaking confidence and anxiety, personality, and attitudes towards Welsh speakers (welsh_learners_osf.csv).

Note that many of these variables were not analysed in the associated manuscript. Age_Group collapses the age bands into three categories, and Proficiency is a category based on the mode of the four self-rated skills. The speaking rating was used where two or more categories were equal, and the ‘advanced’ and ‘proficient’ categories were merged.</p><p dir="ltr">•An R script reproducing all the tables, figures and statistics in the associated manuscript (welsh_learners_analysis.R).</p><p dir="ltr">•The questionnaire in Welsh (welsh_learners_holiadur.pdf) and in English (welsh_learners_questionnaire.pdf).</p><p dir="ltr">The analyses were conducted in R (version 4.4.1) using the ordinal package (version 2025.12-29) for the models and ggplot2 for the figures.

The script runs top to bottom on the deposited data file and writes each table as a CSV file and each figure as a 300 dpi PNG.</p><p dir="ltr">Permission to conduct the research was granted by the School of Social Sciences Research Ethics Committee at Cardiff University following independent review (Reference: SREC/4293).</p>