Open Collections

UBC Community and Partners Publications

Can we compare mean differences across diverse ethnocultural groups? 2009

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


SabrinaWong2008SpringInstituteApr24.pdf [ 206.81kB ]
JSON: 1.0103700.json
JSON-LD: 1.0103700+ld.json
RDF/XML (Pretty): 1.0103700.xml
RDF/JSON: 1.0103700+rdf.json
Turtle: 1.0103700+rdf-turtle.txt
N-Triples: 1.0103700+rdf-ntriples.txt

Full Text

UBC School of Nursing Can we compare mean differences across diverse ethnocultural groups? Sabrina T. Wong, RN PhD Nexus, April 2008 UBC School of Nursing Acknowledgements • Dave Nordstokke-PhD candidate in educational psychology • Steve Gregorich-biostatistician at UCSF • Women who participated in this study • AHRQ: 5P01 HS10856 and NIH RCMAR P30-AG15272 UBC School of Nursing Questions for you…. • Two groups: students and not-students • Do you want to travel more? • Do you want a nicer car? • Do you want to earn more $$? UBC School of Nursing Objectives • Describe the use of quantitative methods to examine if valid group comparisons can be made • Provide an example of an instrument that can be used for meaningful comparisons across groups UBC School of Nursing Social Support Defined • Processes through which social relationships might promote health and well-being • Multi-dimensional: tangible, emotional/companionship, informational UBC School of Nursing Background: Social Support and Aging • Social support keeps people healthy and health them recover from illness faster • Emotional support is specifically associated with: – More positive patterns of cognitive aging – Increased social engagement – Decreased levels of depression UBC School of Nursing Social Support and Ethnocultural Groups • Aging population is becoming increasingly diverse – In US, African Americans, Asians, and Latinos will make-up close to 30% of those aged 65 and older • Can mean social support scores be compared across ethnocultural groups? UBC School of Nursing Purpose • Can we make meaningful comparisons of social support across women from four diverse ethnic groups (African American, Latino, Chinese, and Caucasian) UBC School of Nursing Methods • Telephone recruitment with face-to-face survey administration • Data collected in English, Spanish, and Chinese (Mandarin and Cantonese) • Multidimensional social support instrument (28-items): tangible, emotional/companionship, financial, information, and language support – Language support not included in this analysis UBC School of Nursing Analysis • Tested for evidence of construct validity using LISREL: – Dimensional invariance – Configural invariance – Metric invariance – Scalar invariance UBC School of Nursing Results • Total sample of 1,074 women – African American (n=159), Chinese (353), Latina (n=236), Caucasian (n=326) – Mean age was 61 years, >50% married – Chinese and Latinas had < high school education – All reported lower household income compared to Caucasian UBC School of Nursing Results 0.980.065< 0.001176.71326Non-Latino White 0.970.07< 0.001164.75236Latino 0.980.05< 0.001134.43353Chinese 0.980.0570.002108.21159African-American 0.990.052< 0.001277.031074Total sample CFIRMSEApχ2NEthnic Group UBC School of Nursing Configural Invariance & Metric Invariance (bolded) 0.970.0670.0001325.05562 Latino vs. Non-Latino White 0.970.0590.0001328.9679 Chinese vs. Non-Latino White 0.970.0610.0001310.62589Chinese vs. Latino 0.980.0650.0001 299.97 485 African-American vs. Non-Latino White 0.970.0680.0001282.28395 African-American vs. Latino 0.980.0530.0001252.19512 African-American vs. Chinese CFIRMSEApχ2NModel UBC School of Nursing Discussion • Comparative research requires that instruments measure constructs with the same meaning across groups • Apparent differences across groups may be spurious and due to additive response bias • This study suggests that this instrument could be used for comparing social support across ethnic groups with more work UBC School of Nursing Ideal vs. Practical • Strict factorial invariance is ideal • Strong invariance is more readily attainable • Metric invariance sufficient if focus is restricted to comparing the strength of corresponding regression parameters across groups UBC School of Nursing UBC Centre for Health Services and Policy Research and School of Nursing Advancing world-class health services and policy research, training and data resources on issues that matter to Canadians. » Putting Scholarship into Practice »


Citation Scheme:


Usage Statistics

Country Views Downloads
China 23 19
Japan 2 0
United States 1 1
City Views Downloads
Beijing 23 0
Tokyo 2 0
Overland Park 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}


Share to:


Related Items