AssessingNon-Parametric or T- Tests
AssessingNon-Parametric or T- Tests
At-test is used to test for significance in that one can ascertainwhether there is a significant difference between the means of twosamples of data or populations. As a result, a person can establishthese populations share the same characteristics or if they are quitedifferent from one another (Cohen, Manion, & Morrison, 2013). At-test that is considered to be statistically significant is one thatshows the differences between two groups of data having nopossibility of occurring because the sample was not quite typical.The more the t-test is significant, the more the samples aredifferent from one another (Satorra, & Bentler, 2011). One of theexamples where a t-test could be applied is where one wants to testhow people of the same age group but from different states respond tothe same type of phone from a given company.
Inthe cases where there is no significant difference as shown in thet-test results that could be obtained, this will show that the twopopulations share the same characteristics as it relates to the viewsthat they have regarding the same type of phone. The reverse is alsotrue as it relates to this aspect. This paper shall aim to establishhow a t-test was used to describe the results obtained from a studythat aimed to establish the effect of substance abuse on the academicperformance of students who are in their adolescent stage in Nigeria(Akanbi, Augustina, Theophilus, Muritala, & Ajiboye, 2015).
Thestudy was carried out in the Kwara state of Nigeria where theresearchers aimed to establish whether cases of substance abuse couldhave an effect on their academic performance. This study wasnecessitated by various factors among them including the aspectestablished that a large number of people in their adolescent stagefound in the area angaged in drug abuse (Akanbi et al., 2015).Another issue lay in the fact that there was a general decline in thelevel of academic performance of the students found in the region andas a result, it was important to establish whether this trend was inany way related to the practice of the students as it related tosubstance abuse.
Thestudy applied the use of a survey study design in that theresearchers were interested in how the situation was during the givenperiod when the study was still taking place. The study applied theuse of simple random sampling to obtain the people who would takepart in the study. The respondents included 150 students drawn fromthree college institutions (Akanbi et al., 2015). The researcherswere also keen to ensure that these students belonged to a differentdepartment in these colleges so as to ensure that more representativeand valid results would be obtained in the long-run. The researcherapplied the use of the Drug Habit Inventory to collect data. Thisinstrument is regarded to be valid and offers a high chance ofobtaining reliable data.
Thestatistical analysis conducted on the results applied the use oft-test so as to come up with conclusions that would clearly comparethe populations. The analysis would also establish conclusions thatwould be quite crucial in answering the research question. Thefinding of the study clearly showed that there is a significantdifference between students who abused substances and their academicperformance. As a result, the researchers were able to come up withvarious recommendations that would be quite crucial in helping toalleviate the problem. This lay in the fact that it was important forall the stakeholders involved to carry out sensitization and behaviorchange campaigns against drug and substance abuse among students soas to help protect them from the problems that they face as a resultof being exposed to the use of drugs. They also proposed theestablishment of counseling centers that would go a long way inhelping those students who were already hooked to drug abuse to stop.These centers would be quite effective if set up in every publiccollege so as to reach as many students as possible.
Howthe statistic was used in the study
Thet-test statistic was used in the study to test for the relationshipbetween academic performance and drug abuse. A t value of 2.661 witha p-value of 0.009 measured at 0.05 level of significance showedclearly that the null hypothesis would be rejected and thealternative hypothesis adopted instead. This indicated that it wastrue that cases of drug abuse among the adolescents had a high chanceof affecting the academic performance of the students in a negativeway (Akanbi et al., 2015). The t-statistic was also used to test forthe difference in gender regarding how they abuse drugs and theiracademic performance. A t-value of 3.506 with a p-value of 0.001 wasobtained. This also showed that the null hypothesis would berejected, and instead, the alternative hypothesis would be adopted.This showed that there was a significant difference in genderregarding their abuse of drugs and the effect that such a practicecould have on their academic performance.
Explanationof how this is appropriate
Theuse of the t-statistic to test for significance was quite crucial,especially as it related to finding significant differences in thedifferent populations that were represented in the study among themwhich was gender (Akanbi et al., 2015). As a result, the researcherswere more able to establish the findings that would enable them tocome to appropriate conclusions that they would most likely rely on.
Explanationof how assumptions of the test were met
Theresearch study was based on various assumptions that played a hugerole as it relates to the formulation of the provisions that thestudy would need to follow. One of them lay in the fact that theresearchers came to the field with the assumption that there was highchance that there was some level of association between the practiceof substance abuse among the adolescent and the decrease in thelevels of academic performance that was observed among them (Akanbiet al., 2015). The researchers also acted on the assumption that thesample size chosen would act as a proper representation of theirentire population that was under study. Both of these assumptionswere achieved in that the findings obtained during the study werequite reliable and provided conclusions that were key in theformulation of recommendations that would go a long way in helping todeal with the problems of drug abuse and reduced academic performanceamong the adolescent students.
Levelsof measurement of variables
Variousstudy variables came into the picture during the study. Theseincluded the level of drug abuse among the students as well as theiracademic performance. Data that pertained to the level of drug abusethe students took part in was obtained from the Drug Habit Inventorythat was key in providing information the researchers could rely on(Akanbi et al., 2015). Information that related to the academicperformance of the students was obtained from the General andEducational Studies of the students in that it would be quite easyfor the researchers to compare the level of performance of thestudents over a period and establish a pattern that clearly showedthat there was a significant drop in the performance, an aspect thatshowed that there could have been something that was behind thesignificant drop.
Descriptionof the appropriateness of the level of measurement
Thislevel of measurement of the study variables was quite appropriate asit ensured that the researchers are able to obtain results that wouldbe quite representative of the population that was under study(Akanbi et al., 2015). This was important as the results obtainedwould be considered to be quite factual.
Howthe data was displayed
Datawas displayed in tables that showed the test of significance usingthe t-test (Akanbi et al., 2015). This was quite appropriate as itbrought about some element of making the data presentable.
Thestudy had the strengths of having proper statistical analysis methodsthat were quite easy to follow and, therefore, quite understandable,making it appropriate (Akanbi et al., 2015). The weakness associatedwith the study was that it relied on the use of very few studyvariables, and, therefore, this was not quite appropriate as verylittle information could be obtained from the study. It would be farmuch better if the researchers used more variables to address thisproblem.
Akanbi,M. I., Augustina, G., Theophilus, A. B., Muritala, M., & Ajiboye,A. S. (2015). Impact of Substance Abuse on Academic Performance AmongAdolescent Students of Colleges of Education in Kwara State,Nigeria.Journalof Education and Practice, 6(28),108-112.
Cohen,L., Manion, L., & Morrison, K. (2013). Researchmethods in education.Routledge.
Satorra,A., & Bentler, P. (2011). Scaling corrections for statistics incovariance structure analysis. Departmentof Statistics, UCLA.