Likert scale versus the visual analogue scale in evaluating dentofacial aesthetics: a systematic review
Publicado en línea: 05 jun 2024
Páginas: 158 - 168
Recibido: 01 sept 2023
Aceptado: 01 abr 2024
DOI: https://doi.org/10.2478/aoj-2024-0010
© 2024 Hanis Nabilah Marzuki et al., published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.
The Likert scale is a rating system used in questionnaires and is designed to measure a respondent’s attitudes, opinions or perceptions.1 In health research studies, the Likert scale is a widely utilised scaling technique whose response could range on a scale point of 5 to 7. Possible ranges of responses for a 5-point Likert scale are rated from “strongly agree”, “agree”, “neutral”,” disagree” and “strongly disagree”.1 There are varying viewpoints on the ideal number of response options and the wording of the possible answers has an impact on the responses.2 Likert scales are an ordinal-level measure rather than an interval-level measure, in that the response categories have a rank order but the intervals between values cannot be assumed to be equal.3
In 1921, the Visual Analogue Scale (VAS) was first described and referred to as a “graphic rating method”. Respondents indicated their perceived status by placing a mark along a horizontal line at the most relevant position. A VAS is commonly displayed as a 10 cm long horizontal line, anchored by two verbal descriptors at the extremes.4 The orientation of a VAS may be either horizontal or vertical, and the choice between the two styles should not impact the survey outcome. However, the orientation of the scale may deviate from the respondent’s likely point of view as a horizontal VAS and a vertical 10-cm VAS have shown different reproducibility results.3
Both Likert scales and VAS have their individual advantages and disadvantages that have been previously described.5–10 VAS has been claimed to have better responsiveness than surveys which have used a Likert scale.11 Researchers claim that a Likert scale and a VAS yield a similar result despite being different but comparable in reliability and validity. The disadvantage of a VAS is that it requires more dedication, time and effort to educate respondents who have difficulty in understanding the instructions. The markings of a VAS may also be non-specific and uninterpretable which may be less accurate than using a scale method.
The Likert scale has been noted for its simplicity of use and ease of comprehension as it uses numerical coding for interpretation, which is preferred compared to the VAS.12 Children appear to understand the Likert scale better as the system can facilitate communication more easily than the VAS and is able to discriminate between the maximal intensities.2 A Likert bias can be reduced by increasing the number of response categories.
Few criteria have been used to compare the different types of VAS and Likert scales, but include the consistency of adjectives and a scale ratings relationship, the magnitude of between-subject variability, test-retest reliability, correlation strength with other measurements, factor analysis performance, and achieving a uniform response distribution.2 The consequences of differences in these criteria have caused variations in the conclusion of the method’s relative value. Grant et al. (1999) ascertained that the VAS had greater sensitivity than the Likert scale but there were no significant differences between the two.13
In the dental setting, the Likert scale and VAS have been widely used to measure patient satisfaction regarding treatment, pain, anxiety and more. During orthodontic treatment specifically, patients may evaluate the satisfaction of their changing dentofacial appearance, as in contemporary times, dentofacial aesthetics has become a major concern.14 Clinically, the commonly used Likert scale and VAS rating scale are important and are applied to determine the patient’s perspective and expectation towards treatment. Many studies have been conducted to evaluate the Likert scale and VAS but the results remain inconclusive. However, a study by Dourado et al. (2021) showed that respondents preferred to use the Likert scale compared to the VAS in rating facial attractiveness.15
It is apparent that research comparing the benefits and disadvantages of using the VAS and Likert scales have frequently revealed inconsistent results.16 It has been indicated that the Likert scale is easier to use compared to the VAS while others report the opposite.16 Therefore, the present systematic review aimed to:
identify the validity and reliability of both the Visual Analogue Scale (VAS) and Likert scale in evaluating responses from respondents regarding their perspective of dentofacial aesthetics. identify which scale is preferable for use by clinicians and patients in assessing dentofacial aesthetics.
By comparing the VAS and Likert scales for their advantages and disadvantages in obtaining accurate responses, clinicians can discern the most appropriate scale for incorporation into research questionnaires.
The present systematic review was conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). The review was registered in Prospero with the ID: CRD42023479854.
The eligibility criteria were based on the PICO format summarised in Table I. The present review included studies that recruited lay people and/or dental clinicians as the study population. The type of intervention was the use of the Likert Scale or VAS for comparison in measuring facial aesthetics. In addition, the present systematic review sought studies that reported outcomes related to the validity, reliability, and patient or clinician’s preference between the Likert scale and VAS. The included studies were limited to randomised controlled trials and cross-sectional studies, with no time or language limit. Reviews, case reports, letters to the editor, conference abstracts, meeting abstracts, oral and poster presentations, and all other non-peer-reviewed articles were excluded.
Eligibility criteria
| Criteria | Inclusion criteria | Exclusion criteria |
|---|---|---|
| Population | Laypeople and/or dental clinicians | |
| Intervention | Likert Scale and VAS used in assessing facial aesthetic | |
| Outcome measures | 1.Validity and/or reliability of both scale |
|
| Study design | Randomized controlled trial, cross sectional study | Reviews, case reports, letters to editor, and all other non-peer-reviewed articles |
| Others | All language article |
The database search and identification of studies were conducted independently by two authors (M.H.N. and A.S.I.Z.) in duplicates, up to the 30 November 2023. Any disagreement between the two authors was resolved by discussion with the third author (A.A.). Relevant studies were screened by reviewing the titles and abstracts of all studies identified from six electronic databases: PubMed, Cochrane Central Register of Controlled Trials, Scopus, Web of Science, Ovid, and OpenGrey. A combination of Keywords and MESH terms: (“Likert scale” OR “Likert”) AND (“visual analogue scale” OR “VAS”) AND (“aesthetic*” OR “facial*” OR “dental*” OR “dentist*”) were applied in the electronic database search, with no time or language limit. Furthermore, the reference lists of all the included studies were screened for suitable additional studies to be included in this review. Corresponding authors were contacted to obtain clarifications when necessary.
After database searches and the identification of potential studies, the full texts of identified studies were obtained. The selection of studies was conducted independently by two authors (M.H.N. and A.S.I.Z.) in duplicates, and any disagreement was resolved by discussion with the third author (A.A.) by reviewing the full texts.
Data were extracted from the full text of the selected studies using a standardised data extraction sheet that was designed specifically for the present systematic review. The findings related to study methodology, study population, intervention characteristics, and outcome measure(s) were extracted and are summarised in Table III.
Study methodology, population, intervention characteristics, and outcome measure(s)
| Study | Study Methodology | Study Population | Intervention Charateristics | Outcome Measure(s) |
|---|---|---|---|---|
| Fudalej et al. (2017)15 | Randomized controlled trial | Eight junior orthodontic residents without any experience in treatment of CLP. |
Extra-oral images of 60 non-syndromic complete unilateral cleft lip and palate (CLP) patients were taken from the frontal and profile views and were cropped. |
Reliability was measured by using intraclass correlation coefficient (ICC) in assessing nasolabial appearance. |
| Asik & Kök (2021)13 | Cross-sectional | 195 participants (52 laypersons, 50 patient’s - relatives, 51dentists, and 42 orthodontists) from the Necmettin Erbakan University Faculty of Dentistry, Department of Orthodontics. |
Photographs of a female frontal posed smile were cropped to include only mouth and teeth produced a base image. Dental midline was adjusted 1, 2, 3, and 4mm to the left and right sides using Adobe Photoshop CC 2014 software resulting in 8 modified smile images. |
Effect of dental midline deviation on the perception of smile aesthetics by orthodontist, dentists, patient relatives, and layperson. |
| Hatch et al. (2017)12 | Cross-sectional | Ten university students and staff (5 females and 5 males) identified from convenience sampling approach | Frontal and lateral views of images of 313 adults in Iowa were presented to the raters. |
Reliability of both scales by using test-retest reliability in measuring overall facial attractiveness. |
| Dourado et al. (2021)8 | Cross-sectional | 90 evaluators divided into 3 groups equally: Lay people (20-67 years old) |
Evaluation on facial pleasantness was done on 10 adult patients’ photographs that underwent orthodontic treatment for different facial discrepancies compiled in an album. |
Reliability of Likert scale and VAS was measured through intra examiner error analysis. |
| Eslamipour et al. (2017)14 | Cross-sectional | 993 freshman students from Isfahan University were randomly selected and examined to assess the dental health component (DHC) index of the Index of orthodontic treatment need (IOTN) Examination was done by two examiners. | Questionnaire divide into three sections: Demographic, index of oral aesthetic subjective index scale (OASIS) using 7-point Likert scale, VAS and IOTN. | Determination of most reliable self-perceived indices by using intraclass coefficients. |
All risk of bias assessments were conducted by two authors (M.H.N. and A.S.I.Z.) independently, in duplicates. Any disagreement between these two authors was resolved by discussion with the third author (A.A.).
The risk of bias assessment of the cross-sectional study was conducted using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist. STROBE is a guideline for the report of observational research such as cohort, case-control, and cross-sectional studies. The checklist consists of 22 items that are divided into six categories: (1) the article’s title and abstract, (2) introduction, (3) methods, (4) results, (5), discussion, (6) other information. The guideline helps authors to ensure that selected research papers were of high quality.17 The risk of bias assessment of a randomised controlled trial was assessed using Version 2 of the Cochrane Risk of Bias Tool for Randomised Trials (RoB 2) as shown in Table II. RoB 2 is structured into a series of fixed bias regions that focus on different aspects of the study design, conduct, and reporting. Within each domain, a series of questions (“signaling questions”) aimed to gather information about study characteristics relevant to the risk of bias. Based on the answers to the signalling questions, the algorithm generates suggested judgments regarding the risk of bias attributable to each domain. Verdicts may be at ‘low’ or ‘high’ risk of bias, or express ‘some degree of concern’.18
Risk of bias (RoB) with STROBE checklist
| STROBE | |||
|---|---|---|---|
| Paper | Study | Score | Paper |
| Perception of dental midline deviation and smile attractiveness by eye-tracking and aesthetic ratings | Aşik & Kök (2021)13 | 78.5% | Good |
| Effects of objective 3D measures of facial shape and symmetry on perceptions of facial attractiveness | Hatch et al. (2017)12 | 75.3% | Good |
| Likert scale vs visual analog scale for assessing facial pleasantness | Dourado et al. (2021)8 | 85.18% | Excellent |
| Correlation coefficients of three self-perceived orthodontic treatment need indices | Eslamipour et al. (2017)14 | 85.18% | Excellent |
Figure 1 shows the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram of this systematic review. The six electronic database searches identified 517 potentially relevant articles. After removing 30 duplicates, 487 papers were screened. A total of 449 articles were excluded as they did not meet the inclusion and exclusion criteria. The full texts of the remaining 38 articles were successfully retrieved and assessed for eligibility. A total of 33 papers were subsequently excluded as they did not match the eligibility criteria. Finally, a total of five research papers were included in the systematic review (one randomised controlled trial and four cross-sectional studies).
Figure 1.
Study selection according to the Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flowchart.
Table II and Figure 2 show the risk of bias results of the articles included in the present systematic review.
Figure 2.
Risk of bias according to the Cochrane Risk of Bias Tool for Randomised Trials (RoB 2).
Table III shows the study methodology, population, intervention characteristics, and outcome measure(s) of the included articles.
All of the studies included in the present systematic review presented high heterogeneity which enabled a qualitative review of the use of the Likert scale and VAS in evaluating dentofacial aesthetics. The research outcomes from the papers are summarised in Table IV.
Summary of research outcomes
| Study | Title | Outcomes |
|---|---|---|
| Aşik & Kök (2021)13 | Perception of dental midline deviation and smile attractiveness by eye-tracking and aesthetic ratings | The ICC was 0.805 for the VAS, and 0.760 for the Likert Scale No statistically significant difference between the genders VAS and Likert scores were consistent with the eye-tracking data. No statistically significant difference between the age groups. |
| Hatch et al. (2017)12 | Effects of objective 3D measures of facial shape and symmetry on perceptions of facial attractiveness | The test-retest reliability for overall facial attractiveness ratings was 0.69 on the Likert-scale and 0.75 on the VAS. |
| Dourado et al. (2021)8 | Likert scale vs visual analog scale for assessing facial pleasantness | Intra examiner errors were 0.70 and 0.69 for the Likert scale and VAS, respectively. |
| Eslamipour et al. (2017)14 | Correlation coefficients of three self-perceived orthodontic treatment need indices | The intraclass coefficient for each examiner was over 0.9. |
| Fudalej et al. (2017)15 | Comparison of three methods of rating nasolabial appearance in cleft lip and palate | The VAS method was in turn more reproducible than Likert scale. |
Based on Hatch et al. (2017), the test-retest reliability of the Likert scale and VAS was 0.69 and 0.75, respectively, for overall facial attractiveness ratings which indicated overall acceptable reliability.19 Aşik and Kök (2021) found that, for dental midline deviation and smile attractiveness, the scores of both the Likert scale and VAS were consistent.20 The intraclass correlation coefficient values for the VAS and Likert scales were 0.805 and 0.76, respectively.20
Reliability can also be measured by using an intra-examiner error analysis. In a study evaluating facial pleasantness, the weighted kappa test statistic was applied to the Likert Scale and intraclass correlation coefficient for the VAS to determine intra-examiner reliability. The result showed moderate intra-examiner agreement for the VAS (0.69) and Likert scale (0.70).15 An additional study conducted using intra-examiner reliability testing on a VAS showed acceptable results of an intraclass correlation coefficient which was over 0.90.21
A study conducted by Fudalej et al. (2017) evaluating the nasolabial aesthetics of patients presenting with a unilateral cleft lip and palate (CLP) showed that the VAS was more consistent than the aesthetic index using a Likert scale in rating nasal shape (
When relating gender to validity and reliability scoring, Aşik and Kök (2021) reported that there were no statistically significant differences between genders and ages, observed in both the VAS and the Likert scale evaluations of the effects of midline deviation on smile aesthetic perception.20 Daurado et al. (2021) and Eslamipour et al. (2017) also reported that, regardless of gender and age, the evaluators were able to similarly assess using either method.15,21
In a study conducted to assess facial pleasantness, Dourado et al. (2021) showed that 75% of the examiners, consisting of laypeople, senior orthodontists, and senior oral and maxillofacial surgeons, preferred the Likert scale over the VAS.
The claims made by the respective examiners were that the Likert scale had a better representation of their opinions and was easier and simpler to use.15
However, a further study reported that the VAS had a higher sensitivity (44.6%) compared to the Likert Scale (22.3%), namely the ‘Aesthetic Components of The Index Orthodontic Treatment Need’ and ‘Oral Aesthetic Subjective Index Scale’.
The VAS and Likert Scales have been widely used in evaluating dentofacial aesthetics and much research has been conducted on comparing the Likert Scale and VAS as scoring systems.10,12,23–26 Considerations in choosing between these scales include the subjective nature inherent in both, potential variations in individual interpretations, and respondent comfort with either format. The choice may also hinge on the study objectives, with the VAS being preferred for a more detailed and continuous measurement. Cultural factors, individual preferences, and the research tradition within the field should also influence this decision. Therefore, the present systematic review was conducted to identify the reliability, validity, and preference of clinicians and patients for both the Likert scale and VAS in the evaluation of dentofacial aesthetics. However, due to the heterogeneity of the included studies, a meta-analysis could not be conducted.
Previous research studies have indicated diverse researcher preferences regarding the suitability of the Likert scale or VAS for research evaluation. The preference between Likert scales and VAS among clinicians and laypeople is shaped by distinct factors tied to their professional backgrounds and familiarity with different rating systems. Clinicians, accustomed to precise measurements in their professional training, may gravitate toward the VAS due to its continuous nature, which enables finer gradations. Conversely, laypeople, who often encounter Likert scales in non-clinical assessments, may find them more accessible and user-friendly due to their simplicity and ease of comprehension. However, the present systematic review found that both clinicians and laypersons preferred the Likert scale over the VAS, as it was perceived that the Likert scale had a better representation of their opinions, and was easier and simpler to use to assess facial pleasantness.15
Dourado et al. reported that some participants required 5 to 10 minutes or longer to comprehend the VAS.15 Nevertheless, the VAS was considered to have greater sensitivity compared to the Likert Scale. This finding is in agreement with Eslamipour et al. (2017) who asserted that the VAS has higher sensitivity and serves as an accessible and comprehensible index across diverse social strata.21 Furthermore, Fudalej et al. (2017) chose the VAS over the Likert scale in an assessment of the nasolabial region ratings in cleft lip and palate (CLP) research, as it was argued that the Likert scale lacked reproducibility and yielded variable results within the scoring system. Conversely, the VAS method was deemed more reproducible and preferable by the researchers. However, this claim is not substantiated nor supported by the present systematic review.22 From the perspectives of data collection and analysis, the Likert scale is more straightforward than the VAS because the latter requires additional time to measure the extent of the generated line.
Aesthetics is commonly regarded as both subjective and universal, often assuming that experts comprehend it better than individuals. However, interventions and treatments might be validated based on individual preferences. When evaluating their own aesthetic preferences and necessities, patients should be regarded as experts. The assessment of any deformity by a professional provides another rationale for treatment, although it may not always be essential if the individual does not perceive the deformity as affecting their quality of life. What appears as a discrepancy to dental professionals might be perceived as normal by the layperson. It is crucial to exercise caution against medicalising the spectrum of normalcy, by advocating treatment in the absence of pathology and without the patient’s self-perceived need.
The included studies in this systematic review recruited either university students, young adults, and/or the elderly as their study subjects. The current systematic review found that age groups did not affect the results of the dentofacial aesthetics evaluation on either the Likert scale or the VAS. However, none of the included studies recruited children in the evaluation of dentofacial aesthetics. Children have been reported to prefer the Likert scale and find it easier to complete compared to the VAS.27 The relation between age and preference for VAS and Likert scales is multifaceted, involving cognitive, linguistic, and experiential factors. Researchers and educators should consider these factors when selecting assessment tools for different age groups, and aim for formats that align with the developmental stage and preferences of the target population. Further studies on the preference of individuals from different age groups on the tool used to evaluate dentofacial aesthetics are warranted.
A limitation of this review was that the reviewed papers were taken from many different areas and parts of the world. Therefore, the epidemiology of the conducted research might affect the research results. The level of knowledge and the examiners involved in all the research papers were varied. Preferred scoring systems cannot be concluded as the results from each of the papers were varied.
The papers that have been reviewed were also limited. There were a lack of resources comparing the Likert Scale and VAS from a dentofacial aesthetics perspective. Only five papers were identified for the comparative use of the assessment scales. More online databases could have been explored to find additional related papers. The difficulty in finding the appropriate MESH words is a possible reason for the inability to obtain optimal search papers. Furthermore, some papers had no direct comparison between the Likert scale and VAS which made the interpretation of these papers difficult.
Aspects of aesthetic evaluation were performed differently between the included research papers which generated inconsistent and variable results. From the present review, it was concluded that the Likert Scale and VAS could be used interchangeably based on the suitability of the scoring system for reliable results of the aesthetics evaluation.
Lastly, the heterogeneity of the included studies precluded the possibility of conducting a meta-analysis.
Both the Likert Scale and VAS are valuable tools for the evaluation of dentofacial aesthetics. While both systems have consistently yielded favourable results in objective evaluations of dentofacial aesthetics across the included studies, reaching a consensus on the superiority of one method over the other proved challenging due to heterogeneity which precluded a quantitative analysis and meta-analysis. Therefore, further research is indicated to decisively identify the strengths and weaknesses of these two methods in various applications and considerations within dentofacial aesthetics. Additionally, there is an urgent need for the ongoing development and evaluation of dentofacial aesthetic questionnaires, given the relatively limited current research and literature in this domain.