A Case Study On Potential Correlation Between Specific Student Characteristics And Academic Dishonesty In Writing

A Case Study On Potential Correlation Between Specific Student Characteristics And Academic Dishonesty In Writing Does academic dishonesty in writing differ among various student demographic categories at Emory? The purpose of this study was to examine potential significant associations between certain student characteristics (college year, gender, GPA, credit hours, hours spent on academics and pressure felt to succeed) and their involvement in academically dishonest activities on writing assignments. The data for this analysis were obtained from a voluntary questionnaire attempted by 228 students in QTM100 during the Fall 2014 semester. However, after applying exclusion criteria, only 195 out of the 228 were included in the analysis. Any student that did not have a fully completed survey was excluded; also, any observation with improbable or impossible responses was also not considered. For example, any student listing a GPA above 4.0, enrolling in credit hours above 22 or below 12, and/or spending over 70 hours per week on academics outside of class was excluded. In the questionnaire, cheating in writing was assessed through six parts, including citing fake page numbers and sources, having someone else write the paper, purchasing the paper online, paraphrasing without acknowledging sources, directly copying words without giving credit, and borrowing ideas without crediting the original source. If a student provided an affirmative response to any of these six parts, he/she was considered a â€Å"yes† for having taken part in academic dishonesty in writing; otherwise, he/she was identified as â€Å"no† for not having participated in such activities. Two-sample t-tests were conducted for the quantitative independent variables, namely GPA, credit hours and time spent on academics. The ?2 test was used for the categorical variables (pressure, gender, grade level). The sample of 195 students studied was majority freshman (51.3%), large majority of female students (73.3%), averaged a GPA around 3.41, was enrolled between 16-17 credit hours and spent an average of almost 17 hours every week on academics outside of class. The overall sample appeared to mostly feel a fairly high amount of pressure to succeed. A statistically significant association (a = 0.05) was only found for gender, where there is strong evidence to suggest that cheating in writing differs between males and females. For the other variables, we failed to find any significant association; rather, there is no apparent difference between those that have committed academic dishonesty in writing and those that have not. By comparing the â€Å"Yes† and â€Å"No† columns for each respective independent variable except gender, it can be observed that means for GPA, credit hours, hours spent on academics, and proportions for pressure felt to succeed between the two groups are largely the same. There are several limitations to what we can infer from this analysis. First, we cannot conclude any causal relationships due to the study’s observational nature. Also, this is not a random sample, so it may not be a good representation of all Emory students. The self-reported nature of the study also means that the accuracy of the data cannot be ascertained. However, in a greater sense of the Emory student population, it can be realized that academic dishonesty in writing is not specific to certain student demographics; many people, regardless of their GPA, the amount of pressure they feel or how busy they are with academics, have been academically dishonest in writing assignments at one point or another; oddly enough, though, it appears females are far more likely to cheat in writing than males. Appendix 1: Data cleaning and recoding. There was an extensive process to remove implausible values from the data set, including the following exclusion criteria. For GPA, initially, the minimum was 1.9 and maximum was 4.2. As it is impossible to obtain a GPA above 4.0, all GPA values above 4.0 were removed. For hours spent on academics, values ranged from 3 to 109; however, since the higher end of this range is fairly unlikely, the limit for hours spent on academics outside of class was decreased to 70, which seems to be a reasonable weekly maximum, after factoring in class time and necessary living tasks. Credit hours were adjusted to only contain values between 12 and 22, since this is the acceptable range for number of credit hours enrolled to qualify as an Emory College student. The original question â€Å"On a scale of 1-10, indicate how much pressure you feel to succeed† naturally yielded quantitative results under the â€Å"pressure† variable, but these numerical values were recoded into 3 categoric al intervals of pressure levels: minimal/moderate (0-5), high (6-8) and very high (9-10). Because QTM 100 is considered an introductory college course, it was unreasonable to expect an even distribution of students between the four college years; more than half of the eligible sample consisted of freshmen (100 of 195). As a result, the categorical variable representing year in college was re-coded into a dichotomous categorical variable, with possible categories of freshman and 2nd year and above. Any observation that failed to meet the above limit criteria, or originally did not have the entire questionnaire complete, were coded as missing and excluded from the analysis. Appendix 2: Statistical analysis. *All mentions of â€Å"cheating† and â€Å"academic dishonesty† refer only to academic dishonesty in writing. â€Å"Yes† and â€Å"No† groups refer to those that have been academically dishonest in writing and those that have not, respectively. Association between college year (freshman, 2nd year and above) and academic dishonesty ?2 test of independence H0: college year and academic dishonesty are independent HA: college year is associated with academic dishonesty. ?2 = 1.45; p-value = 0.2279 At the a = 0.05 significance level, we fail to reject H0 and conclude that academic dishonesty is independent of college year, and there is no statistically significant evidence to suggest there is an association. This result is quite practically meaningful, as it appears that cheating is independent of college grade level. Association between gender (male, female) and academic dishonesty ?2 test of independence H0: academic dishonesty is independent of gender HA: gender is associated with academic dishonesty ?2 = 10.64; p-value = 0.0011. At the a = 0.05 significance level, we reject H0 and conclude that academic dishonesty is associated with gender. Specifically, there is statistically significant evidence to suggest that academic dishonesty is far more prevalent among female than male students. It is hard to say if there is any practical significance in this result, given the largely disproportionate amount of females in the sample, but the very low p-value suggests that in practice, perhaps females are far more susceptible than males to participate in academic dishonesty. Association between GPA and academic dishonesty â€Å"Yes† group: mean GPA, 3.41; standard deviation, 0.36 â€Å"No† group: mean GPA, 3.42; standard deviation, 0.40 Two-sample t-test for difference of means H0:  µYES =  µNO, HA:  µYES ?  µNO, where  µ represents true mean GPA t = 0.34; p-value = 0.7352 There appears to be no apparent difference in GPA between the two groups. Unsurprisingly, at the a = 0.05 significance level, we fail to reject H0 and conclude that academic dishonesty is independent of student GPA, and there is no statistically significant evidence to suggest an association. This result is quite practically meaningful, as it is surprising to observe that students with lower GPAs are not more susceptible to cheating. Association between number of credit hours taken and academic dishonesty â€Å"Yes† group: mean credit hours, 16.37; standard deviation, 2.17 â€Å"No† group: mean credit hours, 16.44; standard deviation, 2.12 2-sample t-test for difference of means H0:  µYES =  µNO, HA:  µYES ?  µNO, where  µ represents true mean number of credit hours taken t = 0.24, p-value = 0.8119 At the a = 0.05 significance level, we fail to reject H0 and conclude that academic dishonesty is independent of number of credit hours taken, and there is no statistically significant evidence to suggest an association. This result is quite surprising as well, as it would be expected that students with a greater course load would have less time on average to allot to individual assignments, leading to a higher amount of cheating; however, this does not appear to be the case. Association between hours spent on academics outside of class and academic dishonesty â€Å"Yes† group: mean hours spent on academics, 16.29; standard deviation, 7.18 â€Å"No† group: mean hours spent on academics, 17.19, standard deviation, 8.71 Two-sample t-test for difference of means H0:  µYES =  µNO, HA:  µYES ?  µNO, where  µ represents true mean hours spent on academics outside of class t = 0.78, p-value = 0.4336. At the a = 0.05 significance level, we fail to reject H0 and conclude that academic dishonesty is independent of number of hours spent on academics outside of class, and there is no statistically significant evidence to suggest an association. This is also surprising, as in practice, it would be expected that those that spend less time on academics tend to be more lazy and find shorter ways to get their writing assignments done. Association between pressure felt to succeed and academic dishonesty ?2 test of independence H0: academic dishonesty is independent of pressure felt HA: pressure felt is associated with academic dishonesty ?2 = 0.50, p-value = 0.7791. At the a = 0.05 significance level, we fail to reject H0 and conclude that academic dishonesty is independent of pressure felt to succeed, and there is no statistically significant evidence to suggest an association. This is quite practically meaningful, as it would be expected that those that felt a lot of pressure to succeed would be more likely to succumb to cheating, but in fact, it is those that feel little/moderate pressure that have the highest proportion of cheaters.