Why Ordinal Variables Can (Almost) Always Be Treated as Continuous Variables: Clarifying Assumptions of Robust Continuous and Ordinal Factor Analysis Estimation Methods

Robitzsch, Alexander

doi:10.3389/feduc.2020.589965

Cited by 189 publications

(126 citation statements)

References 83 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Predictors age and TIA were both categorical questions with ranges in the survey, and converted to continuous predictor variables preserving the ordinal ranks [ 49 ] where age groups and TIA categories were labelled in sequence from lowest to highest (0, 1, 2, 3, 4, 5). While the categories for age and TIA may not have perfectly equal spacing, Pasta [ 50 ] and Robitzsch [ 49 ] specify that they can be treated as continuous in almost all cases, even when spacing is not equal across categories. This was done to assess for a trend in the predictor variables age, gender, and TIA for each of the outcome variables.…”

Section: Methodsmentioning

confidence: 99%

Identifying risk factors and implications for beach drowning prevention amongst an Australian multicultural community

2022

View full text Add to dashboard Cite

Multicultural communities in Australia are recognised as a priority area for drowning prevention, but no evidence-based study has addressed their knowledge of beach safety. This study used an online survey tool to identify and examine risk factors relating to swimming ability, beach visitation characteristics and behaviour, and beach safety knowledge of the Australian Southern Asian community to assist in the development of future beach safety interventions. Data was obtained through 249 online and in-person surveys of people aged > 18 years. Most respondents reported poor swimming ability (80%), often swam in in the absence of lifeguards (77%), did not understand the rip current hazard (58%), but reported that they entered the water (76%) when visiting beaches. Close to one-quarter (28%) had not heard, or didn’t know the purpose, of the red and yellow beach flags, which identify lifeguard supervised areas on Australian beaches. Length of time living in Australia is an important beach safety consideration for this community, with minimal differences in terms of gender and age. Those who have lived < 10 years in Australia visit beaches more frequently and are less likely to have participated in swimming lessons, be able to swim, heard of the flags or swim between them, understand rip currents, or have participated in a beach safety program. Very few (3%) respondents received beach safety information from within their own community. The importance of beach safety education and swimming lessons within the Southern Asian community should be prioritised for new and recent migrants to Australia.

show abstract

Section: Methodsmentioning

confidence: 99%

Identifying risk factors and implications for beach drowning prevention amongst an Australian multicultural community

2022

View full text Add to dashboard Cite

show abstract

“…The bivariate normality of pairs of LRVs can be tested against observed ordinal data when at least one of the variables has ≥3 categories [66], but there are no nonnormality corrections for normal-theory estimators that could accommodate nonnormal LRVs, as there are for nonnormal observed indicators [67]. Arguably, assuming continuity of ordinal data rather than assuming their LRVs are normal simply trades off one tenuous assumption for another [68], and the latter can have less predictable consequences.…”

Section: Future Directions For Discrete Datamentioning

confidence: 99%

How to Estimate Absolute-Error Components in Structural Equation Models of Generalizability Theory

Jorgensen

2021

Psych

View full text Add to dashboard Cite

Structural equation modeling (SEM) has been proposed to estimate generalizability theory (GT) variance components, primarily focusing on estimating relative error to calculate generalizability coefficients. Proposals for estimating absolute-error components have given the impression that a separate SEM must be fitted to a transposed data matrix. This paper uses real and simulated data to demonstrate how a single SEM can be specified to estimate absolute error (and thus dependability) by placing appropriate constraints on the mean structure, as well as thresholds (when used for ordinal measures). Using the R packages lavaan and gtheory, different estimators are compared for normal and discrete measurements. Limitations of SEM for GT are demonstrated using multirater data from a planned missing-data design, and an important remaining area for future development is discussed.

show abstract

“…Lastly, it should be mentioned that the analyses were performed on the basis that ordinal variables can be treated as numerical variables. While there are some arguments in favor of exclusively considering these variables as categorical, there is also a strong line supporting dealing with ordinal variables as continuous, according to Robitz [52].…”

Section: Discussionmentioning

confidence: 99%

Vocational Value Profiles of Students with Preferential Vocational Interest in Sport and Their Relationship with Personal and Academic Wellbeing

Álvarez

Ocete

Franco

2021

IJERPH

View full text Add to dashboard Cite

In recent years there has been a significant increase in the number of students who choose to pursue university studies related to the field of sports. However, there are no studies that have investigated the existence of differentiated profiles within students whose preferred vocational area is sport. The main objective of this study was to establish the profiles of students in the second year of Spain Bachillerato whose preferred vocational interest is sport, according to the two representative vocational values: (a) “to have a fun professional activity”; and (b) “to have a professional activity whose schedule makes it possible to reconcile personal and professional life”. In addition, the resulting groups were compared according to their perception of general and academic wellbeing and their identification with the other vocational values. Two hundred and thirty participants (MAge = 17.47; DTAge = 0.669; N = 171; 74.3% male and N = 59; 25.7% female) completed some validated measures. Three clusters emerged which did not differ in terms of general and academic wellbeing. Differences were found though in terms of some vocational values such “to help people”, “to develop one’s entrepreneurial initiative” or “to be self-employed”. The findings invite us to rethink the differences in the specific profiles of vocational values and their impact on employability opportunities, and to consider these approaches in the orientation of students who have among their priority options to study sport sciences.

show abstract

Why Ordinal Variables Can (Almost) Always Be Treated as Continuous Variables: Clarifying Assumptions of Robust Continuous and Ordinal Factor Analysis Estimation Methods

Cited by 189 publications

References 83 publications

Identifying risk factors and implications for beach drowning prevention amongst an Australian multicultural community

Identifying risk factors and implications for beach drowning prevention amongst an Australian multicultural community

How to Estimate Absolute-Error Components in Structural Equation Models of Generalizability Theory

Vocational Value Profiles of Students with Preferential Vocational Interest in Sport and Their Relationship with Personal and Academic Wellbeing

Contact Info

Product

Resources

About