Reliability Coefficients Accounting for Correlation between Responses for Family Members

2.50
Hdl Handle:
http://hdl.handle.net/10755/160225
Type:
Presentation
Title:
Reliability Coefficients Accounting for Correlation between Responses for Family Members
Abstract:
Reliability Coefficients Accounting for Correlation between Responses for Family Members
Conference Sponsor:Midwest Nursing Research Society
Conference Year:2007
Author:Knafl, Kathleen, PhD
P.I. Institution Name:Oregon Health & Science University
Contact Address:, 3455 Southwest US Veterans Hospital Road, Portland, OR, 97239, USA
Co-Authors:G.J. Knafl, J.P. O'Malley, <u>K. Knafl, </u>School of Nursing, Oregon Health & Science University, Portland, OR; J.K. Dixon, M. Grey, School of Nursing, Yale University, New Haven, CT;
In family research, survey responses for family member dyads are correlated, so standard reliability coefficients are not appropriately computed using the combined data for all family members. Therefore, new methods were developed for calculating composite reliability coefficients that account for such intra-familial correlation. These methods are described and demonstrated. Linear mixed modeling was used to generalize internal consistency reliability, test-retest reliability, and the item-total correlation to scale responses for one or two family members per participating family. These methods are illustrated with analyses of responses to the Family Management Measure (FMM), based on the Family Management Style Framework, from fathers and mothers on their family's management of their child's chronic condition. Specifically, composite reliability is investigated for the FMM mutuality items assessing how partnered parents share the management of their child's condition. The FMM data included responses to 9 mutuality items (out of 65) from 514 partnered parents (not including 65 single mothers) from 352 families, including 324 (63%) parents with participating partners (162 families). Retests were conducted within 2-4 weeks for 61 of these parents from 40 families, including 42 (69%) parents with participating partners (21 families). Intra-familial correlation for the scale was .46 (p<.01) and ranged from .17 to .42 for the items, with all but one significant (p<.05). Mutuality had composite internal consistency reliability of .77, increasing with removal of one item to .78, and composite test-retest reliability of .72, increasing to .75 with removal of that item. Composite item-total correlation was lowest at .23 for that item and otherwise ranged from .28 to .64. In conclusion, correlations between family members may be substantial, so in the past reliability coefficients could be appropriately computed only for each type of family member separately, not for both together. However, composite reliability coefficients, which account for such intra-familial correlation, now can be computed using data for all family members. These provide more appropriate summary measures of the overall reliability of scale responses for all surveyed family members.
Repository Posting Date:
26-Oct-2011
Date of Publication:
17-Oct-2011
Sponsors:
Midwest Nursing Research Society

Full metadata record

DC FieldValue Language
dc.typePresentationen_GB
dc.titleReliability Coefficients Accounting for Correlation between Responses for Family Membersen_GB
dc.identifier.urihttp://hdl.handle.net/10755/160225-
dc.description.abstract<table><tr><td colspan="2" class="item-title">Reliability Coefficients Accounting for Correlation between Responses for Family Members</td></tr><tr class="item-sponsor"><td class="label">Conference Sponsor:</td><td class="value">Midwest Nursing Research Society</td></tr><tr class="item-year"><td class="label">Conference Year:</td><td class="value">2007</td></tr><tr class="item-author"><td class="label">Author:</td><td class="value">Knafl, Kathleen, PhD</td></tr><tr class="item-institute"><td class="label">P.I. Institution Name:</td><td class="value">Oregon Health &amp; Science University</td></tr><tr class="item-address"><td class="label">Contact Address:</td><td class="value">, 3455 Southwest US Veterans Hospital Road, Portland, OR, 97239, USA</td></tr><tr class="item-email"><td class="label">Email:</td><td class="value">knaflk@ohsu.edu</td></tr><tr class="item-co-authors"><td class="label">Co-Authors:</td><td class="value">G.J. Knafl, J.P. O'Malley, &lt;u&gt;K. Knafl, &lt;/u&gt;School of Nursing, Oregon Health &amp; Science University, Portland, OR; J.K. Dixon, M. Grey, School of Nursing, Yale University, New Haven, CT;</td></tr><tr><td colspan="2" class="item-abstract">In family research, survey responses for family member dyads are correlated, so standard reliability coefficients are not appropriately computed using the combined data for all family members. Therefore, new methods were developed for calculating composite reliability coefficients that account for such intra-familial correlation. These methods are described and demonstrated. Linear mixed modeling was used to generalize internal consistency reliability, test-retest reliability, and the item-total correlation to scale responses for one or two family members per participating family. These methods are illustrated with analyses of responses to the Family Management Measure (FMM), based on the Family Management Style Framework, from fathers and mothers on their family's management of their child's chronic condition. Specifically, composite reliability is investigated for the FMM mutuality items assessing how partnered parents share the management of their child's condition. The FMM data included responses to 9 mutuality items (out of 65) from 514 partnered parents (not including 65 single mothers) from 352 families, including 324 (63%) parents with participating partners (162 families). Retests were conducted within 2-4 weeks for 61 of these parents from 40 families, including 42 (69%) parents with participating partners (21 families). Intra-familial correlation for the scale was .46 (p&lt;.01) and ranged from .17 to .42 for the items, with all but one significant (p&lt;.05). Mutuality had composite internal consistency reliability of .77, increasing with removal of one item to .78, and composite test-retest reliability of .72, increasing to .75 with removal of that item. Composite item-total correlation was lowest at .23 for that item and otherwise ranged from .28 to .64. In conclusion, correlations between family members may be substantial, so in the past reliability coefficients could be appropriately computed only for each type of family member separately, not for both together. However, composite reliability coefficients, which account for such intra-familial correlation, now can be computed using data for all family members. These provide more appropriate summary measures of the overall reliability of scale responses for all surveyed family members.</td></tr></table>en_GB
dc.date.available2011-10-26T22:44:33Z-
dc.date.issued2011-10-17en_GB
dc.date.accessioned2011-10-26T22:44:33Z-
dc.description.sponsorshipMidwest Nursing Research Societyen_GB
All Items in this repository are protected by copyright, with all rights reserved, unless otherwise indicated.