Article Summary
J. Waters

Research Review of Student Factors in Post-Secondary Academic Achievement

During the lit review for my own research on tech distractions & academic achievement, I came across a journal article by Schneider and Preckel (2017) on their impressive systematic literature review of meta-analysis research – thousands of empirical studies on post-secondary student academic achievement. I’ve summarized here the evidence-based student factors they found are strongly related to students’ academic achievement in higher (post-secondary) education. I’ve also written some tips based on psychology research, on this e-portfolio site.

Their systematic review included 105 variables related to students, to instructors, and to instructional methods. Only the factors related to students are included here. If you are interested in the original article, you can go into the PsycINFO database from our Library page and access it directly. The Reference citation for the article is:

Schneider, M. & Preckel, F. (2017). Variables associated with achievement in higher education: A systematic review of meta-analyses. Psychological Bulletin, 143(6), 565-600. doi:

Some Definitions of their Research Method:

A systematic review investigates a research question by collecting, analyzing, and synthesizing the many research studies that have investigated that topic (Uman, 2011). Researchers comb through the research databases to collect these hundreds, or thousands, of studies, according to “inclusion criteria” such as population, the number of participants, outcomes, etc. After collecting the research studies, the researchers then critically analyze the mass of the pooled findings for us. A systematic literature review is very valuable as an extensive survey of the research findings in a single topic. Like a single swallow in spring or a single leaf fall in autumn, a single research study doesn’t signal a conclusion on its own. Research must be replicated! It can take many studies to give us solid, reliable and generalizable data for a theory or conclusion. When you read a particular research finding in a textbook, it usually has been based on evidence from many studies.

A meta-analysis accomplishes this goal using quantitative statistical techniques to combine data from multiple studies of the same variables conducted by different researchers (Cozby & Rawn, 2012). These statistical procedures use effect sizes (Cozby & Rawn, 2012) to compare the findings of a wide assortment of studies, and allow statistical conclusions. Standardizing and pooling the data expands the limited research sample sizes of a single study, and allows an overall understanding of variables, increasing generalizability and reliability. So a systematic review of meta-analysis research, such as this study by Schneider and Preckel (2017) of 38 meta-analyses, can potentially review and include the findings of thousands of research studies on dozens or even hundreds of variables (105 in this case), and a greatly expanded participant sample size (1,920,239 in this case).

The effect size is a quantitative measure of the strength of an association between variables (Cozby & Rawn, 2012; Uman, 2011) (for example, between class attendance and GPA). Translating experimental and correlational research into effect size allows us to compare the effects of variables across these different research methods. A meta-analysis will typically show the effect sizes for a variable from a number of studies, with a summary of the average.

One example of an effect size is a correlation. The larger the absolute value of an effect size, the stronger the relationship or effect. In the case of a correlation, calculated with a Pearson’s correlation (symbol is r), r = 0 is no effect, r = .1 is a very small effect size, r = .3 is a medium effect size, r = .4 is a large effect size, and r = 1.0 is a perfect relationship (Cosby & Rawn, 2012). The correlational measure of effect size is the coefficient of determination, defined as r2.

Even a very small correlation may be statistically significant, if the sample size is huge, as statistical significance just lets us know the probability we found that result by chance. And a small but significant correlation may still be important to know, as it may predict a small but reliable effect.

Some associations are positively correlated, meaning a high score in one variable (such as GPA) relates to and predicts a high score in the other variable (such as a lot of classes attended) and a low score in one predicts a low score in the other (such as lower high school GPA predicts lower university GPA). A negative relationship means that a high score in one variable (such as test anxiety) predicts a low score in the other (such as GPA).

Caution: A correlation shows variables are related, but can never prove that one of the variables causes the other. But knowing that class attendance has a strong correlation with GPA allows us to predict that where we see a group of students with high GPAs we will often see lots of classes attended, and vice versa.

In Schneider and Preckel’s review, a statistic called Cohen’s d was used to compare and summarize different statistics. In an experiment or meta-analysis, Cohen’s d can describe the degree of the effect of the independent variable on the dependent variable (Cosby & Rawn, 2012). In this study, the authors categorized d of 0 to .11 as no effect, a d of .11 to .35 as a small effect, a d of .35 to .66 as a medium effect, and a d larger than .66 as a large effect.

Reasons for the study: (this includes a summary of Schneider and Preckel’s (2017) arguments with additional relevant research):

Although much educational research has previously been done on factors related to K-12 (Kindergarten to grade 12) student achievement, Schneider and Preckel’s (2017) study may be the first systematic review of  post-secondary student achievement. Of course, some of the factors important in predicting university or college student achievement will be similar to ones in K-12 education, but Schneider and Preckel (2017) point out that there are also crucial differences between K-12 and post-secondary education, such as the increased cognitive challenge of higher education as well as the much larger classes and different instructional methods. These factors can significantly affect student achievement, especially in first year, which generally has the highest drop-out rates (Finnie, Childs, & Martinello, 2014). In addition, post-secondary education has tuition costs that are steeply increasing year by year, compared to free public school education in Canada, so the costs of failure or low GPA leading to probation or requirement to withdraw are financially costly as well as costly to the student’s self-esteem and self-efficacy.

The very strong correlation between high school GPA and university GPA (Olani, 2009) suggest that in general, students who have been accepted at university and college are capable of finishing their degrees. However, not all students who enroll in post secondary education complete their degrees, or complete them without repeating courses or within four years. The fact that some don’t complete to graduation or have low GPAs (and lower GPAs have been found to be associated with lower employment and earnings (Finnie, Pavlic, Afshar, Bozkurt & Miyairi, 2016)) underscores the need for this research. According to an OECD Report (2017), there is an impressive employment and earnings advantage for graduates, especially for those with higher GPAs (Finnie et al., 2016). This supports the need for an examination of variables that might help post-secondary students achieve and graduate in a timely way.

Schneider and Preckel’s Systematic Review Process

See Uman’s (2011) on-line article for a review of the eight stages of a systematic review and meta-analysis (

In their systematic review of meta analysis research, Schneider and Preckel (2017) began with a keyword search of the PsycINFO database using the search string (achievement or grades or competence or performance or learning or GPA) and (higher education or college or university or tertiary).

From this they selected and summarized 38 meta-analyses (Schneider & Preckel, 2017) covering 105 variables, 3,330 effect sizes and nearly 2 million students in university or college, published between 1980 and 2014 (most (60%) of the studies were conducted since 2005). They then ordered the 105 variables by effect sizes and summary statistics to determine the factors that are most strongly related to student academic achievement, in terms of student behaviour, instructor behaviour, and instructional method.

Significant factors found by the systematic review were ones with effect sizes that were at or above the overall mean effect size (that is, had a Cohen’s d above the average).

The Research Findings

The following are factors Schneider and Preckel (2017) found were medium to strong effects on students’ academic achievement. These are related to the students’ motivation, ability, behaviour (such as study strategies) or personality. Explanations of these variables follow the Table.

Table 1: Top Student Variables (in rank order) Associated with Academic Achievement

Rank in 105 variables Category Variable Degree of effect size
2 Motivation Performance self-efficacy Very strong
5 Motivation Student set grade goal Very strong
6 Strategies Class attendance Very strong
7 Ability/Prior achievement HS GPA Very strong
10 Ability/Prior achievement Admission test results Strong
13 Strategies Effort regulation Strong
17 Strategies Uses a strategic approach to learning Medium to strong
19 Motivation Achievement motivation Medium to strong
21 Motivation Academic self-efficacy Medium to strong
21 Ability/Prior achievement Content of recommendation letter Medium to strong
30 Ability/Prior achievement Cognitive ability (assessed by standardized tests) Medium
30 Personality Conscientiousness Medium
37 Strategies Time and study management Medium
37 Context Financial support Medium
37 Strategies Peer learning Medium
37 Strategies Learning strategy/organization Medium
43 Strategies Concentration Medium
45 Motivation Academic goals Medium
47 Strategies Help seeking Medium
47 Personality Emotional intelligence Medium
47 Personality Need for cognition Medium
54 Strategies Critical thinking
54 Strategies Time spent studying
54 Strategies Academic skills
54 Motivation Academic intrinsic motivation
58 Personality Locus of control

 Student Factors:

  1. Motivation: five variables have a medium to large effect. In rank order:
    1. Self-efficacy: Performance self-efficacy (d = 1.81, rank 2) A very large to huge effect. This type of self-efficacy would be the students’ expectation of their ability to perform a specific task, an assignment or an exam. It is found to be causally related to achievement, which is itself causally related to self-efficacy.
      1. And Academic self-efficacy (d = 0.58, rank 21). A medium effect. This would be a more global student self-assessment of their academic competency in general.
    2. Grade goals (d = 1.12, rank 5) the students’ minimum standards for their grades.
    3. High achievement motivation (d = 0.64, rank 19) and achievement motivation in general
    4. Academic goals (d = 0.36, rank 45), such as the importance of a particular course to their overall academic goals.
  2. Ability (Intelligence) and prior achievement: these variables have large effect sizes and high ranking among the 105 variables. Ability and prior achievement are strongly related to achievement. in rank order:
    1. High School GPA: (d = 0.9, rank 7). Likely a foundation of knowledge and prior learning demonstrated by these first two variables helps the student make connections to the course material enabling a deeper level of understanding and therefore better achievement.
    2. Standardized admission test results: (d = 0.79, rank 10)
    3. Content of professors’ recommendation letters: (d = 0.58, rank 21) as the professors might be describing the students’ motivation and persistence better than a course grade.
    4. Cognitive standardized tests (e.g. IQ): (d = 0.47, rank 30). This is only a medium-large effect, possibly because IQ tests are more abstract and generic than the specific course learning outcomes assessed in the courses.
  3. Strategies: eleven variables related to student study and learning strategies were found to show strong effect sizes. in rank order:
    1. Class Attendance: (d = 0.98, rank 6). This is a well-established association with academic achievement, and this effect has been repeatedly found over years of research. This is especially interesting in the light of recent online learning additions to the classroom such as mixed mode (blended learning) classes, course management systems, and online tutoring systems. Face to face in-class instruction has a unique and powerful association with student achievement.
    2. Student effort regulation: (d = 0.73, rank 13) This includes student strategies such as increasing persistence and effort when faced with a challenging assignment.
    3. Student strategic approach to learning: (d = 0.65, rank 17) such as using learning strategies that were task dependent. Using a surface approach (with shallow information processing and a focus on external rewards) was negatively related to achievement (d = – 0.39).
    4. Resource management (time management, help-seeking, etc.) and cognitive strategies: such as metacognitive strategies such as organization and critical thinking, and memory strategies such as rehearsal and elaboration, have a small but positive effect, while maladaptive strategies such as procrastination have a negative effect.
  4. Personality: four personal attributes have a large association with achievement:
    1. Conscientiousness: (d = 0.47, rank 30) The rank of this Big 5 trait is as strong as intelligence. It reflects the students’ tendency to be organized, disciplined and industrious. This trait relates to many different behaviours that would lead to academic achievement, such as self-control, planning, persistence, class attendance, goal setting, meeting assignment deadlines, etc.
    2. Test anxiety: (d = – 0.43, rank 104) Negatively related to achievement. Ergene’s meta-analysis of 77 studies (as cited in Schneider & Preckel, 2017), across a total of 2,482 students suggests that cognitive behavioural programs as brief as 4 to 6 hours long, teaching test-taking skills and changing maladaptive cognitions, can significantly reduce test anxiety.
    3. Emotional intelligence: (d = 0.35, rank 47)
    4. Need for cognition: (d = 0.35, rank 47)
    5. Six other variables have a positive but low correlation to achievement: locus of control, optimism, self esteem, Big 5 personality traits of openness to experience and agreeableness, and gender, while 5 other variables have no association to achievement: general self-concept, Big 5 traits of extroversion and emotional stability, depression, and biological age.
  1. Context Variables: These variables have small effect sizes, with only one a medium effect size: whether the student receives financial resources from the institution. (d = 0.41, rank 37).
    1. Surprisingly, factors often assumed to be related to achievement had very low effect sizes, showing little relationship (d of .29 or lower). These factors include students’ feeling of connections within and to the institution, relationships with peers and faculty, students’ social integration with fellow students and social support in general, perception of support from faculty, institutional commitment, involvement in campus activities, or their socio-economic status. These factors were not correlated with academic achievement.

See the page below for some cognitive behaviour tips based on these findings.


Cosby, P. C. & Rawn, C. D. (2012). Methods in behavioural research. (Canadian ed.). US: McGraw Hill.

Engler, B. (2014). Personality theories (9th ed.). Cengage Learning.

Finnie, R., Childs, S. & Martinello, F. (2014). Post-secondary student persistence and pathways: Evidence from the YITS-A in Canada. Education Policy Research Initiative (EPRI).

Finnie, R., Pavlic, D., Afshar, K., Bozkurt, E., Miyairi, M. (2016). Do better grades in post-secondary education lead to better labour market outcomes? Education Policy Research Initiative (EPRI).

Krause, M., Corts, D., Smith, S., & Dolderman, D. (2018). An introduction to psychological science (2nd Canadian ed.). Toronto, ON: Pearson Canada.

Organization for Economic Cooperation and Development (OECD). (2017). ‘Canada’ in Education at a glance 2017: OECD Indicators. OECD Publishing, Paris. doi:

Olani, A.  (2009). Predicting first year university students’ academic success. Electronic Journal of Research in Educational Psychology, 7(3). Special Issue: Monographs on: Motivation and competencies in higher education. 1053-1072. Accession Number: 2010-17405-007

Reisberg, D. (2016). Cognition: Exploring the science of the mind (6th ed.). New York: W.W. Norton & Co.

Schneider, M. & Preckel, F. (2017). Variables associated with achievement in higher education: A systematic review of meta-analyses. Psychological Bulletin, 143(6), 565-600. doi:

Uman, L. S. (2011). Systematic reviews and meta-analyses. Journal of the Canadian Academy of Child and Adolescent Psychiatry, 20(1), 57-59.