Skip to main content
Download PDF
Download ePub

Rasch analysis of the new general self efficacy scale: an evaluation of its psychometric properties in older adults with low vision

Health and Quality of Life Outcomes volume 22, Article number: 90 (2024) Cite this article

Abstract

Background

While general self-efficacy is known to relate to achievement in many areas, it has rarely been evaluated in individuals with low vision. Here we explore the psychometric properties and targeting of the New General Self Efficacy Scale (NGSES) using Rasch analysis in data collected from older adult clinical trial participants with low vision.

Methods

Participants (n = 121) completed pre-intervention telephone questionnaires (i.e., NGSES, Activity Inventory (AI), Beck Depression Inventory (BDI), SF-36, and the Telephone Interview for Cognitive Status (TICS). Rasch analysis using the Method of Successive Dichotomizations (MSD) was applied to NGSES, AI, and BDI datasets to estimate person and item measures, and ordered rating category thresholds.

NGSES infit mean square statistics and standard errors were analyzed to test whether data fit Rasch model requirements, and targeting was assessed by comparing distributions of person and item measures. Multiple linear regression evaluated the influence of participant characteristics on NGSES person measures.

Results

There was a significant difference (p = 0.01) in the distributions of NGSES person measures (mean = 0.85, range -2.1-3.2) and item measures (mean = 0, range -1.0-0.72). Infit mean square statistics and standard errors for item and person measures conformed to expectations of the Rasch model.

Greater NGSES person measure was related to lower BDI person measure (β = -0.48, p < 0.0005, 95% CI -0.67—-0.28) and higher AI person measure (β = 0.11, p = 0.047, 95% CI 0.0013 – 0.22) after controlling for cognitive status (TICS). Other variables including SF-36, visual acuity, and patient demographics were not related to NGSES person measure.

Conclusions

While NGSES person measures estimated from a low vision patient population conformed to basic Rasch model requirements, the significant differences in the person and item measure distributions indicate poor NGSES targeting, specifically a ceiling effect. Therefore, there is limited discrimination between persons at the upper end of the scale. While further work can evaluate how self-efficacy may be associated with other factors in the low vision patient population, the ceiling effect found in this study cautions interpretation of NGSES results for those with high general self-efficacy.

Trial registration

Pre-registered on clinicaltrials.gov, identifier NCT04926974.

Introduction

As the U.S. population continues to age, the prevalence of age-related eye diseases is increasing, leading to a rise in low vision; i.e., permanent visual impairment that cannot be improved through further intervention [6]. Consequently, addressing the needs of older individuals with low vision has become an important public health priority. Low vision rehabilitation improves vision-related quality of life [2] through prescription of visual assistive devices (such as magnifiers), and comprehensive training of alternative methods to perform daily activities that would otherwise be too visually difficult to perform. For example, low vision rehabilitation for a person reporting difficulty reading medication labels could include use of large print medication labels from the pharmacy, a hand-held magnifier for reading, or use of Braille labels or other adaptive medication marking strategies to allow non-visual medication identification.

General self-efficacy, defined as an individual’s belief in their ability to persevere and perform across a wide range of activities [16], could influence a person’s willingness to seek low vision rehabilitation, and the outcomes of those rehabilitation services. Self-efficacy could also influence other associated factors in the low vision patient population, such as visual ability – an individual’s self-reported ability to perform visually mediated activities [23, 24]. General self-efficacy has previously been evaluated using rating scale questionnaires in educational and organizational settings [1, 7, 15, 32], as well as in elderly adults receiving healthcare services [8, 17, 28, 35, 38]. A body of literature has identified lower general self-efficacy in elderly adults who receive more intensive healthcare services [38]. Given the low vision patient population’s older average age, higher prevalence of chronic health conditions [13], and need for more frequent eye exams, it is plausible that these individuals could share characteristics with elderly adults receiving intensive healthcare services. However, general self-efficacy in older adults with low vision has not been thoroughly evaluated. Previous investigations have only explored general self-efficacy as a potential mediator of fatigue [31] or self-management program outcomes in low vision patients [27]. Mean general self-efficacy raw score [31] and person measure [27] were high in these two studies. A different study compared general self-efficacy in members of the Norwegian Association of the Blind and Partially Sighted versus the general population and found higher general self-efficacy in organization members [5]. While general self-efficacy is not well understood in individuals with low vision (i.e., partially sighted), greater general self-efficacy may improve low vision rehabilitation outcomes by increasing the likelihood of an individual’s acceptance and application of new low vision rehabilitation strategies including visual-assistive aids for use in daily living activities. As such, further exploration of self-efficacy and validation of its measurement in individuals with low vision is an important endeavor.

The majority of the literature on general self-efficacy has utilized raw scores – sums or means of the actual Likert scale ratings – as an outcome measure [15, 33]. When respondents complete the New General Self-Efficacy Scale (NGSES), they respond to 8 items on a 1 to 5 rating scale, and their selections are then summed to produce a raw score that can range from 8 to 40. Raw score outcome measures are problematic because they are not on an equal interval scale [18, 20]. For example, the difference in amount of latent trait, in this case self-efficacy, between a rating of “1” versus a “2” on a Likert scale is not necessarily the same as the difference between a “2” versus a “3.” An equal interval scale requires that each point change on a measurement scale corresponds to the same amount of latent trait [20]. Raw scores cannot account for differences in item difficulty in score (i.e., the most and least difficult items in an instrument impact the raw score in the same way), and presence of missing data can skew outcome measures [18].

Item Response Theory (IRT) and Rasch analysis were developed to resolve issues in the use of raw scores as outcome measures by estimating parameters on an equal interval scale [20]. IRT and Rasch models estimate person measures (i.e., “person-derived score,” the amount of the latent trait possessed by a respondent based on their responses to all items in the instrument), item measures (i.e., “item-derived score,” the amount of latent trait required for a particular item based on the pattern of all persons’ responses to all items in the instrument), and rating category thresholds (i.e., the boundary between rating categories for items) on an equal interval scale measured in units of logit [19, 22].

Historically, IRT models have had certain limitations. Notably, IRT models include an item discrimination parameter which changes the measurement scale for different subsets of items [20]. This means that people responding to different subsets of items will have person measures estimated on slightly different measurement scales, which is problematic for studies wanting to compare differences between groups or persons. In contrast, Rasch analysis always estimates parameters on the same scale [20]. However, Rasch analysis has historically had problems with disordering of thresholds – the boundaries between rating categories which, by definition, should never be disordered. The Method of Successive Dichotomizations (MSD) is both a Rasch and IRT model that estimates parameters on the same scale regardless of the subset of items chosen, and always estimates ordered thresholds [4].

By estimating all parameters on the same equal interval scale, item measures and person measures can also be directly compared to one another to evaluate instrument targeting – whether the range of latent trait assessed by instrument items is consistent with the amount of latent trait possessed by respondents. The distribution of item measures should be similar to the distribution of person measures if targeting is good [3]. An instrument with many items requiring too much or too little latent trait will have poor precision in estimating the amount of latent trait possessed by respondents. As such, evaluation of instrument targeting is important to understand suitability of an instrument for use in a specific population. To our knowledge, targeting has not previously been evaluated in general self-efficacy scales, and has not been explored for self-efficacy in older adults with low vision.

Previous work evaluating general self-efficacy in individuals with low vision has underutilized Rasch analysis. Of studies exploring general self-efficacy as a variable that could influence low vision rehabilitation outcomes [5, 27, 31], only one study applied Rasch analysis to general self-efficacy data, but did not look at targeting or other psychometric properties of the instrument [27]. Most studies evaluating general self-efficacy across any cohort rely on raw score data or IRT analysis.

In this study, we applied Rasch analysis to NGSES data acquired from a sample of older adults with low vision upon enrolling in the Community Access through Remote Eyecare (CARE) randomized clinical trial, which was designed to evaluate low vision patients’ use of visual assistive smartphone applications (apps) following individualized training. We evaluated the psychometric properties of NGSES, its targeting in this cohort of older adults with low vision, and explored how NGSES may be related to other associated factors in this patient population.

Methods

Participants enrolled in a randomized clinical trial to evaluate changes in visual ability with use of visually-assistive smartphone apps between June 2021 and March 2023. Participant recruitment occurred through the New England College of Optometry (NECO) and University of California Los Angeles (UCLA). Data collected for use in this study conformed to the tenets of the Declaration of Helsinki, and approval was obtained through the UCLA Institutional Review Board, which served as the single IRB for this multicenter trial. UCLA recruited participants from the state of California (n = 56 completed baseline measures; 40%) and NECO recruited participants from the New England region, which included Massachusetts, Rhode Island, Connecticut and New Hampshire (n = 84 completed baseline measures; 60%). The study protocol was pre-registered on clinicaltrials.gov (identifier NCT04926974, registration date June 8, 2021) prior to enrolling the first participant.

Study eligibility was as follows: participants were required to be 55 years of age or older with visual acuities between 20/40 and 20/800 or visual field constriction to less than 20 degrees. Participants were required to have a score of greater than or equal to 20 on the Telephone Interview for Cognitive Status (TICS) indicating mild to no cognitive impairment [9], and were naïve to the three study apps they would be randomized to use during the trial (Seeing AI, Aira, and SuperVision +).

Study outcome measures

All study outcome measures included in this analysis were evaluated by telephone questionnaires prior to intervention.

The primary outcome measure in this analysis is person measure estimated from the NGSES, an 8-item instrument asking participants to rate their agreement, on a 1 to 5 scale, with a range of statements reflecting belief in their overall ability to achieve goals. This instrument was developed in a sample of college students, and has previously undergone IRT analysis to evaluate psychometric properties [7]. NGSES is able to accurately discriminate between individuals with different levels of general self-efficacy, capturing similar data to other general self-efficacy scales but using fewer items [32]. Previous use of this instrument in college student populations made use of written or computerized administration [7, 32]. However, these administration modalities may not be feasible in a population of older adults with low vision who may be unable to read and complete hard copy or digital forms. Telephone administration of a general self-efficacy scale has previously been employed in individuals who are blind [5], and was utilized in this study.

Secondary outcome measures in this analysis were acquired by telephone to be explored for potential associations with general self-efficacy. The Telephone Interview for Cognitive Status (TICS) was administered to evaluate cognition [9], and SF-36 was administered to evaluate general wellbeing [10]. Both instruments have been previously administered by telephone to evaluate the influence of comorbidities on low vision rehabilitation outcome measures [12]. Raw score TICS and SF-36 data were used in our analyses as these data are not scored on a Likert scale and are therefore incompatible with Rasch analysis. The Beck Depression Inventory (BDI) [30, 34] and Activity Inventory (AI) [18, 19, 21] are patient-reported outcome measures that have previously been administered via telephone to low vision populations and validated with Rasch analysis. These instruments evaluate depressive symptoms (BDI) and visual ability (AI) – a participant’s self-reported ability to achieve a range of visually-mediated goals and tasks. Rasch analysis using the Method of Successive Dichotomizations has been applied to AI responses to produce item measure and rating category threshold calibrations, which can be used to anchor the scale based on responses of approximately 3,000 low vision patients [11].

We also collected demographic information for the study cohort, which included age, race, gender, self-reported education level, and self-reported residence status (e.g., lives alone, lives with spouse, etc.). Binocular best-corrected visual acuity was measured using letter-by-letter scoring with an ETDRS chart [37].

Method of successive dichotomizations

The Rasch model employed herein, MSD, estimates items on an equal interval scale regardless of the subset of items each person responds to, and always produces ordered thresholds [4]. MSD estimates parameters by dichotomizing the data as many times as there are rating categories (i.e., between rating category 1 and below versus 2 and higher, between rating categories 2 and below versus 3 and higher, etc.). This model was derived based on the mathematical definition of a rating scale, and avoids limitations of other Rasch and IRT models which have problems with threshold disordering or scale invariance [4, 20].

In this work, we applied MSD to NGSES to estimate person measures, item measures, and rating category thresholds using the R package ‘msd’ (R version 4.2.1, R core Team 2022). Standard error, infit and outfit mean square statistics (“infit” and “outfit”) and reliability were also determined for each parameter.

MSD was also applied to BDI data, as well as to AI data. For AI data, item measure and rating category threshold calibrations from previous work were used to anchor the scale to the responses of approximately 3,000 low vision patients [11, 25].

Rasch validation of NGSES

We applied MSD to NGSES data from older adults with low vision, analyzed NGSES parameter precision (measured by standard error) and explored the fit of NGSES data to the Rasch model (captured by infit). As a Rasch model, MSD requires that underlying variability is normally distributed and that variance is identical for each person-item combination [20]. Standard errors and infits of person measures and item measures were examined to test whether these requirements were met. Item infits were transformed to item infit z-scores by Wilson-Hilferty transformation.

NGSES targeting was evaluated by comparing the distributions of item and person measures to ensure that they fall within the same neighborhood of values. Similar person and item measure distributions would indicate that the items were well targeted to the persons, which is necessary for estimate precision. The Kolmogorov-Smirnov test was used to determine whether there was a significant difference in distributions of person measures and item measures.

Finally, a principal component analysis was performed to determine whether NGSES produced unidimensional person measure estimates. An eigenvalue cutoff of 1 was used to determine which factors should be included in the model.

Associations between NGSES and participant characteristics

Relationships were evaluated between NGSES person measures and secondary outcome measures. Associations were initially explored using Pearson correlation coefficients, and Q-Q plots were used to evaluate normality of data. Simple and multiple linear regressions were used to explore relationships between NGSES and other study variables including: AI person measure, BDI person measure, SF-36 subscales, age, TICS score, and best-corrected visual acuity (BCVA).

MSD Rasch analyses were performed using R Statistical Software (v4.2.1; R Core Team 2022) while regressions were performed using Stata statistical software ( v15.1; StataCorp LLC, College Station, TX) and principal component analysis was performed using IBM SPSS Statistics (v29; IBM Corp., Armonk, NY).

Results

Of the 145 individuals recruited for this study, five withdrew prior to data collection, resulting in 140 participants who contributed NGSES data, 19 (14%) of whom were excluded from analysis due to insufficient variance in their responses to allow estimation of person measures with Rasch analysis (e.g., each participant used only one rating category for all items: 15 participants only used rating category 5, three participants only used rating category 4, and one participant only used rating category 3). Table 1 shows demographic information for the 121 participants included in this analysis.

Table 1 Demographic information of participants
Full size table

Mean NGSES raw score was 31.6 (SD 5.7), with only 35 participants (25%) achieving a raw score less than 30, a previously proposed cutoff for low self-efficacy [14]. Raw NGSES scores and MSD-estimated person measures were linearly related and moderately correlated (y = 0.37x + 28.5, R2 = 0.54).

Evaluation of targeting using MSD for NGSES data

The distributions of NGSES item measures and person measures were compared to assess targeting — whether items were appropriate in difficulty level for the range of persons included in this study (Fig. 1) [3]. The person measure distribution (mean = 0.85, range -2.1-3.2) and item measure distribution (mean is 0 by Rasch model convention, range -1.0-0.72) were significantly different (Kolmogorov-Smirnov test p = 0.01). Few item measures were present in the more positive tail of the distribution, which corresponded to greater self-efficacy. The differences in these distributions indicate a ceiling effect, making NGSES poorly targeted for persons with high self-efficacy. Additionally, as shown in Fig. 1, there is a greater proportion of individuals with high general self-efficacy in the sample.

Fig. 1
figure 1

Construct or Wright map of item measures and person measures estimated from New General Self Efficacy Scale

Full size image

Next, item and person measures were plotted relative to their respective standard errors (Fig. 2). Person measure standard errors (mean = 0.70, range 0.40-1.13) followed an approximately u-shaped distribution when plotted relative to person measures, with greater standard errors for more extreme person measures. Item measure standard errors were consistently small (mean = 0.25, range 0.22-0.29) for each item measure. The larger average person measure standard error than average item measure standard error can be explained by the larger number of persons that were used to estimate items than the number of items that were used to estimate persons. Table 2 details mean fit statistics and standard errors, as well as reliabilities for both person and item measures.

Fig. 2
figure 2

Person measures and item measures plotted relative to their standard error

Full size image
Table 2 Fit statistics, standard errors, and reliabilities for NGESES item and person measures
Full size table

Item infit z-scores were plotted relative to their item measure (Fig. 3), and fell within ± 1.5 standard deviations of 0, demonstrating that item measures were well fit to the model. The distribution of person infits was plotted relative to the expected chi-square function with 7 degrees of freedom (Fig. 4). The distribution of person measure infits was significantly different from that expected by the model (Kolmogorov-Smirnov test, P < 0.005), with a greater proportion of infits falling less than or close to the mean of the chi square distribution, and fewer infits falling within the right tail of the distribution, indicating that person measures were better fit than the model anticipates.

Fig. 3
figure 3

Item infit z-score relative to item measure

Full size image
Fig. 4
figure 4

Relative frequency of person infit mean square statistics plotted alongside the expected chi-square distribution

Full size image

Dimensionality of the NGSES instrument was also explored using principal component analysis based on the correlation matrix of raw scores. One component explained 55.4% of the variance, with the next residual component explaining 11% of variance, and all remaining components explaining less than 9% of variance.

General self-efficacy and participant characteristics

Normality of data was confirmed using Q-Q plots. Pearson correlation coefficients were used to evaluate relationships between variables. Significant correlations were observed between NGSES person measures and BDI person measures (r = -0.44, p < 0.001), AI person measures (r = 0.22, p = 0.020), SF-36 fatigue subscale (r = 0.26, p = 0.004), SF-36 emotional wellbeing subscale (r = 0.26, p = 0.005), and SF-36 general health subscale (r = 0.23, p = 0.011). No significant correlation was observed between NGSES and age, TICS score, BCVA, or other SF-36 subscales.

To further explore whether patient characteristics were associated with higher NGSES person measures, multiple linear regression was used. When controlling for TICS scores, higher NGSES person measure was related to lower BDI person measure and higher AI person measure (Table 3, Fig. 5). Non-significant regression factors included: SF36-fatigue (p = 0.98) and SF-36 emotional wellbeing (p = 0.86) subscales, TICS (p = 0.33), BCVA (p = 0.63), college educated (p = 0.97), non-White race (p = 0.96), and self-reported living with a spouse (p = 0.94).

Table 3 Multiple linear regression of New General Self Efficacy Scale person measure with Beck Depression Inventory (BDI) person measure, Activity Inventory (AI) person measure, and TICS raw score
Full size table
Fig. 5
figure 5

Scatterplot of New General Self Efficacy Scale (NGSES) person measure relative to Beck Depression Inventory (BDI) person measure (A) and relative to Activity Inventory (AI) person measure (B)

Full size image

Discussion

Here, Rasch analysis was applied to NGSES data collected from older low vision patients enrolled in a randomized clinical trial. As NGSES was originally developed and validated in college students [7], it was necessary to ensure that the instrument is appropriate to use in our patient sample of older adults with low vision. In our study, person measures, item measures, and rating category thresholds were estimated with good fit to the Rasch model. However, we detected a significant difference in the distributions of item measures and person measures estimated from our sample, indicating a ceiling effect. This ceiling effect is caused by the insufficient number of items that require high levels of general self-efficacy included in NGSES. In Rasch analysis, item measures and person measures are estimated conjointly based on one another [20]. As such, insufficient item measures in a particular neighborhood of values yields a lack of precision for person measures in that region. This is reflected by the higher standard errors for persons with the highest NGSES scores seen in Fig. 2. As NGSES person measures were, overall, high in our patient sample, indicating higher levels of self-efficacy, it is particularly troublesome that there are insufficient items in this range of values. This finding of poorer precision in person measure estimates for those with very high general self-efficacy is consistent with previous research [32], but here, our use of MSD allows direct comparison of person and item measures to clearly evaluate the cause of this problem: poor instrument targeting due to a ceiling effect.

Few prior studies have evaluated general self-efficacy in a low vision patient population [5, 27, 31]. Direct comparison of our results to other studies in individuals with low vision is limited by their use of the Schwarzer and Jerusalem 10-item General Self Efficacy Scale [15] rather than the NGSES [7], as well as their use of classic test theory. In contrast to raw scores, our MSD-estimated person measures are on an equal interval scale. The one prior study that evaluated general self-efficacy in low vision patients using Rasch analysis investigated factors predicting outcomes with a novel self-management program, and did not find general self-efficacy to be a significant associated factor [27]. Another study explored factors related to fatigue in individuals with low vision, and found an indirect effect from general self-efficacy raw scores that was mediated by depression [31]. An additional study compared raw general self-efficacy scores between members of a Norwegian blindness organization and the general population, finding higher general self-efficacy in blind individuals [5]. While these studies make use of a different general self-efficacy instrument thereby preventing direct comparison of their results to ours, our finding of high general self-efficacy in older adult low vision patients is consistent with these studies.

Our work is also consistent with previous literature identifying a relationship between self-efficacy and depression [29]. While low vision patients are known to have a higher prevalence of depression than the general population [26, 36], our particular sample had a low prevalence of depressive symptoms as indicated by the negative average BDI person measure: the majority of our sample had no or mild depressive symptoms. Despite the limited range of depressive symptoms in our sample, the strong negative correlation between depressive symptomatology and general self-efficacy in this study (i.e., fewer depressive symptoms in those with higher general self-efficacy) is consistent with previous work.

Additionally, a significant relationship was observed between participants’ visual ability – their self-reported ability to perform visually-mediated activities (i.e., AI results) – and general self-efficacy despite the lack of a significant relationship between visual acuity and general self-efficacy. Individuals with greater visual ability have less difficulty performing daily activities, which may be related to better vision or pre-existing mastery of low vision rehabilitation strategies that make activities less difficult to perform. Successful use of existing low vision rehabilitation strategies could be related to high self-efficacy, and is an area worthy of future study. Previous work has not evaluated relationships between visual ability and self-efficacy, although one study [27] did evaluate both variables.

A limitation of this study is that participants were enrolled in a larger randomized clinical trial evaluating use of new smartphone apps. Clinical trial participation may introduce a self-selection bias, selected for individuals with higher general self-efficacy. Reassuringly, patient demographics in this study were consistent with those reported from a large multi-center low vision outcomes study [13]. We also only recruited participants who were 55 years of age or older, so results cannot be generalized to younger individuals with low vision. Future work could evaluate general self-efficacy in all patients presenting to low vision clinic. Longitudinal outcomes related to general self-efficacy before and after an intervention could also be the focus of future work.

Conclusions

In this study, Rasch analysis was used to estimate valid person measures, item measures, and rating category thresholds in a sample of older adults with low vision participating in a clinical trial. This work explored the psychometric properties of the NGSES in low vision patients, and identified a ceiling effect which precludes reliable interpretation of outcomes for individuals with high general self-efficacy. Further work could validate additional NGSES items or a modified rating scale to improve targeting by adding more difficult items or increasing the difficulty of existing items. However, in its present state NGSES does not produce reliable measurements for low vision clinical trial participants with high self-efficacy.

Availability of data and materials

No datasets were generated or analysed during the current study.

References

  1. Bandura A, Barbaranelli C, Caprara GV, Pastorelli C. Multifaceted impact of self-efficacy beliefs on academic functioning. Child Dev. 1996;67(3):1206–22.

    Article  CAS  PubMed  Google Scholar 

  2. Binns AM, Bunce C, Dickinson C, Harper R, Tudor-Edwards R, Woodhouse M, Linck P, Suttie A, Jackson J, Lindsay J. How effective is low vision service provision? A systematic review. Surv Ophthalmol. 2012;57(1):34–65.

    Article  PubMed  Google Scholar 

  3. Bond TG, Fox CM. Applying the Rasch model: fundamental measurement in the human sciences. Psychology Press; 2013.

  4. Bradley C, Massof RW. Method of successive dichotomizations: An improved method for estimating measures of latent variables from rating scale data. PLoS One. 2018;13(10):e0206106.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Brunes A, Hansen MB, Heir T. General self-efficacy in individuals with visual impairment compared with the general population. PLoS One. 2021;16(7):e0254043. https://doiorg.publicaciones.saludcastillayleon.es/10.1371/journal.pone.0254043.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Chan T, Friedman DS, Bradley C, Massof R. Estimates of incidence and prevalence of visual impairment, low vision, and blindness in the United States. JAMA Ophthalmol. 2018;136(1):12–9.

    Article  PubMed  Google Scholar 

  7. Chen G, Gully SM, Eden D. Validation of a new general self-efficacy scale. Organ Res Methods. 2001;4(1):62–83.

    Article  CAS  Google Scholar 

  8. Chirico A, Lucidi F, Merluzzi T, Alivernini F, De Laurentiis M, Botti G, Giordano A. A meta-analytic review of the relationship of cancer coping self-efficacy with distress and quality of life. Oncotarget. 2017;8(22):36800.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Cook SE, Marsiske M, McCoy KJ. The use of the Modified Telephone Interview for Cognitive Status (TICS-M) in the detection of amnestic mild cognitive impairment. J Geriatr Psychiatry Neurol. 2009;22(2):103–9.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Framework IC. The MOS 36-item short-form health survey (SF-36). Med Care. 1992;30(6):473–83.

    Article  Google Scholar 

  11. Gobeille M, Bradley C, Goldstein JE, Massof R. Calibration of the Activity Inventory item bank: a patient-reported outcome measurement instrument for low vision rehabilitation. Transl Vis Sci Technol. 2021;10(6):12–12.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Goldstein JE, Chun MW, Fletcher DC, Deremeik JT, Massof RW, Group LVRNS. Visual ability of patients seeking outpatient low vision services in the United States. JAMA Ophthalmol. 2014;132(10):1169–77.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Goldstein JE, Massof RW, Deremeik JT, Braudway S, Jackson ML, Kehler KB, Primo SA, Sunness JS, Group LVRNS. Baseline traits of low vision patients served by private outpatient clinical centers in the United States. Arch Ophthalmol. 2012;130(8):1028–37.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Haugland T, Wahl AK, Hofoss D, DeVon HA. Association between general self-efficacy, social support, cancer-related stress and physical health-related quality of life: a path model study in patients with neuroendocrine tumors. Health Qual Life Outcomes. 2016;14:1–7.

    Article  Google Scholar 

  15. Jerusalem M, Schwarzer R. Self-efficacy as a resource factor in stress appraisal processes. In Self-efficacy (pp. 195-214). Taylor & Francis. 2014.

  16. Judge TA, Erez A, Bono JE. The power of being positive: the relation between positive self-concept and job performance. Hum Perform. 1998;11(2–3):167–87.

    Article  Google Scholar 

  17. Li X, Yang K, An Y, Liu M, Yan C, Huang R. General self-efficacy and frailty in hospitalized older patients: the mediating effect of loneliness. Geriatr Nurs. 2022;48:308–12.

    Article  Google Scholar 

  18. Mallinson T. Why measurement matters for measuring patient vision outcomes. Optom Vis Sci. 2007;84(8):E675–82.

    Article  Google Scholar 

  19. Massof RW. Application of stochastic measurement models to visual function rating scale questionnaires. Ophthalmic Epidemiol. 2005;12(2):103–24.

    Article  PubMed  Google Scholar 

  20. Massof RW. Understanding Rasch and item response theory models: applications to the estimation and validation of interval latent trait measures from responses to rating scale questionnaires. Ophthalmic Epidemiol. 2011;18(1):1–19.

    Article  PubMed  Google Scholar 

  21. Massof RW, Ahmadian L, Grover LL, Deremeik JT, Goldstein JE, Rainey C, Epstein C, Barnett GD. The Activity Inventory: an adaptive visual function questionnaire. Optom Vis Sci. 2007;84(8):763.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Massof RW, Bradley C. An adaptive strategy for measuring patient-reported outcomes: incorporating patient preferences relevant to cost-benefit assessments of vision rehabilitation. Person Center Outcome Metrol. 2023;107:107–50.

  23. Massof RW, Hsu CT, Baker FH, Barnett GD, Park WL, Deremeik JT, Rainey C, Epstein C. Visual disability variables. I: the importance and difficulty of activity goals for a sample of low-vision patients. Arch Phys Med Rehabil. 2005;86(5):946–53.

    Article  PubMed  Google Scholar 

  24. Massof RW, Hsu CT, Baker FH, Barnett GD, Park WL, Deremeik JT, Rainey C, Epstein C. Visual disability variables. II: the difficulty of tasks for a sample of low-vision patients. Arch Phys Med Rehabil. 2005;86(5):954–67.

    Article  PubMed  Google Scholar 

  25. Massof RW, Stelmack JA. Interpretation of low vision rehabilitation outcomes measures. Optom Vis Sci. 2013;90(8):788.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Parravano M, Petri D, Maurutto E, Lucenteforte E, Menchini F, Lanzetta P, Varano M, van Nispen RM, Virgili G. Association between visual impairment and depression in patients attending eye clinics: a meta-analysis. JAMA Ophthalmol. 2021;139(7):753–61.

    Article  PubMed  Google Scholar 

  27. Rees G, Xie J, Chiang PP, Larizza MF, Marella M, Hassell JB, Keeffe JE, Lamoureux EL. A randomised controlled trial of a self-management programme for low vision implemented in low vision rehabilitation services. Patient Educ Couns. 2015;98(2):174–81. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.pec.2014.11.008.

    Article  PubMed  Google Scholar 

  28. Remm S, Halcomb E, Hatcher D, Frost SA, Peters K. Understanding relationships between general self-efficacy and the healthy ageing of older people: an integrative review. J Clin Nurs. 2023;32(9–10):1587–98.

    Article  PubMed  Google Scholar 

  29. Remm SE, Halcomb E, Peters K, Hatcher D, Frost SA. Self-efficacy, resilience and healthy ageing among older people who have an acute hospital admission: a cross-sectional study. Nurs Open. 2023;10(11):7168–77.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Sauer S, Ziegler M, Schmitt M. Rasch analysis of a simplified Beck Depression Inventory. Personal Individ Differ. 2013;54(4):530–5.

    Article  Google Scholar 

  31. Schakel W, Bode C, Van De Ven PM, Van Der Aa HP, Hulshof CT, Van Rens GH, Van Nispen RM. Understanding fatigue in adults with visual impairment: a path analysis study of sociodemographic, psychological and health-related factors. PLoS One. 2019;14(10):e0224340.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Scherbaum CA, Cohen-Charash Y, Kern MJ. Measuring general self-efficacy: a comparison of three measures using item response theory. Educ Psychol Measur. 2006;66(6):1047–63.

    Article  Google Scholar 

  33. Sherer M, Adams C. The self-efficacy scale: a construct validity study. 1983.

    Google Scholar 

  34. Siegert RJ, Tennant A, Turner-Stokes L. Rasch analysis of the Beck Depression Inventory-II in a neurological rehabilitation sample. Disabil Rehabil. 2010;32(1):8–17.

    Article  PubMed  Google Scholar 

  35. Torrisi M, De Cola MC, Buda A, Carioti L, Scaltrito MV, Bramanti P, Manuli A, De Luca R, Calabro RS. Self-efficacy, poststroke depression, and rehabilitation outcomes: is there a correlation? J Stroke Cerebrovasc Dis. 2018;27(11):3208–11.

    Article  PubMed  Google Scholar 

  36. van der Aa HP, Comijs HC, Penninx BW, van Rens GH, van Nispen RM. Major depressive and anxiety disorders in visually impaired older adults. Invest Ophthalmol Vis Sci. 2015;56(2):849–54.

    Article  PubMed  Google Scholar 

  37. Vanden Bosch ME, Wall M. Visual acuity scored by the letter-by-letter or probit methods has lower retest variability than the line assignment method. Eye. 1997;11(3):411–7.

    Article  PubMed  Google Scholar 

  38. Whitehall L, Rush R, Górska S, Forsyth K. The general self-efficacy of older adults receiving care: a systematic review and meta-analysis. Gerontologist. 2021;61(6):e302–17.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

We would like to thank Chris Bradley PhD for his contribution to this work. We would also like to acknowledge the following research assistants for their work in administering baseline study questionnaires: Bridget Peterson, Meghan Knizak, Chris Yeung, Priyanshi Patel, Erika Pacheco, Phoebe Hu, Jewel Chu, Sarah Zoe Bui, Joyce Kuo, Pia Gentapanan, Angelina Chen, Benjamin Zietz, and Cindy Pabla.

Funding

Funding for this trial was provided by NIDILRR Grant 90DPGE0012-02-01, NIH T35EY007149, and support to AKB from an unrestricted award from Research to Prevent Blindness to the Department of Ophthalmology at UCLA.

Author information

Authors and Affiliations

  1. New England College of Optometry, 424 Beacon St., Boston, MA, 02115, USA

    Micaela Gobeille, Alexis G. Malkin, Jeffrey Ho, Cecilia Idman-Rait & Nicole C. Ross

  2. Department of Ophthalmology, University of California, Stein Eye Institute, 100 Stein Plaza, Los Angeles, CA, USA

    Ava K. Bittner & Max Estabrook

  3. Department of Psychology, Northeastern University, 360 Huntington Ave, Boston, MA, USA

    Nicole C. Ross

Authors
  1. Micaela Gobeille
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Ava K. Bittner
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Alexis G. Malkin
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Jeffrey Ho
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Cecilia Idman-Rait
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Max Estabrook
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Nicole C. Ross
    View author publications

    You can also search for this author inPubMed Google Scholar

Consortia

for the CARE Study Team

Contributions

MG: data analysis and interpretation, manuscript drafting, manuscript revision. AB: study design, data acquisition, data analysis and interpretation, manuscript drafting, manuscript revision. AM: study design, data acquisition, manuscript drafting, manuscript revision. JH: data acquisition, manuscript drafting, manuscript revision. CR: study design, data acquisition, manuscript drafting, manuscript revision. ME: data acquisition, manuscript drafting, manuscript revision. NR: study design, data acquisition, data analysis and interpretation, manuscript drafting, manuscript revision.

Corresponding author

Correspondence to Micaela Gobeille.

Ethics declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board at UCLA. All participants provided informed consent prior to enrolling.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Gobeille, M., Bittner, A.K., Malkin, A.G. et al. Rasch analysis of the new general self efficacy scale: an evaluation of its psychometric properties in older adults with low vision. Health Qual Life Outcomes 22, 90 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12955-024-02306-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12955-024-02306-2

Keywords

Health and Quality of Life Outcomes

ISSN: 1477-7525

Contact us