Testing Reliability of a Research Instrument

TESTING RELIABILITY 1

TestingReliability of a Research Instrument

UniversityAffiliation

The study was designed to evaluate the test-retest reliability andinternal consistency of the scales of the Spanish version of theMinnesota Multiphasic Personality Inventory-Adolescent (MMPI-A). Inthis regard, the researchers used two samples of 939 and 109 Spanishadolescents aged 14-18 years (Zubeidat et al., 2011). All thestudents were assessed with the MMPI-A within their schoolenvironment. Notably, the first sample of participants responded tothe inventory once. On the other hand, the second sample reacted toit on two occasions with a two-week interval between sessions(Zubeidat et al., 2011). The study results showed no significantdifferences in variances or means between the first and the secondtest administration for most MMPI-A scales.

Furthermore, test-retest reliability ranged between .62 (Amorality,Ma1) and .92 (Immaturity, IMM) while most correlationsexceeded .70 (Zubeidat et al., 2011). Generally, internal consistencyvalues for the MMPI-A scales in the pretest and posttest were quitesimilar. Conversely, external validity of the MMPI-A was demonstratedthrough various significant correlations between its scales andYSR/11-18 syndromes and measures of social interaction. Consequently,the highest correlations were established between theAnxious/Depressed YSR/11-18 scale and other MMPI-A scales such asAdolescent-Alienation, Adolescent-Anxiety, Welsh’s Anxiety, andSchizophrenia (Zubeidat et al., 2011). Other correlations wereobserved between the Social Avoidance and Distress Scale and theMMPI-A Adolescent-Social Discomfort scale.

The study evaluated the reliability of the MMPI-A clinical scale. Inthis regard, the coefficient was estimated using the test-retestprocedure. The researchers calculated the Pearson product-momentcorrelation value between the two sets of scores of the sameindividuals on both occasions. I have confidence in the researchdesign due to several reasons. For example, nonstudent samples werenot used in the study. Although 26 secondary schools were randomlyselected, only 13 agreed to participate (Zubeidat et al., 2011). Allassessments were done within the school environment to limit theinfluence of external stimuli. The inclusion criterion was alsoadhered to since the study did not deliberately exclude any school.However, the researchers excluded adolescents not attending school.Such an assumption was based on the fact that education wascompulsory in Spain until the age of 16 (Zubeidat et al., 2011).Consequently, there are very few marginalized adolescents aged 16-18years.

Besides, I have confidence in the research analyses of thereliability of the instrument. The study used several statisticaltechniques to examine the relationships between the two samples. Forexample, equality of variances was evaluated using the Pitman-Morgantest. The Bonferroni correction was also used for the same 67variables (Zubeidat et al., 2011). In addition, Hotelling’s testwas applied to examine the equality of means.

Indeed, I would have confidence in the instrument based on thereliability of the instrument as demonstrated in the study. The studyresults proved that correlations existed between various tests. Themethodology was also sufficient to determine that similar resultswould be obtained using different sets of data. Furthermore, I wouldhave no reservations concerning the use of the instrument in futureresearch. Similar studies have made significant contributions towardsthe understanding of the psychopathology of adolescents (Leeth,2011). Therefore, the instrument could help to study the psychometriccharacteristics of future participants (Handel et al., 2011). Infuture research, I would like to determine the reliability of theinstrument using adolescents not attending school. Furthermore, Iwould also examine the impact of a test-retest analysis using morethan two points in time.

It is crucial to demonstrate reliability in the population andculture sampled in the study. The participants of the study wereselected based on convenience. Although several schools were chosenrandomly, only 13 agreed to participate in the study (Zubeidat etal., 2011). Moreover, the researchers excluded a sample ofadolescents not attending school. In this regard, it would be quiteimpractical to generalize the results to other populations orcultures. Future researchers may have to conduct the study usingrandom participants with diverse backgrounds.

References

Handel, R. W., Archer, R. P., Elkins, D. E., Mason, J. A., &ampSimonds-Bisbee, E. C. (2011). Psychometric properties of theMinnesota Multiphasic Personality Inventory–Adolescent (MMPI–A)clinical, content, and supplementary scales in a forensic sample.Journal of personality assessment, 93(6), 566-581.

Leeth, C. (2011). Minnesota Multiphasic PersonalityInventory-Adolescent (MMPI-A). In Encyclopedia of Child Behaviorand Development (pp. 956-957). New York, NY: Springer.

Zubeidat, I., Sierra, J., Salinas, J., &amp Rojas-García, A.(2011). Reliability and Validity of the Spanish Version of theMinnesota Multiphasic Personality Inventory–Adolescent (MMPI–A).Journal of Personality Assessment, 93(1), 26-32.http://dx.doi.org/10.1080/00223891.2011.529011