RESEARCH METHODOLOGY
Evaluating the psychometric properties of the Jacelon Attributed Dignity Scale Cynthia S. Jacelon & Jeungok Choi Accepted for publication 25 January 2014
Correspondence to C.S. Jacelon: e-mail: [email protected] Cynthia S. Jacelon PhD RN FAAN Associate Professor College of Nursing, University of Massachusetts Amherst, Amherst, MA, USA Jeungok Choi PhD MPH RN Associate Professor College of Nursing, University of Massachusetts Amherst, Amherst, MA, USA
J A C E L O N C . S . & C H O I J . ( 2 0 1 4 ) Evaluating the psychometric properties of the Jacelon Attributed Dignity Scale. Journal of Advanced Nursing 70(9), 2149–2161. doi: 10.1111/jan.12372
Abstract Aim. To develop and psychometrically test the Jacelon Attributed Dignity Scale (JADS). Background. The JADS was designed to measure self-perceived attributed dignity in community-dwelling older adults. Attributed dignity was conceived of as a state characteristic of the self. The JADS is a short, positively scored, normreferenced, evaluation index designed to measure self-perceived attributed dignity during the last week. Design. Instrument development and testing including psychometric properties, internal consistency, factor structure, temporal stability and construct validity. Method. Using a quota sample, 289 older adults (65–99 years old) were recruited from senior centres in western New England to complete the JADS, demographic information, the Self-Esteem Scale and the Social Desirability Scale during 2010– 2011. Descriptive statistics, exploratory factor analysis, construct validity and temporal stability were evaluated. Results. The resulting positively scored 18-item scale has four factors with high internal consistency for each factor and the entire scale. Construct validity was established by examining correlations with instruments that measured self-esteem and social desirability. Attributed dignity is a unique concept that is stable over time. Conclusion. The JADS is an 18-item Likert-scaled instrument designed to measure attributed dignity. Attributed dignity is a concept with four factors and is defined as a cognitive component of the selfconnoting self-value, perceived value from others, self in relation to others and behaving with respect. The importance of attributed dignity for older adults in relation to health, function, independence, quality of life and successful ageing can now be evaluated. Keywords: community dwelling, dignity, instrument development, nursing, older adults, theory
© 2014 John Wiley & Sons Ltd
2149
C.S. Jacelon and J. Choi
Why is this research or review needed? ● Dignity is important to the well-being of older adults. ● To date, there is no instrument to measure dignity in community-dwelling older adults.
What are the three key findings? ● The Jacelon Attributed Dignity Scale is a short, valid and reliable Likert-like scale to measure attributed dignity. ● Attributed dignity is a unique concept distinct from selfesteem and social desirability. ● Attributed dignity consists of four factors: the individual’s self-value,
their
perceived
value
from
others,
their
self-reflection on how they interact with others and their perception of the extent to which their behaviour demonstrates respect for others.
How should the findings be used to influence policy/ practice/research/education? ● The Jacelon Attributed Dignity Scale can be used to evaluate the efficacy of interventions designed to enhance dignity in older adults. ● The Jacelon Attributed Dignity Scale can be used to explore the importance of attributed dignity in the lives of older adults.
Introduction There is increasing interest in the concept of dignity for older adults in urgent care (Bridges & Nugus 2010), hospital care (Baillie 2009), long-term care (Franklin et al. 2006) and end-of-life care (Chochinov et al. 2006). There has been less exploration of dignity for community-dwelling older adults. Dignity, an intrinsic quality of being human that is manifested as an attributed, dynamic quality of the selfconnoting self-value and behaviour that demonstrates respect for self and others (Jacelon et al. 2004), varies among older adults and is sensitive to human interaction. Developing a way to measure attributed dignity in community-dwelling older adults will advance the understanding of this concept in the lives of these individuals. The purpose of the study was to test the psychometric properties, internal consistency, factor structure, temporal stability and construct validity of the Jacelon Attributed Dignity Scale (JADS).
Background Research on dignity and older adults has been conducted in Australia (Henderson et al. 2009), Canada (Jacobsen 2009, 2150
Chochinov et al. 2011, Montross et al. 2011), Germany (Pleschberger 2007), the UK (Arino-Biasco et al. 2005, Baillie & Gallagher 2011, Brown et al. 2011), Scandinavia (Jakobsen & Sorlie 2010), Taiwan (Lin et al. 2011) and the USA (Periyakoil et al. 2010). The research has two main foci: dignity in health care and dignity at the end of life. A third, lesser, focus of research is the dignity of communitydwelling older adults (Jacelon et al. 2004, Tadd 2004, Bayer et al. 2006, Calnan et al. 2006). Most research exploring dignity and older adults has been qualitative, using case method (Baillie 2009), ethnography (Tadd et al. 2011), phenomenology (Matiti & Trorey 2008) and qualitative description (Bridges & Nugus 2010). Other scientists have used surveys (Chochinov et al. 2006), randomized clinical trials (Chochinov et al. 2011) and instrument development (Chochinov et al. 2008, Jacelon et al. 2009, Periyakoil et al. 2010, Albers et al. 2011, Vlug et al. 2011). Dignity in health care Dignity is important for older adults who need health care (Fenton & Mitchell 2002, Jacelon 2003). Lack of dignity may lead to poorer health outcomes (Walsh & Kowanko 2002). Researchers have explored the concept of dignity in healthcare settings from the perspective of patients (Matiti & Trorey 2008, Henderson et al. 2009, Webster & Bryan 2009, Tadd et al. 2011), nurses and other care providers (Arino-Biasco et al. 2005, Baillie et al. 2009a,b, Baillie & Gallagher 2011, Lin & Tsai 2011), or both (Berg et al. 2006, Baillie 2008, 2009, Jacobsen 2009, Woolhead et al. 2009, Tadd et al. 2011). Much of this work has been conducted in the UK as the government has an initiative to maintain dignity in the care of older adults (Philip 2006). The most important concepts in preserving dignity in care settings were as follows: the patient’s sense of control and privacy, and appropriate communication between the care provider and the patient. Other findings included: the need for patients to be able to make choices about care, be comfortable and treated with respect by care providers and to have an advocate. Dignity at the end of life Many qualitative studies have been conducted (Franklin et al. 2006, Pleschberger 2007, Dwyer et al. 2009, Brown et al. 2011, Karlsson & Berggren 2011) and significant progress has been made in developing and testing a framework for dignity at the end of life (Chochinov et al. 2002, Hall et al. 2009). Efficacy of an intervention for promoting dignity at the end of life (Chochinov et al. 2005, McClement et al. 2007, Montross et al. 2011) has been supported in a © 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY
randomized clinical trial (Chochinov et al. 2011). These studies describe the patient’s need for dignity as they approach death. These needs include: the need for care for physical needs, frailty, dependence; the need for adequate communication; and the need to leave a legacy. Dignity and older adults Over time, many authors have emphasized the importance of dignity in old age (Maclaren 1977, Grau 1984, Erikson et al. 1986, Haug & Ory 1987, Volicer 1997, George 1998, Moody 1998, Lothian & Philp 2001, Bayer et al. 2006). A group of European scientists conducted a large focus group study across six European countries (Tadd 2005) to explore the importance of dignity for older adults. The 89 focus groups were comprised of older adults, family members, and health and social care professionals (Bayer et al. 2006, Calnan et al. 2006). Findings indicated that ‘Dignity was seen as a relevant and important concept which, if maintained, enhanced self-esteem, self-worth and well-being’ (Bayer et al. 2006, p. 24). Experiencing dignity involved having respect for oneself and being shown respect from others. Measuring dignity Three instruments to measure dignity at the end of life were identified (Chochinov et al. 2002, Chochinov et al. 2008, Periyakoil et al. 2010, Albers et al. 2011, Vlug et al. 2011). Chochinov et al. (2008) developed the Patient Dignity Inventory (PDI) to ‘measure sources of dignity distress’ for individuals nearing the end of life (p. 559). The five factors identified for the PDI are symptom distress, existential distress, dependency, peace of mind and social support (Chochinov et al. 2008, p. 564). Albers et al. (2011) tested the PDI on a large sample (N = 4941) in the Netherlands. Their findings supported the PDI as useful, but lacking content on aspects of both communication and care-related issues. Using a portion of the same sample (N = 2537) as Albers et al., a team of researchers led by Vlug used responses to open-ended questions to develop an instrument to identify factors that affect a patient’s self-perceived dignity with the goal of developing strategies for dignity-conserving care (Vlug et al. 2011). Four domains were identified including ‘evaluation of self in relation to others, functional status, mental state and care and situational aspects’ (p. 578). Although this scale measures similar concepts to the PDI, it provides extra value in two ways. First, the participants in this study were not at the end of life, but asked to reflect on desires as if they were at the end of life. Using a different population contributes to the generalizability of the findings. Second, the structure of the instrument indicates the presence of an issue, but also the importance. © 2014 John Wiley & Sons Ltd
Attributed Dignity Scale
In another effort to measure dignity at the end of life, Periyakoil et al. conducted an open-ended written survey with 100 care providers (Periyakoil et al. 2010) to ascertain key factors ‘thought to influence preservation and erosion of dignity’ (p. 496). They used the identified themes to create the preservation of dignity card sort tool (pDCT). The researchers defined dignity as having intrinsic factors and extrinsic factors. The intrinsic factors are inherent in all individuals; the extrinsic factors rest outside the individual and are affected by the behaviour of others (p. 496). The pDCT consists of 10 cards, each with an intrinsic or extrinsic factor from the developed model of dignity. The instruments described above are used to assess the interface between the terminally ill patient and the care provider. Themes across the instruments include: intrinsic factors such as existential distress, peace of mind or spirituality, and autonomy; extrinsic factors such as dignity-preserving situations and care, self in relation to others and respect; and corporeal factors such as functional status, mental status and symptom distress. These instruments are focused on end of life care for patients of any age and are not useful for independent older individuals without terminal illness. The JADS is designed to measure the attributed dignity of community-dwelling older adults. To evaluate the importance of dignity in the lives of older adults, it is essential to have a measure of dignity that has been developed for the general population.
Theoretical framework The first author developed the concept of attributed dignity. Her interest in dignity arose from a study of hospitalized older adults (Jacelon 2003, 2004). Following that study, the author found that the concept of dignity was poorly developed. This led to a hybrid concept analysis (Schwartz-Barcott & Kim 2001) employing a review of the literature and focus groups (Jacelon et al. 2004). Attributed dignity consisting of three factors, (SV) behaviour with respect to self (BRS) and behaviour with respect to others (BRO), was hypothesized to be a state characteristic, sensitive to human interaction and environmental stimuli (Miller & Keys 2001). Based on symbolic interaction (Blumer 1969), it was proposed that an individual’s attributed dignity would fluctuate in response to the meaning that the individual ascribed to a situation. This three-factor (SV, BRS, BRO) conceptualization of attributed dignity was the basis for development of the JADS. The JADS is proposed to measure attributed dignity based on the previous week of the individual’s life (Jacelon et al. 2009). 2151
C.S. Jacelon and J. Choi
Attributed dignity and related concepts Early in the development of the concept of attributed dignity, reviewers suggested that attributed dignity was a proxy for self-esteem or social desirability. Consequently, these concepts were selected for evaluating construct validity. Self-esteem is defined as ‘. . . a positive or negative attitude toward . . . the self’ (Rosenberg 1965, p. 30), the evaluative component of self-concept and a trait that is consistent over time (Blascovich & Tomaka 1991). Social desirability is defined as the extent to which an individual acts in ways to enhance positive responses of others (Crowne & Marlowe 1960).
The Study Aims To develop and test, psychometrically, the JADS.
Design Instrument development and testing to determine the psychometric properties, internal consistency, factor structure, temporal stability and construct validity of the JADS.
Methodology The JADS was conceived as a short, positively scored, norm-referenced, evaluation index (Mishel 1998) to measure an individual’s self-perceived attributed dignity using a five-point Likert scale response set. Exploratory factor analysis was used to determine the best number of factors for the JADS.
Sample/participants The participants were adults at least 65 years of age who could speak and understand English and were able to give informed consent. The middle and oldest old are thought to be most vulnerable and were deliberately over-sampled. The plan was to use quota sampling to create a sample that was one-third young old (65– 74 years old), one-third middle old (75–84 years old) and one-third oldest old (85 years old or older) participants. The goal was to recruit at least 230 participants, 10 subjects per item, based on the most commonly agreed on minimum sample size for factor analysis with unknown variability (Pedhazur & Schmelkin, 1991). To test temporal stability, 10% of the sample participants, plus 5 to compensate for possible missing data (n = 28), were tar2152
geted to repeat completion of the JADS 2 to 3 days after the initial data completion. An additional 10%, plus five (n = 28) participants, were to complete the JADS 30 days later.
Instruments The JADS was developed using a two-stage process outlined by Lynn (1986). The stages were item development and item testing. Original items were developed from statements made by community-dwelling older adults during focus groups and informed by literature review (Jacelon et al. 2004). Two review panels of experts, a review panel of older adults, cognitive appraisal interviews (Knafl et al. 2007) and a pilot test of the instrument (Jacelon et al. 2009) were conducted. The items are written in a 6th grade English language reading level (DeVellis 2003) and printed in large print (Burnside et al. 1998). At the conclusion of the pilot testing, the JADS was a 23-item scale (SV 8 items, BRS 5 items, BRO 10 items) with a five-point Likert response set ranging from 1 ‘totally false’–5 ‘totally true’. Attributed dignity has been operationalized as the total score, or scores achieved on the JADS. Higher scores indicated greater attributed dignity. In addition to the JADS, participants completed a demographic data sheet (Table 2), the Rosenberg SelfEsteem Scale (SES) (1965) and the Crowne and Marlow Social Desirability Scale (SDS) (1960). The SES is a 10item positively scored scale that has been used to explore the concept of self-esteem in widely varying populations including older adults. Cronbach’s a (Cronbach 1951) of the SES ranges from 077–083 (Dobson et al. 1979, Murrell et al. 2003, Skultety & Whitbourne 2004). Self-esteem as scored on the SES is a continuous variable. The SES is rated as a summed Likert scale with a maximum score of 40 and a minimum score of 10 (Morris Rosenberg Foundation 2004). Participants also completed the SDS (Crowne & Marlowe 1960). The SDS measures the likelihood of response distortion on self-report measures. For this study, a shortened version, the SDS 2 (10) was used (Strahan & Gerbashi 1972). The 10-item scale has five items keyed true and five keyed false. Scores can range from 0–10 with higher scores indicating less likelihood of response distortion on selfreport measures. The SDS 2 (10) has been used with samples of older adults (Taren et al. 1999, Horner et al. 2002, Consedine et al. 2004) with reported internal consistency of 088 with test–retest reliability also of 088 (Taren et al. 1999). © 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY
Ethical considerations The Institutional Review Board of the authors’ university approved the study. Permission to recruit participants at senior centres was obtained from the State Council on Aging. There were no unexpected events during data collection.
Data collection Older adults were recruited from urban and rural senior centres in western New England, USA, during 2010–2011. Flyers were posted at senior centres explaining the study and posting the dates a researcher would be at the site. Volunteers came to the centre and read the consent form as the researcher read it aloud. After the researcher answered any questions, the participants signed the consent document and completed the questionnaires. The researcher read the instruments verbatim to participants who had vision problems or other health problems that required assistance. Participants were given a $1000 thank you gift card to a local grocery store.
Data analysis All data were entered into an Excel 2004 for Mac (Version 11.6.4) spreadsheet. Data analysis proceeded in several steps using Excel and SPSS Version 19.0 software (Armonk, NY, USA). The sample was described by calculating frequency distributions, means and SD as appropriate for all demographic information. Descriptive statistics were calculated for each item on every scale. These statistics included frequency distributions, means, SD and the extent of missing data. Missing data did not exceed 4% on any scale item and no patterns were discerned; therefore, we did not impute missing data. Scale scores and Cronbach’s a were calculated for the JADS, SES and SDS. Psychometric analysis of the JADS followed the descriptive analysis. The functioning of individual items was explored for variation in responses, skewness and correlations with other items and with the total scale. This process informed the selection and retention of items in the creation of the final form of the JADS. Next, an inter-item correlation matrix was constructed to explore patterns of relationships between items. Items that performed poorly loading on more than one factor or weakly loading on any factor were deleted. Analysis proceeded using exploratory factor analysis (EFA) to identify the basic numerical structure of the data and to explore the most logical number of factors for the JADS. EFA is used when the number of factors in an instrument has not © 2014 John Wiley & Sons Ltd
Attributed Dignity Scale
been predetermined. In the case of the JADS, the items were generated based on the dimensions identified during concept analysis using qualitative data. These dimensions had not yet been tested empirically; therefore, an exploratory strategy was most appropriate. Based on the proposed relationships among factors, principal components analysis with oblique rotation was selected for use (Tabachnick & Fidell 2008). We chose oblique rotation because the factors are theoretically connected (Figure 1) and the subscale correlation matrix indicated moderate correlations among factors. The number of factors was decided using several strategies. Factors whose eigenvalues were close to or greater than 10 were retained in the analysis. A Scree Test was plotted using the eigenvalues of the factors (Nunnally & Bernstein 1994) to determine the best factor solution. The best explanation of the factors was retained. Each factor of this ‘best’ solution was inspected. This inspection included evaluating internal consistency reliability (Cronbach’s a coefficient), the contribution of specific items as indicated by factor loading (>04) (Dixon 2005) and change in alpha if that item was deleted. The factors identified by EFA were then compared with the proposed dimensions of the JADS and named based on the theme of the items in each factor. Temporal stability was assessed at two points to determine if attributed dignity was a state or trait characteristic. The first five participants at each data collection session who agreed to be retested were alternately assigned to either the 2–3-day or the 30-day retest group until each group included 28 subjects. To test the assumption that attributed dignity is a state characteristic varying over time and as an additional measure of construct validity and a preliminary exploration of the variability in attributed dignity over time, an additional retest sample was recruited. In general, a period of 2–14 days is suggested for testing the stability of state concepts (Streiner Intrinsic dignity
Attributed dignity factors of the ADS Perceived value from other (PVO)
Human Manifested dignity by
Self value (SV)
Self in relation to others (SRO)
Behavioral respect others (BRO)
Figure 1 Attributed dignity with four-factor solution. 2153
C.S. Jacelon and J. Choi
& Norman 2003). The first test–retest was to test the stability with respect to the same reference period; the second retest was to test the stability with respect to a difference reference period. If attributed dignity were a state characteristic, it would be stable during the first retest period and vary during the second, whereas if it is a trait characteristic, it was anticipated that the scores on the JADS would be stable across both reference periods. In both cases, a reliability coefficient of ≥070 was accepted as evidence of test–retest reliability (Aday 1996). Test–retest was analysed by computing intraclass correlation coefficients (ICC) using a two-way random-effect model and type consistency (Shrout & Fleiss 1979, McGraw & Wong 1996).
Results Sample The sample consisted of two hundred and eighty-nine (N = 289) older adults (Table 1). The sample was larger than planned (N = 230) to increase the number of participants in the oldest age groups. As planned, we over-sampled the older age groups, with an average age of 774 years (range 65–99) and, as expected, the sample was 74% (n = 210) women. The sample was less diverse than the US population, with only 9% (n = 25) of respondents self-identifying as other than white, while nationally 16% of the older population self-identify as diverse (US Bureau of Census 2000). Other sample characteristics can be found in Table 2.
Item response For all items, no response was endorsed more than 80% of the time. The response choices, ‘totally false’ and the ‘mostly false’ were rarely endorsed. Across the 18 items in the final solution, ‘totally false’ was chosen as a response only three times. ‘Mostly false’ was endorsed 11 times.
Table 2 Participant characteristics. Education
Female (%)
Male (%)
Total (%)
Evaluating the psychometric properties of the Jacelon Attributed Dignity Scale Cynthia S. Jacelon & Jeungok Choi Accepted for publication 25 January 2014
Correspondence to C.S. Jacelon: e-mail: [email protected] Cynthia S. Jacelon PhD RN FAAN Associate Professor College of Nursing, University of Massachusetts Amherst, Amherst, MA, USA Jeungok Choi PhD MPH RN Associate Professor College of Nursing, University of Massachusetts Amherst, Amherst, MA, USA
J A C E L O N C . S . & C H O I J . ( 2 0 1 4 ) Evaluating the psychometric properties of the Jacelon Attributed Dignity Scale. Journal of Advanced Nursing 70(9), 2149–2161. doi: 10.1111/jan.12372
Abstract Aim. To develop and psychometrically test the Jacelon Attributed Dignity Scale (JADS). Background. The JADS was designed to measure self-perceived attributed dignity in community-dwelling older adults. Attributed dignity was conceived of as a state characteristic of the self. The JADS is a short, positively scored, normreferenced, evaluation index designed to measure self-perceived attributed dignity during the last week. Design. Instrument development and testing including psychometric properties, internal consistency, factor structure, temporal stability and construct validity. Method. Using a quota sample, 289 older adults (65–99 years old) were recruited from senior centres in western New England to complete the JADS, demographic information, the Self-Esteem Scale and the Social Desirability Scale during 2010– 2011. Descriptive statistics, exploratory factor analysis, construct validity and temporal stability were evaluated. Results. The resulting positively scored 18-item scale has four factors with high internal consistency for each factor and the entire scale. Construct validity was established by examining correlations with instruments that measured self-esteem and social desirability. Attributed dignity is a unique concept that is stable over time. Conclusion. The JADS is an 18-item Likert-scaled instrument designed to measure attributed dignity. Attributed dignity is a concept with four factors and is defined as a cognitive component of the selfconnoting self-value, perceived value from others, self in relation to others and behaving with respect. The importance of attributed dignity for older adults in relation to health, function, independence, quality of life and successful ageing can now be evaluated. Keywords: community dwelling, dignity, instrument development, nursing, older adults, theory
© 2014 John Wiley & Sons Ltd
2149
C.S. Jacelon and J. Choi
Why is this research or review needed? ● Dignity is important to the well-being of older adults. ● To date, there is no instrument to measure dignity in community-dwelling older adults.
What are the three key findings? ● The Jacelon Attributed Dignity Scale is a short, valid and reliable Likert-like scale to measure attributed dignity. ● Attributed dignity is a unique concept distinct from selfesteem and social desirability. ● Attributed dignity consists of four factors: the individual’s self-value,
their
perceived
value
from
others,
their
self-reflection on how they interact with others and their perception of the extent to which their behaviour demonstrates respect for others.
How should the findings be used to influence policy/ practice/research/education? ● The Jacelon Attributed Dignity Scale can be used to evaluate the efficacy of interventions designed to enhance dignity in older adults. ● The Jacelon Attributed Dignity Scale can be used to explore the importance of attributed dignity in the lives of older adults.
Introduction There is increasing interest in the concept of dignity for older adults in urgent care (Bridges & Nugus 2010), hospital care (Baillie 2009), long-term care (Franklin et al. 2006) and end-of-life care (Chochinov et al. 2006). There has been less exploration of dignity for community-dwelling older adults. Dignity, an intrinsic quality of being human that is manifested as an attributed, dynamic quality of the selfconnoting self-value and behaviour that demonstrates respect for self and others (Jacelon et al. 2004), varies among older adults and is sensitive to human interaction. Developing a way to measure attributed dignity in community-dwelling older adults will advance the understanding of this concept in the lives of these individuals. The purpose of the study was to test the psychometric properties, internal consistency, factor structure, temporal stability and construct validity of the Jacelon Attributed Dignity Scale (JADS).
Background Research on dignity and older adults has been conducted in Australia (Henderson et al. 2009), Canada (Jacobsen 2009, 2150
Chochinov et al. 2011, Montross et al. 2011), Germany (Pleschberger 2007), the UK (Arino-Biasco et al. 2005, Baillie & Gallagher 2011, Brown et al. 2011), Scandinavia (Jakobsen & Sorlie 2010), Taiwan (Lin et al. 2011) and the USA (Periyakoil et al. 2010). The research has two main foci: dignity in health care and dignity at the end of life. A third, lesser, focus of research is the dignity of communitydwelling older adults (Jacelon et al. 2004, Tadd 2004, Bayer et al. 2006, Calnan et al. 2006). Most research exploring dignity and older adults has been qualitative, using case method (Baillie 2009), ethnography (Tadd et al. 2011), phenomenology (Matiti & Trorey 2008) and qualitative description (Bridges & Nugus 2010). Other scientists have used surveys (Chochinov et al. 2006), randomized clinical trials (Chochinov et al. 2011) and instrument development (Chochinov et al. 2008, Jacelon et al. 2009, Periyakoil et al. 2010, Albers et al. 2011, Vlug et al. 2011). Dignity in health care Dignity is important for older adults who need health care (Fenton & Mitchell 2002, Jacelon 2003). Lack of dignity may lead to poorer health outcomes (Walsh & Kowanko 2002). Researchers have explored the concept of dignity in healthcare settings from the perspective of patients (Matiti & Trorey 2008, Henderson et al. 2009, Webster & Bryan 2009, Tadd et al. 2011), nurses and other care providers (Arino-Biasco et al. 2005, Baillie et al. 2009a,b, Baillie & Gallagher 2011, Lin & Tsai 2011), or both (Berg et al. 2006, Baillie 2008, 2009, Jacobsen 2009, Woolhead et al. 2009, Tadd et al. 2011). Much of this work has been conducted in the UK as the government has an initiative to maintain dignity in the care of older adults (Philip 2006). The most important concepts in preserving dignity in care settings were as follows: the patient’s sense of control and privacy, and appropriate communication between the care provider and the patient. Other findings included: the need for patients to be able to make choices about care, be comfortable and treated with respect by care providers and to have an advocate. Dignity at the end of life Many qualitative studies have been conducted (Franklin et al. 2006, Pleschberger 2007, Dwyer et al. 2009, Brown et al. 2011, Karlsson & Berggren 2011) and significant progress has been made in developing and testing a framework for dignity at the end of life (Chochinov et al. 2002, Hall et al. 2009). Efficacy of an intervention for promoting dignity at the end of life (Chochinov et al. 2005, McClement et al. 2007, Montross et al. 2011) has been supported in a © 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY
randomized clinical trial (Chochinov et al. 2011). These studies describe the patient’s need for dignity as they approach death. These needs include: the need for care for physical needs, frailty, dependence; the need for adequate communication; and the need to leave a legacy. Dignity and older adults Over time, many authors have emphasized the importance of dignity in old age (Maclaren 1977, Grau 1984, Erikson et al. 1986, Haug & Ory 1987, Volicer 1997, George 1998, Moody 1998, Lothian & Philp 2001, Bayer et al. 2006). A group of European scientists conducted a large focus group study across six European countries (Tadd 2005) to explore the importance of dignity for older adults. The 89 focus groups were comprised of older adults, family members, and health and social care professionals (Bayer et al. 2006, Calnan et al. 2006). Findings indicated that ‘Dignity was seen as a relevant and important concept which, if maintained, enhanced self-esteem, self-worth and well-being’ (Bayer et al. 2006, p. 24). Experiencing dignity involved having respect for oneself and being shown respect from others. Measuring dignity Three instruments to measure dignity at the end of life were identified (Chochinov et al. 2002, Chochinov et al. 2008, Periyakoil et al. 2010, Albers et al. 2011, Vlug et al. 2011). Chochinov et al. (2008) developed the Patient Dignity Inventory (PDI) to ‘measure sources of dignity distress’ for individuals nearing the end of life (p. 559). The five factors identified for the PDI are symptom distress, existential distress, dependency, peace of mind and social support (Chochinov et al. 2008, p. 564). Albers et al. (2011) tested the PDI on a large sample (N = 4941) in the Netherlands. Their findings supported the PDI as useful, but lacking content on aspects of both communication and care-related issues. Using a portion of the same sample (N = 2537) as Albers et al., a team of researchers led by Vlug used responses to open-ended questions to develop an instrument to identify factors that affect a patient’s self-perceived dignity with the goal of developing strategies for dignity-conserving care (Vlug et al. 2011). Four domains were identified including ‘evaluation of self in relation to others, functional status, mental state and care and situational aspects’ (p. 578). Although this scale measures similar concepts to the PDI, it provides extra value in two ways. First, the participants in this study were not at the end of life, but asked to reflect on desires as if they were at the end of life. Using a different population contributes to the generalizability of the findings. Second, the structure of the instrument indicates the presence of an issue, but also the importance. © 2014 John Wiley & Sons Ltd
Attributed Dignity Scale
In another effort to measure dignity at the end of life, Periyakoil et al. conducted an open-ended written survey with 100 care providers (Periyakoil et al. 2010) to ascertain key factors ‘thought to influence preservation and erosion of dignity’ (p. 496). They used the identified themes to create the preservation of dignity card sort tool (pDCT). The researchers defined dignity as having intrinsic factors and extrinsic factors. The intrinsic factors are inherent in all individuals; the extrinsic factors rest outside the individual and are affected by the behaviour of others (p. 496). The pDCT consists of 10 cards, each with an intrinsic or extrinsic factor from the developed model of dignity. The instruments described above are used to assess the interface between the terminally ill patient and the care provider. Themes across the instruments include: intrinsic factors such as existential distress, peace of mind or spirituality, and autonomy; extrinsic factors such as dignity-preserving situations and care, self in relation to others and respect; and corporeal factors such as functional status, mental status and symptom distress. These instruments are focused on end of life care for patients of any age and are not useful for independent older individuals without terminal illness. The JADS is designed to measure the attributed dignity of community-dwelling older adults. To evaluate the importance of dignity in the lives of older adults, it is essential to have a measure of dignity that has been developed for the general population.
Theoretical framework The first author developed the concept of attributed dignity. Her interest in dignity arose from a study of hospitalized older adults (Jacelon 2003, 2004). Following that study, the author found that the concept of dignity was poorly developed. This led to a hybrid concept analysis (Schwartz-Barcott & Kim 2001) employing a review of the literature and focus groups (Jacelon et al. 2004). Attributed dignity consisting of three factors, (SV) behaviour with respect to self (BRS) and behaviour with respect to others (BRO), was hypothesized to be a state characteristic, sensitive to human interaction and environmental stimuli (Miller & Keys 2001). Based on symbolic interaction (Blumer 1969), it was proposed that an individual’s attributed dignity would fluctuate in response to the meaning that the individual ascribed to a situation. This three-factor (SV, BRS, BRO) conceptualization of attributed dignity was the basis for development of the JADS. The JADS is proposed to measure attributed dignity based on the previous week of the individual’s life (Jacelon et al. 2009). 2151
C.S. Jacelon and J. Choi
Attributed dignity and related concepts Early in the development of the concept of attributed dignity, reviewers suggested that attributed dignity was a proxy for self-esteem or social desirability. Consequently, these concepts were selected for evaluating construct validity. Self-esteem is defined as ‘. . . a positive or negative attitude toward . . . the self’ (Rosenberg 1965, p. 30), the evaluative component of self-concept and a trait that is consistent over time (Blascovich & Tomaka 1991). Social desirability is defined as the extent to which an individual acts in ways to enhance positive responses of others (Crowne & Marlowe 1960).
The Study Aims To develop and test, psychometrically, the JADS.
Design Instrument development and testing to determine the psychometric properties, internal consistency, factor structure, temporal stability and construct validity of the JADS.
Methodology The JADS was conceived as a short, positively scored, norm-referenced, evaluation index (Mishel 1998) to measure an individual’s self-perceived attributed dignity using a five-point Likert scale response set. Exploratory factor analysis was used to determine the best number of factors for the JADS.
Sample/participants The participants were adults at least 65 years of age who could speak and understand English and were able to give informed consent. The middle and oldest old are thought to be most vulnerable and were deliberately over-sampled. The plan was to use quota sampling to create a sample that was one-third young old (65– 74 years old), one-third middle old (75–84 years old) and one-third oldest old (85 years old or older) participants. The goal was to recruit at least 230 participants, 10 subjects per item, based on the most commonly agreed on minimum sample size for factor analysis with unknown variability (Pedhazur & Schmelkin, 1991). To test temporal stability, 10% of the sample participants, plus 5 to compensate for possible missing data (n = 28), were tar2152
geted to repeat completion of the JADS 2 to 3 days after the initial data completion. An additional 10%, plus five (n = 28) participants, were to complete the JADS 30 days later.
Instruments The JADS was developed using a two-stage process outlined by Lynn (1986). The stages were item development and item testing. Original items were developed from statements made by community-dwelling older adults during focus groups and informed by literature review (Jacelon et al. 2004). Two review panels of experts, a review panel of older adults, cognitive appraisal interviews (Knafl et al. 2007) and a pilot test of the instrument (Jacelon et al. 2009) were conducted. The items are written in a 6th grade English language reading level (DeVellis 2003) and printed in large print (Burnside et al. 1998). At the conclusion of the pilot testing, the JADS was a 23-item scale (SV 8 items, BRS 5 items, BRO 10 items) with a five-point Likert response set ranging from 1 ‘totally false’–5 ‘totally true’. Attributed dignity has been operationalized as the total score, or scores achieved on the JADS. Higher scores indicated greater attributed dignity. In addition to the JADS, participants completed a demographic data sheet (Table 2), the Rosenberg SelfEsteem Scale (SES) (1965) and the Crowne and Marlow Social Desirability Scale (SDS) (1960). The SES is a 10item positively scored scale that has been used to explore the concept of self-esteem in widely varying populations including older adults. Cronbach’s a (Cronbach 1951) of the SES ranges from 077–083 (Dobson et al. 1979, Murrell et al. 2003, Skultety & Whitbourne 2004). Self-esteem as scored on the SES is a continuous variable. The SES is rated as a summed Likert scale with a maximum score of 40 and a minimum score of 10 (Morris Rosenberg Foundation 2004). Participants also completed the SDS (Crowne & Marlowe 1960). The SDS measures the likelihood of response distortion on self-report measures. For this study, a shortened version, the SDS 2 (10) was used (Strahan & Gerbashi 1972). The 10-item scale has five items keyed true and five keyed false. Scores can range from 0–10 with higher scores indicating less likelihood of response distortion on selfreport measures. The SDS 2 (10) has been used with samples of older adults (Taren et al. 1999, Horner et al. 2002, Consedine et al. 2004) with reported internal consistency of 088 with test–retest reliability also of 088 (Taren et al. 1999). © 2014 John Wiley & Sons Ltd
JAN: RESEARCH METHODOLOGY
Ethical considerations The Institutional Review Board of the authors’ university approved the study. Permission to recruit participants at senior centres was obtained from the State Council on Aging. There were no unexpected events during data collection.
Data collection Older adults were recruited from urban and rural senior centres in western New England, USA, during 2010–2011. Flyers were posted at senior centres explaining the study and posting the dates a researcher would be at the site. Volunteers came to the centre and read the consent form as the researcher read it aloud. After the researcher answered any questions, the participants signed the consent document and completed the questionnaires. The researcher read the instruments verbatim to participants who had vision problems or other health problems that required assistance. Participants were given a $1000 thank you gift card to a local grocery store.
Data analysis All data were entered into an Excel 2004 for Mac (Version 11.6.4) spreadsheet. Data analysis proceeded in several steps using Excel and SPSS Version 19.0 software (Armonk, NY, USA). The sample was described by calculating frequency distributions, means and SD as appropriate for all demographic information. Descriptive statistics were calculated for each item on every scale. These statistics included frequency distributions, means, SD and the extent of missing data. Missing data did not exceed 4% on any scale item and no patterns were discerned; therefore, we did not impute missing data. Scale scores and Cronbach’s a were calculated for the JADS, SES and SDS. Psychometric analysis of the JADS followed the descriptive analysis. The functioning of individual items was explored for variation in responses, skewness and correlations with other items and with the total scale. This process informed the selection and retention of items in the creation of the final form of the JADS. Next, an inter-item correlation matrix was constructed to explore patterns of relationships between items. Items that performed poorly loading on more than one factor or weakly loading on any factor were deleted. Analysis proceeded using exploratory factor analysis (EFA) to identify the basic numerical structure of the data and to explore the most logical number of factors for the JADS. EFA is used when the number of factors in an instrument has not © 2014 John Wiley & Sons Ltd
Attributed Dignity Scale
been predetermined. In the case of the JADS, the items were generated based on the dimensions identified during concept analysis using qualitative data. These dimensions had not yet been tested empirically; therefore, an exploratory strategy was most appropriate. Based on the proposed relationships among factors, principal components analysis with oblique rotation was selected for use (Tabachnick & Fidell 2008). We chose oblique rotation because the factors are theoretically connected (Figure 1) and the subscale correlation matrix indicated moderate correlations among factors. The number of factors was decided using several strategies. Factors whose eigenvalues were close to or greater than 10 were retained in the analysis. A Scree Test was plotted using the eigenvalues of the factors (Nunnally & Bernstein 1994) to determine the best factor solution. The best explanation of the factors was retained. Each factor of this ‘best’ solution was inspected. This inspection included evaluating internal consistency reliability (Cronbach’s a coefficient), the contribution of specific items as indicated by factor loading (>04) (Dixon 2005) and change in alpha if that item was deleted. The factors identified by EFA were then compared with the proposed dimensions of the JADS and named based on the theme of the items in each factor. Temporal stability was assessed at two points to determine if attributed dignity was a state or trait characteristic. The first five participants at each data collection session who agreed to be retested were alternately assigned to either the 2–3-day or the 30-day retest group until each group included 28 subjects. To test the assumption that attributed dignity is a state characteristic varying over time and as an additional measure of construct validity and a preliminary exploration of the variability in attributed dignity over time, an additional retest sample was recruited. In general, a period of 2–14 days is suggested for testing the stability of state concepts (Streiner Intrinsic dignity
Attributed dignity factors of the ADS Perceived value from other (PVO)
Human Manifested dignity by
Self value (SV)
Self in relation to others (SRO)
Behavioral respect others (BRO)
Figure 1 Attributed dignity with four-factor solution. 2153
C.S. Jacelon and J. Choi
& Norman 2003). The first test–retest was to test the stability with respect to the same reference period; the second retest was to test the stability with respect to a difference reference period. If attributed dignity were a state characteristic, it would be stable during the first retest period and vary during the second, whereas if it is a trait characteristic, it was anticipated that the scores on the JADS would be stable across both reference periods. In both cases, a reliability coefficient of ≥070 was accepted as evidence of test–retest reliability (Aday 1996). Test–retest was analysed by computing intraclass correlation coefficients (ICC) using a two-way random-effect model and type consistency (Shrout & Fleiss 1979, McGraw & Wong 1996).
Results Sample The sample consisted of two hundred and eighty-nine (N = 289) older adults (Table 1). The sample was larger than planned (N = 230) to increase the number of participants in the oldest age groups. As planned, we over-sampled the older age groups, with an average age of 774 years (range 65–99) and, as expected, the sample was 74% (n = 210) women. The sample was less diverse than the US population, with only 9% (n = 25) of respondents self-identifying as other than white, while nationally 16% of the older population self-identify as diverse (US Bureau of Census 2000). Other sample characteristics can be found in Table 2.
Item response For all items, no response was endorsed more than 80% of the time. The response choices, ‘totally false’ and the ‘mostly false’ were rarely endorsed. Across the 18 items in the final solution, ‘totally false’ was chosen as a response only three times. ‘Mostly false’ was endorsed 11 times.
Table 2 Participant characteristics. Education
Female (%)
Male (%)
Total (%)
