The psychometric properties of pen-and-pencil and internet versions of the midlife women's symptom index (MSI)
Introduction
Internet research allows people in different geographical areas across the US to communicate and exchange their experiences and opinions at their convenience without long-distance travel (Fawcett and Buhle, 1995). Studies have supported the feasibility of Internet surveys, richness of the electronically collected data, informative and prompt responses of research participants, decreased human errors, and nearly identical results with mail and telephone surveys (Fawcett and Buhle, 1995; Curl and Robinson, 1997; Lakeman, 1997; Pew Research Center for the People and the Press, 1999; Stanton, 1998).
Truell et al. (2002) even reported that Internet surveys were more effective than mail surveys in settings when the target populations had both e-mail and Internet access, although Internet and mail surveys achieved a similar response rate. McCabe et al. (2002) reported that the Internet survey resulted in a final sample that more closely matched the target sample in gender mix than did the mail survey, and the response rate for the Internet survey was significantly higher than for the mail survey. Andersson et al. (2002) recently reported that the data collection format did not result in any substantial difference between the Internet survey and the mail survey. Despite the recent positive findings on the Internet research compared with the traditional research methods including mail surveys and telephone surveys, very little is still known about differences in psychometric properties of the Internet survey format and the traditional survey format of an identical instrument (Riva et al., 2003).
Recently, a measurement scale for symptoms experienced during the menopausal transition, the Midlife Women's Symptom Index (MSI), was developed with an overall goal of development of a general measurement scale to measure symptoms experienced by diverse ethnic groups of women during the menopausal transition. The purposes of the study presented in this paper were to determine comparability of an Internet format to a PP format of the MSI. Specific aims of this study were to:
Aim #1. Determine reliability (internal consistency, inter-item correlations, and item-total correlations) of the PP format and the Internet format of the MSI.
Aim #2. Determine convergent, construct, and face validity of the PP format and the Internet format of the MSI.
Aim #3. Compare item omission, response rate, and completeness of the survey between the PP format and the Internet format.
Section snippets
Methods
This study utilized a cross-sectional comparative design with three sections: (a) a PP survey (Group 1); (b) an Internet survey (Group 2); and (c) an expert review (Group 3).
Findings
The findings supported the internal consistency reliability of the PP format and the Internet format of the MSI (see Table 3). K-R20 for the PP format was 0.94 among all participants, 0.94 among Whites, 0.88 among Asians, and 0.97 among Hispanics. Yet, K-R20 for the PP format among African Americans could not be calculated because of the small sample size of this subgroup (n=4). K-R20 for the Internet format was 0.86 among all the participants, 0.89 among Whites, 0.78 among Hispanics, and 0.86
Discussions
Studies have reported that Internet surveys generate comparable data to PP surveys. Pettit (2002) reported that there was no statistically significant difference between the PP survey data and the Internet survey data. Davis (1999) asserted that findings from Internet-based questionnaire research are comparable with results obtained using standard procedures. Fouladi et al. (2002) indicated that the magnitude of mode of administration effects was in general very small. The findings of the study
Implications and conclusions
In the study presented in this paper, psychometric properties of an Internet format and a PP format of the MSI were determined. The findings supported the reliability and validity of the Internet and PP format of the MSI, and suggested the benefits of the Internet format of the MSI in terms of item omissions and completeness. Based on the discussions on the findings, we want to conclude this paper with the following implications.
First, because the Internet format and the PP format of the MSI
Acknowledgements
This study was funded by the Expedited Proposal-Enhancement Grant (EP-EG) program, the Center for Health Promotion Research: a NIH-funded Research Center, NINR (P30 NR05051).
References (33)
- et al.
Is there a menopausal syndrome? Menopausal status and symptoms across racial/ethnic groups
Social Science and Medicine
(2001) - et al.
The Cornell Medical Indexhealth questionnaire in the identification of neurotic patients in general practice
Journal of Psychosomatic Research
(1962) - et al.
Cross-cultural validityethnocentrism in health studies with special reference to the Vietnamese
Social Science and Medicine
(1984) - et al.
The menopausal syndromea factor analytic replication
Maturitas
(1991) - et al.
Researching the symptoms of menopausean exercise in methodology
Maturitas
(1988) - et al.
Symptom experiences of midlife womenobservations from the seattle midlife women's health study
Maturitas
(1996) - et al.
Hypersensitivity to sound (hyperacusis)a prevalence study via the internet and post
International Journal of Audiology
(2002) Preface
- et al.
Vitamin E and the climacteric syndrome
Archives of Internal Medicine
(1953) - et al.
Manual of Cornell Medical Indexhealth questionnaire. New York hospital and the department of medicine and psychiatry
(1956)
The reliability and validity of the Chinese version of the caregiver burden inventory
Nursing Research
Handheld computers in clinical audita comparison with established pen and paper method
International Journal of Health Care Quality Assurance
Web-based administration of a personality questionnairecomparison with traditional methods
Behav Res Methods Instrum Comput
A comparison of cigarette smokers recruited through the internet or by mail
International Journal of Epidemiology
Using the internet for data collection. An innovative electronic strategy
Computers in Nursing
Cited by (38)
The Psychometric Properties of the Midlife Women's Symptom Index
2015, JOGNN - Journal of Obstetric, Gynecologic, and Neonatal NursingCitation Excerpt :The MSI was developed from a feminist perspective to measure symptoms of menopause relevant to women’s experiences, with consideration given to the participants’ race/ethnicity and cultural background (Im, 2009). The first version of the MSI consisted of 88 items (Im, 2006, 2009; Im et al., 2005), but 15 items were eliminated based on findings of item-total correlations of greater than 0.7 or less than 0.3, which yielded the most updated version of the instrument (Im et al., 2010). The current version of the MSI contains 73 items and three subscales, including physical (51 items), psychological (18 items), and psychosomatic (four items) symptoms of menopause, and has two sections: prevalence and severity (Im et al., 2010; Lee et al., 2010).
Effects of the Body Mass Index on Menopausal Symptoms Among Asian American Midlife Women Using Two Different Classification Systems
2014, JOGNN - Journal of Obstetric, Gynecologic, and Neonatal NursingReliability and validity of self-reported burnout in college students: A cross randomized comparison of paper-and-pencil vs. online administration
2011, Computers in Human BehaviorCitation Excerpt :In an earlier study on the equivalence of computerized vs. traditional research methods, Whitener and Klein (1995) identified several factors that may explain the lack of factor equivalence and threaten validity. These include socio-demographic differences between samples as has been acknowledged in other investigations (Bowling, 2005; Herrero & Meneses, 2006; Hewson & Charlton, 2005; Im et al., 2005; Klovning, Sandvik, & Hunskaar, 2009); single vs. group effects; mode of administration and instruments presentation and social desirability (Bowling, 2005; Fan & Yan, 2010; Hewson & Charlton, 2005; Thorndike et al., 2009; Whitener & Klein, 1995). To minimize these effects, namely single vs. group effects and mode of administration, we adopted a cross-over design with all the items being presented simultaneously, in the same order, and with the possibility to review and change answers if desired for both paper-and-pencil and internet formats.
Methodological and ethical considerations in designing an Internet study of quality of life: A discussion paper
2009, International Journal of Nursing StudiesWomen's health: The view as we near 2010
2008, International Journal of Nursing StudiesFactors Associated With Perceived Health Status of Multiracial/Ethnic Midlife Women in the United States
2016, JOGNN - Journal of Obstetric, Gynecologic, and Neonatal NursingCitation Excerpt :Menopausal symptoms were evaluated according to the total number of symptoms using the Midlife Women's Symptom Index (Im, 2006). In this study, the reliability of the Midlife Women's Symptom Index (Cronbach's alpha = .95) was good, as shown previously (Im et al., 2005). The behavioral factors assessed were smoking and perceived level of physical activity.