Validity assessment of mibos by a.w. harrison et. al


Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Validity assessment of mibos by a.w. harrison et. al

  1. 1. Research MethodsTheme: Class Room PresentationDate: 20th June 2013.Topic: The Ingratiation Construct: An Assessment of the Validity of theMeasure of Ingratiatory Behaviour in Organisational Settings(MIBOS)Research Scholars: Allison W. Harrison, Wayne A. Hochwarter,Pamela L. Perrewe, David A. Ralston.Source: Journal of Applied Psychology, 1998, Vol.83, No.6, 932-943.Presented by : Abdul Rahman, EFPM 2013.20th June 2013 Abdul Rahman, EFPM 2013 1
  2. 2. PreludeFundamental postulates of this paper:- “Substantive Research without proper construct validationcan lead to dysfunctional consequences” *Schwab, 1980+.- “Excessive measurement error has led researchers toinaccurately reject or accept hypothesis instead ofhaving these outcomes based on theoretical arguments”.[Bagozzi, Yi & Phillips, 1991].20th June 2013 Abdul Rahman, EFPM 2013 2
  3. 3. What is Ingratiatory Behaviour“ A class of strategic behaviours illicitlydesigned to influence a particular otherconcerning the attractiveness of one’s personalqualities” (Jones, 1964, P.11)20th June 2013 Abdul Rahman, EFPM 2013 3
  4. 4. What is MIBOS- Tool / Instrument for measuring Ingratiatory Behaviour in Organisationalsettings.- Developed by Kamalesh Kumar and Michael Bayerlein, 1991.- Salient Features of MIBOS:MIBOSOPINIONCONFIRMITY(7 ITEMS)FAVOURRENDERING (6ITEMS)SELFPRESENTATION(4 ITEMS)OTHERENHANCEMENT(7 ITEMS)20th June 2013 Abdul Rahman, EFPM 2013 4
  5. 5. Checklist of MIBOS developed by KK & MB(1991)Description Y/N RemarksConceptualisation Stage:a) Whether extensive lit. analysis done Y Initial studies refers 1964b) Whether Construct clarity was established Y MIBOS to measuresubordinate’s behaviourtowards superior.c) Whether expert panels / other valid measures usedduring item constructionYd) Whether pretesting of items done Y Pretesting was done withlearned population to validate67 items.e) Whether screening of unrelated items deletedwhile pretestingY 67 items reduced to 55 items.f) Whether individual to total correlation tested Y 55 items reduced to 27items20th June 2013 Abdul Rahman, EFPM 2013 5
  6. 6. Checklist of MIBOS developed by KK & MB(1991) .. contdDescription Y/N Remarksg) Whether Social Desirability Test done Y 27 items reduced to 24 itemsh) Whether Factor Analysis done Y Upto second factor analysis done.i) Whether Reliability Test done for MIBOS Y Internal Consistency and Test – RetestConsistency establishedj) Whether Acquiescence Response Test (ART)done to measure the general tendency factorY (less than 0.11)k) Whether content validity test done Y Done with different set of employeesl) Whether convergent validity test done Y Correlation of 4 factors taken: a) Selfmonitoring Skillb) Need for powerc) Locus of Controld) Machiavellianismm) Whether discriminatory validity done tocheck whether any other factor has significanceinfluence on constructY Factors correlated wereAssertiveness, rationality, exchange ofbenefits, upward appeal andcoalition. (less sf)20th June 2013 Abdul Rahman, EFPM 2013 6
  7. 7. ASSESSMENT OF THE VALIDITY OF MIBOSExhaustive work done by A.W Harrison & others, 1998.Approaches used:1. Confirmatory Factor Analysis (CFA).2. Convergent Validity3. Discriminant Validity.20th June 2013 Abdul Rahman, EFPM 2013 7
  8. 8. Areas of Improvement / shortcomings targetedto address by this research work- To address the psychometric properties of the MIBOS scale(Shrienshein & Hinkin, 1990)- Though MIBOS is a valid measure there are few alternative full scalemeasure of ingratiation is available ( as per Watt, 1993).- In the area of Construct Validity,“Kacman & Valle (1997) reported that 15 of the 24 items doubleloaded and hence stability of the factor structure was not fullyestablished.”- GFI (0.792) and AGFI(0.785) used to check double loading – notacceptable results.- Double loading items were checked which resulted GFI and AGFI as0.836 and 0.785 respectively and concluded as lack of fit.20th June 2013 Abdul Rahman, EFPM 2013 8
  9. 9. Alternate conceptualisation work done by A.W. H& others1. Examined 4 factor model of Kumar & Beyerlein (1991).2. Examined K & B conceptualisation as a second order structurein model taking into consideration the foll. Study:“Kacman and Valle (1997) found that the 4 factors of theMIBOS were highly correlated and therefore may possibly beindicating only one construct”.3. MIBOS as unidimensional.20th June 2013 Abdul Rahman, EFPM 2013 9
  10. 10. Convergent Validity of A.W.H & others“ Campbell and Firske (1959) suggested that Convergent Validity of ameasure is demonstrated when independent methods of measurementlead to comparable results”.Hence A.W.H & others decided to include more factors and assessed therelationship between MIBOS and the foll:1. Locus of Control 7. Self monitoring2. Machiavellianism 8. Self liking3. Resource Scarcity 9. Fear of negative evaluation4. Role clarity and role ambiguity 10. More comparable measures.5. Management Style (autocratic) With the above factors, H1 – H11 wasformed.6. Need for Power20th June 2013 Abdul Rahman, EFPM 2013 10
  11. 11. Alternate Sampling work done by A.W. H &othersExtended the sample group to 4 different set.Sample set 1 288 Managers from manufacturingorganisationSample set 2 144 members of professional organisation.Sample set 3 452 clerical personnel of UniversitySample set 4 279 employed undergraduates.20th June 2013 Abdul Rahman, EFPM 2013 11
  12. 12. Reliability Assessment of MIBOS done by A.W. H& othersSynopsis of Reliability assessment:1. Reliability estimates are satisfactory2. Hinkin (1995) recommended the use of Confirmatory Factor Analysis(CFA) as a rigorous method of examining factor structure.Hence three alternate conceptualisation of ingratiation were analysed withCFA using LISREL 8.3. Harris & Schaubroeck, 1998, James & James 1989 recommends for studyof absolute fit. Hence GFI, AGFI and NFI were used to measure the absolutefit.4. Sweeney & Mc Farlin (1993) recommended the use of “Squared Multiplecorrelation” for construct validity and since it is the “best practice” the samehas been followed.20th June 2013 Abdul Rahman, EFPM 2013 12
  13. 13. Analysis of Results of work done by A.W. H &othersCriteria DescriptionConstruct Validity GFI – Poor for all the three construct models.Sq. Multiple Correlation – Consistently low for all the threemodels.Convergent Validity Hypo: (H1 to H10) not favorable correlation with MIBOS.H11: Somewhat favorable correlation with MIBOS(H11: MIBOS scores should be positively correlated withcomparable measures of ingratiatory tactics)Social Desirability No Correlation existed except few in students set of samplewhich has been ignored.Discrimant Analysis In addition to 5 discriminant factors used by KK & B, threemore are used viz. job focused, supervisor focused and selffocused impression tactics.Final correlation is as large as 0.61 and hence DV to bequestioned.20th June 2013 Abdul Rahman, EFPM 2013 13
  14. 14. Inferences and identified future area of researchof work done by A.W. H & othersInferences FRA1. None of the structural models wereadequate1. Critically assess and perhaps modifyitem working for the existing scale.2. Internal consistency though exists,but not sufficient for construct validity2. Perhaps there are more than fourpossible sets of ingratiation tactics.3. Kacmar & Valley (1997) concludedthat MIBOS might suffer due to itemwording.3. Attempts to be made to compareresults of MIBOS with constructsmeasured with alternate methods.4. All assessments of validity producedweak results.4. May chose to compare MIBOS to asmall number of carefully wordedconstructs that have undergone rigorousconstruct validation.20th June 2013 Abdul Rahman, EFPM 2013 14
  15. 15. Thanks20th June 2013 Abdul Rahman, EFPM 2013 15