Factors Influencing Academic Performance: A Study Among University Students

Background

Performance in academic, usually measured through the examination means, is a crucial goal for universities, colleges and schools (Sampson, 2004). Hoyle (1986) maintained that learning institutions are established with a sole purpose of imparting knowledge as well as different skill sets to all those who go through them with an aim of giving quality performance. There has been a major concern among the quality assurance committee as well the deans on the performance of some students who don’t do well. Their major concern is that the poor performance of some students might affect the overall confidence that people have about their schools. The aim of this study was to investigate the factors that might be contributing to the performance.

Method of data collection

This study was a survey that was conducted among university. A sample of 50 students was collected from university students. A small questionnaire that sought to collect a set of six questions was prepared and administered to the students. The researcher sought to get responses for the following questions;

• What is your age?
• Which gender do you associate yourself with?
• How many roommates do you stay with?
• How many hours do you spend studying in a week?
• What was your SAT score?
• What did you score in last semester exams?

Summary of the data set

As earlier mentioned, a sample of 50 students was included in the study. The dataset is given below;

Table 1: Dataset

Age	gender	Number of roommates	Weekly study hrs.	SAT	Average score
27	2	0	28	1498	84
30	1	1	9	1586	69
30	2	0	23	548	76
21	1	1	20	503	64
27	2	1	25	1086	81
26	1	1	30	1331	70
21	1	1	27	1259	84
30	2	0	33	614	97
28	1	3	29	1390	95
30	1	3	25	1100	72
26	1	2	15	954	58
28	2	0	24	634	52
29	2	0	34	1471	97
26	1	0	12	1489	93
29	2	3	6	566	65
30	2	3	23	1204	54
29	1	1	8	546	55
27	1	1	5	543	50
26	2	1	22	1243	79
28	2	0	25	1528	82
21	2	0	23	1245	80
26	2	0	18	1350	98
30	2	0	9	870	75
22	2	0	19	1046	96
30	2	2	20	1350	55
24	2	1	17	1400	94
25	2	1	14	1356	78
30	1	3	9	852	52
29	2	2	24	1274	76
26	2	1	4	548	61
30	2	2	14	1418	50
28	2	3	8	906	64
29	1	0	10	502	56
21	1	1	35	785	72
21	2	1	13	948	52
29	2	1	20	1250	82
21	1	0	18	731	64
29	2	2	21	652	78
21	1	3	20	1432	53
29	2	1	5	1277	60
28	1	2	23	1201	82
30	1	2	10	649	50
23	2	0	29	1360	80
24	1	1	8	837	50
28	2	1	31	1004	73
25	1	3	26	1495	55
26	2	0	35	1580	98
24	1	2	22	656	60
29	2	0	28	1466	98
30	1	2	10	1355	53

 Gender distribution

42% (n = 21) of the student participants were male while the remaining 58% (n = 29) were female students. 

Descriptive data analysis

In this section, we present the descriptive statistics for the continuous variables. The average age of the participants was found to be 26.72 years old with the oldest participant being 30 years old and the youngest participant being 21 years old. The median age was 28 years old.

Table 2: Descriptive statistics

	Age	Number of roommates	Weekly study hrs.	SAT	Average score
Mean	26.720	1.180	19.320	1077.760	71.440
Standard Error	0.439	0.150	1.240	49.857	2.251
Median	28.000	1.000	20.000	1202.500	72.000
Mode	30.000	1.000	23.000	548.000	50.000
Standard Deviation	3.104	1.063	8.766	352.540	15.918
Sample Variance	9.634	1.130	76.834	124284.104	253.394
Kurtosis	-0.733	-0.974	-1.019	-1.372	-1.191
Skewness	-0.734	0.474	-0.058	-0.315	0.240
Range	9.000	3.000	31.000	1084.000	48.000
Minimum	21.000	0.000	4.000	502.000	50.000
Maximum	30.000	3.000	35.000	1586.000	98.000
Sum	1336.000	59.000	966.000	53888.000	3572.000
Count	50	50	50	50	50

 In terms of the weekly study hours, participants on average spent 19.32 hours to study in a week. The maximum number of hours spent by the participants to study was 35 hours a week while the lowest was 4 hours a week. On average, the students included in the study had a score of 1077.76 in SAT.

Lastly, we found out that the average score of the students was 71.44 with the highest performing student getting 98% while the lowest score being 50%.

Interval estimation  95% confidence interval for average score

Under this we considered average score for the students in their last semester exams. The 95% confidence estimation is given as follows;

CI:

Margin of error (ME)

, ,

We can then replace the values in the equation of the confidence interval as follows;

Methods

CI:

CI:

Lower bound:

Lower bound:

From the above calculations, we are 95% confident that the true mean of average score is between 67.0259 and 75.8541.

95% confidence interval for SAT scores

Under this we considered SAT scores. The 95% confidence estimation is given as follows;

CI:

Margin of error (ME)

, ,

We can then replace the values in the equation of the confidence interval as follows;

CI:

CI:

Lower bound:

Lower bound:

From the above calculations, we are 95% confident that the true mean of SAT scores is between 980.0409 and 1175.4791.

95% confidence interval for Weekly study hours

Under this we considered weekly study hours. The 95% confidence estimation is given as follows;

CI:

Margin of error (ME)

, ,

We can then replace the values in the equation of the confidence interval as follows;

CI:

CI:

Lower bound:

Lower bound:

From the above calculations, we are 95% confident that the true mean of Weekly study hours is between 16.8902 and 21.7498.

Results

We applied a multiple regression model technique which is basically a statistical approach used to predict or rather to estimate the dependent (response) variable pegged on two or more independent (explanatory) variables (Armstrong, 2012). For this section, we sought to investigate the relationship between the years of residency in the current place and three independent variables. The explanatory variables are; religious level of the participant, age of the participant when he/she stopped full-time education and living standard level of the participant.

The description of the variables is given in table 3 below;

Table 3: Description of the variables

Variable	Type of variable	Description
Average score	Dependent	This is a numerical variable such that the participants in the study are required to give the score they obtained in their last semester exam
Age	Independent	This is a numerical variable such that the participants in the study are required to give their age in years
Gender	Independent	Categorical variable that was coded as a dummy variable (male = 1, female = 0)
Roommates	Independent	This is a numerical variable such that the participants in the study are required to give the number of roommates they stay with in their hostel
SAT scores	Independent	This is a numerical variable such that the participants in the study are required to give the score they obtained in their SAT
Hours	Independent	Numerical where the respondents were to give their weekly study hours (amount of hours they spent)

 For this analysis, we sought to predict the following regression model;

Where,

Results of the regression analysis are given in table below;

Table 4: Regression Statistics

Multiple R	0.729952
R Square	0.532831
Adjusted R Square	0.479743
Standard Error	11.48173
Observations	50

 Table 4 above presents the summary of the regression model where we observe the value of R-Squared to be 0.5328; this simply implies that 53.28% of the variation in the dependent is explained by the five independent variables in the model. 

Table 5: ANOVA
	df	SS	MS	F	Significance F
Regression	5	6615.796	1323.159	10.03685	1.86E-06
Residual	44	5800.524	131.8301
Total	49	12416.32

 In the table below (see table 5), we check on the adequacy of the model. From the table, we observe the p-value for the F-Statistics to be 0.000 (this value is definitely less than 5% which is the significance level), as such, we fail to accept the null hypothesis and so conclude that the model is statistically different from zero.

Table 5: regression coefficients

	Coefficients	Standard Error	t Stat	P-value	Lower 95%	Upper 95%
Intercept	38.171	15.841	2.410	0.020	6.245	70.097
Age	0.236	0.568	0.415	0.680	-0.910	1.382
gender	4.588	3.677	1.248	0.219	-2.822	11.999
Number of roommates	-5.264	1.712	-3.076	0.004	-8.714	-1.815
Weekly study hrs	0.689	0.207	3.333	0.002	0.273	1.106
SAT	0.012	0.005	2.332	0.024	0.002	0.022

Looking at table 5, we see that 3 out of the 5 independent variables are significant in the model (p-value < 0.05). The 2 significant variables are number of roommates, weekly hours and Sat scores. However age and gender of the student were found to be insignificant.

Results

The coefficient of number of roommates was found to be -5.264; this means that a unit increase in the number of roommates would result to a decrease in the average score of the student by 5.264. Similarly, a unit decrease in the number of roommates would result to an increase in the average score of the student by 5.264.

The coefficient of weekly study hours is 0.689; this means that a unit increase in the weekly hours would result to an increase in the average score by 0.689. Similarly, a unit decrease in the weekly hours would result to a decrease in the average score by 0.689.

The coefficient of SAT scores is 0.012; this means that a unit increase in the SAT scores would result to an increase in the average score by 0.012. Similarly, a unit decrease in the SAT score would result to a decrease in the average score by 0.012.

The coefficient of the intercept is 38.171; this means that holding other factors constant (zero values for the independent variables) we would expect the average score to be 38.171.

From the above results and analysis, assuming we consider the three variables in the model, we would have the model to be as follows;

Replacing the values obtained we have the model as;

Where,

Conclusion and recommendations

The aim of this study was to investigate the factors that influence on the performance of the university students. Results revealed that number of roommates, weekly study hours and SAT scores significantly affect the student performance. Age and gender on the other hand had no significant influence on the average score. Based on this, it would be recommendable that the university management look at ways of creating more space for the students so that the issue of sharing rooms can be reduced since it has significant influence on the student performance. The students also need to be advised on the need to be serious with their studies and commit more time for the study since the number of hours spent on studying significantly influences on the average score. 

References

Armstrong, J. S., 2012. Illusions in Regression Analysis. International Journal of Forecasting (forthcoming), 28(3), p. 689.

Farrell, P. J. & Rogers-Stewart, K., 2006. Comprehensive study of tests for normality and symmetry: extending the Spiegelhalter test. Journal of Statistical Computation and Simulation, 76(9), p. 803 – 816.

Good, P. I. & Hardin, J. W., 2009. Common Errors in Statistics.

Razali, N. & Wah, Y. B., 2011. Power comparisons of Shapiro–Wilk, Kolmogorov–Smirnov, Lilliefors and Anderson–Darling tests. Journal of Statistical Modeling and Analytics, 2(1), p. 21–33.

Rice, J. A., 2006. Mathematical Statistics and Data Analysis.

Sampson, D., 2004. Academic performance, mental health and school factors.

Sawilowsky, S. S., 2005. Misconceptions Leading to Choosing the t Test Over The Wilcoxon Mann–Whitney Test for Shift in Location Parameter. Journal of Modern Applied Statistical Methods, 4(2), p. 598–600.

Shapiro, S. S. & Wilk, M. B., 1965. An analysis of variance test for normality (complete samples). Biometrika, 52 (3-4), p. 591–611.

Tofallis, C., 2008. Investment Volatility: A Critique of Standard Beta Estimation and a Simple Way Forward. European Journal of Operational Research, 187(3), p. 1358–1367.

Tofallis, C., 2009. Least Squares Percentage Regression. Journal of Modern Applied Statistical Methods, 7(6), p. 526–534.

Tofallis, C., 2009. Least Squares Percentage Regression. Journal of Modern Applied Statistical Methods, 7(5), p. 526–534.

YangJing, L., 2009. Human age estimation by metric learning for regression problems. International Conference on Computer Analysis of Images and Patterns, p. 74–82.