The relationship between study skills and learning outcomes: A meta-analysis.

This paper reports the results of a meta-analysis of 52 studies that investigated the relationship between a range of study strategies and outcomes measures. Low correlations were found between a range of different types of study skills and various outcome measures. Having many study skills (i.e. versatility), as assessed by total study skills scores, produced the largest correlations with both cognitive and affective outcomes. Various deep and achieving approaches were positively related to outcomes. Surface approaches were negatively related to outcomes, although many surface strategies such as inflexibility and reproducing were unrelated to outcomes. Merely increasing time-on-task was not correlated to outcomes. Results are interpreted not as an indictment of the usefulness of engaging in good study behaviour. It is suggested that students do not do this of their own accord; however, when they are taught to implement effective learning behaviours, then cognitive and affective outcomes are enhanced.

There have been many studies investigating the relationship between various study skills and learning outcomes. The typical study of this genre specifies a variety of study skills and then correlates the scores on tests that measure students' use of these skills with some achievement outcome, typically GPA. The results of studies that have found a positive association between achievement and the use of a particular strategy or set of strategies have been used by writers of many study skills programs to justify teaching students a pot pourri of study methods. The skills typically stressed are related to organisation and management of time, setting goals, textbook study methods (such as scanning, underlining, SQ3R), memorising, using the library, essay writing, and preparing for and taking examinations.

Not all methods that students use in learning situations are viewed as wise choices. The term `learning pathologies' is applied to non-strategic behaviours that hinder rather than help in learning, often because they are the antithesis of those behaviours that have been shown to assist in learning. For instance, some study skills instruments measure the extent to which students are disorganised, test anxious, absent from school, work avoidant, or globetrotting (over-ready to jump to conclusions). There are conflicting views about how much other learning behaviours can be described as pathological. Strategies of memorisation, for example, are promoted in some instances as being appropriate (such as when studying for a vocabulary test in a foreign language), whereas in other situations the use of memorisation strategies leads students to focus on surface detail at the cost of seeing relationships between different pieces of information and ideas (for example, when trying to memorise the `structure' of a novel).

Regardless of whether a study skill is perceived to be helpful or detrimental to academic performance, the theory of study skills is more sophisticated than implied by this `dust bowl of empiricism' approach. It is increasingly becoming clear that there is not a best set of study skills. Nist, Simpson, and Hogrebe (1985) criticised some of the studies that have compared study methods on the grounds that they were not conducted in naturalistic settings in which students were free to select and employ the strategies they preferred. Often the experimental conditions were such that students were taught a particular method or methods and then placed in a situation in which they were instructed to use what they had been taught. Their achievement was then compared with that of students who had been instructed in a different method, or who had received no training at all.

In a previous meta-analysis, we investigated the effects of study skills interventions on outcomes (Hattie, Biggs, & Purdie, 1996). The meta-analysis of 270 effect-sizes from 51 studies demonstrated that study skill intervention programs, in general, do work most of the time. The results supported the notion of situated cognition (Brown, Collins, & Duguid, 1989) whereby it is recommended that training other than for simple mnemonic performance should be in context, use tasks within the same domain as the target content, and promote a high degree of learner activity and metacognitive awareness. Strategy training should be seen as a balanced system in which the individual's abilities and insights and sense of responsibility are brought into use so that the strategies that are appropriate to the task at hand can be used. Of course, the student will need to know what those strategies are, and also become aware of the conditional knowledge that empowers them: the how, when, where, and why of their use.

Theories and research related to situated cognition and to self-regulated learning demonstrate that it is not some inherent quality in the study skill itself that will lead to improved learning outcomes (in particular, higher grades), but rather the decisions that students make about how and when they are going to use which particular set of skills to achieve a particular purpose or goal. Metacognitive awareness allows students to assess task requirements and situational restraints, and to be flexible in their choice of strategy to suit those conditions. Accordingly, we were particularly interested in this study to explore the relationship between versatility in study behaviour and outcomes. Instead of focusing on each study skill as a separate action, versatility or flexibility can be seen to represent a student's ability to engage in a range of study or learning behaviours, depending on the demands of task and context. Such a notion of versatility is signified in Derry's (1988) distinction between specific learning tactics and learning strategies. Derry proposed that a learning strategy is a complete plan formulated to accomplish a learning goal, whereas a learning tactic is any individual processing technique (such as rehearsal, imaging, and outlining) used in service of the plan. Devising an appropriate plan or strategy involves more than an indiscriminate combining of the tactics at one's disposal. Bloom (1984) argued that to be successfully combined, methods for learning must deal with different aspects of the learning task or learning context. It is possible that some methods compete with each other, particularly in clearly defined contexts or when applied to specific tasks. Nevertheless, if the most appropriate learning plan is to be made, students must be aware of, and able to use as wide a range of specific skills as possible.

The research on self-regulated learning has demonstrated that effective learners possess high levels of metacognitive awareness in relation to the planning, instigating, and monitoring of learning behaviours (e.g. Zimmerman & Martinez-Pons, 1986, 1990). Such research suggests that it is not reliance on a single strategy or small set of strategies that is associated with good academic performance. When strategy use has been measured, the students with higher overall strategy scores achieve more highly than those with lower overall strategy scores. Cantwell and Beamish (1994) specifically examined the question of versatility in self-regulated learning in a study that investigated the planning and orchestration of strategy choices in secondary and tertiary populations. They maintained that the conception of self-regulated learning has generally been unidimensional -- the possession or non-possession of the positive attributes of effective self-regulation. Their study was based on the proposal that self-regulated learning may embrace qualitatively distinct conceptualisations of self-regulation: (a) adaptive self-regulation, marked by purposeful planning, and flexibility in the choice of strategies; (b) inflexible self-regulation, marked by an unwillingness to depart from routine behaviours; and (c) ambivalent self-regulation, marked by an inability to coherently generate or orchestrate processing options in the face of uncertainty. The results of their study indicated that flexibility was associated with better performance in academic learning, whereas both inflexibility and ambivalence were associated with markedly less successful learning outcomes.

As well as a student's academic performance possibly being related to the ability to use a range of specific study skills, some have argued that the amount of time a student spends on a task will be related to learning outcomes. Over three decades ago, Carroll (1963) conceived the degree of learning to be a function of the ratio of the amount of time the learner actually spends on the learning task to the total amount of time needed. Since then, and particularly in the decade spanning the 1980s, models of educational productivity have included time on task as an important predictor of academic achievement and attitude (e.g. Walberg, Fraser, & Welch, 1986). In particular, homework time has been shown to have a positive impact on achievement (Keith, Reimers, Fehrmann, Pottebaum, & Aubey, 1986). On the other hand, Krantz (1983) argued that certain qualitative aspects of studying (such as the particular strategies selected to successfully complete a task) may be more important than the duration of studying. Similarly Karweit (1984) argued:

   Time is a necessary, but not sufficient, condition for learning. Learning
   takes time, but providing time does not in itself ensure that learning will
   take place. More time may result in more learning -- if adequate time was
   the major cause of the problem in the first place. If other factors were
   the real cause, then providing more time will not be an effective strategy.
   (p.33)

In a synthesis of research on time and learning (Walberg, 1988), the notion of productive time (that fraction of lesson and study time that students spend on appropriate learning activities) emerged as even more important than engaged time or time-on-task. Although the effects of time allocation and engagement in learning tasks were consistent (but modest), Walberg suggested that more can be achieved by focusing on productive time. Walberg also noted that time could be negatively correlated with achievement `if for example, students were pressed beyond exhaustion, or a school imposed a rigorous standard of achievement and slower students studied more but attained less' (p.84).

The relationship between time and learning outcome has been explored from the perspective of Lerner's (1971) `just-world' hypothesis which proposes that people have a need to believe that they live in a just world -- a world in which people get what they deserve and deserve what they get. When there is evidence to counter this belief, it is typically ignored. Hill (1990) proposed and tested a corollary to the just-world hypothesis; that is, effort (time spent studying) leads to reward (improved academic achievement). After two investigations that examined the correlation between study time and a range of achievement measures, Hill concluded that such a correlation was, for all practical purposes, non-existent. Nevertheless it seems that Lerner's hypothesis is alive and well -- that is, evidence to counter the belief that more time spent on learning leads to better learning outcomes has been largely ignored.

The intent of the present study was to apply meta-analytic techniques to further explore the issues of versatility and time in relation to study skills. Specifically the aims of the study were threefold: (a) to assess which study strategies or study motives are most related to outcomes; (b) to investigate whether versatility is the most successful strategy; and (c) to compare the effectiveness (as measured by correlation with academic achievement) of a range of study strategies with that of time spent in studying.

Method

Sample of studies

We first searched various computer-based information sources using the keywords: study skills, learning strategies, learning processes, cognitive style, study habits, cognitive strategies, cognitive processes, learning style, metacognitive skills, and thinking skills. These keywords were crossed with achievement, ability and various subject topics. The keywords were searched using Psychological Abstracts (1985 to 1992), and the Educational Resources Information Center (1985 to 1992). After locating various articles, we then searched the references for further studies. Criteria for including studies in the sample were that (a) they were concerned with study skills but were not designed specifically to assess the effectiveness of an intervention (this was the focus of our other meta-analysis, which used a different set of articles); (b) it was possible to calculate a correlation; and (c) the outcome measure was related to either academic performance or affect. This yielded the present sample of 52 studies. (A list of the studies used in the meta-analysis is available from the authors upon request.) There were some studies with more than one sample, and most had multiple indicators of the variables of interest. As a consequence, 653 correlations could be coded.

Variables coded from each study

The following general information was coded from each study: publication year, publication form (journal article, book chapter, or thesis), and sample size. The study skills measures were coded into two levels of categories, primarily using the Biggs (1987) classification scheme as follows. At the more general level, the categories included achieving approach (divided further into achieving motive and achieving strategy), deep approach (deep motive and deep strategy), surface approach (surface motive and surface strategy), general study skills, and learning pathologies (which included globetrotting, negative attitudes, improvidence, disorganised, work avoidance, procrastinative metacognition, and negative attitudes).

At the second level many of these were further subdivided. Achieving strategy included organisation (including scheduling, organising strategy), searching (e.g. visual cueing), note taking (and also including highlighting, identifying main ideas, underlining, reviewing notes), examination technique (also focusing on test relevance), and other (such as textbook reading, methodological study). Deep motivation included attitude and self-efficacy (e.g. non-dogmatism, self-efficacy, study attitudes, independence), internal locus of control, intrinsic motivation (also task motivation). Deep strategy included depth and complexity (e.g. tolerance of ambiguity, wide reading, academic complexity, adaptive strategy management, comprehension learning) deep and elaborative processing (such as hyperprocessing, elaboration, constructive processing, synthesis analysis), and other deep strategies. Surface strategies included memorisation (and recall, rehearsal, fact rote learning), reproducing (fact retention, duplicative processing, labelling, simplifying), inflexibility (e.g. intolerance of ambiguity, cognitive simplicity), and other surface strategies (usually unspecified). The study skills category related to self-regulation (also monitoring, metacognition), time-on-task (e.g. hours studied), and many were a combination of many study skills which we term `total study skills'. A high score on total study skills indicates that the students used a variety of strategies and thus were most versatile.

The outcome measures were coded into eight major classifications: ability, general achievement, subject-based achievement (in mathematics, science, language skills), increasing memory, changing self-efficacy or self-concept, attitude, and study skills. This last category was included to accommodate one study in which locus of control measures were correlated with study skills outcome measures.

Two characteristics of the research design were coded. The studies were graded according to quality (coded independently and agreed to by both authors and classified as low, medium, and high). Where there were disagreements about this coding, a third person also coded the article, and then all disagreements were resolved by discussion. The purpose of study (specifically related to study or learning skills, or study skills was secondary) was also coded.

Three characteristics of the participants were coded: age (primary/ elementary, junior secondary, secondary, college/university, adults); ability level (low, medium, high, mixed, underachieving); and socioeconomic status (low, middle, upper, mixed). The latter two categories were reported on too few occasions, and where reported, tended to be mixed socioeconomic groups or mixed ability.

The correlation indicates the magnitude and direction of association between the study skill and outcome. In order to normalise the distribution of r and to stabilise the variance, Fisher's z-transformation can be used. The correlations were converted to zrs using Fisher's transformation. After the various statistics were calculated (mean, sd, homogeneity), the zrs were converted back to correlations to facilitate interpretation of information in the tables presented in this paper. While it is more common to use effect-sizes in meta-analyses, the use of correlations is desirable and defensible in the present case as the fundamental question is one of the magnitude of relationships (for further discussion of the adjustment of effect size estimates using Fisher's transformation of r, see Hedges & Olkin, 1985).

For summaries involving breakdowns by study skills, a negative correlation would be expected when relating learning pathologies to outcome measures. For summaries involving all outcomes, however, the negative correlation would artificially decrease the mean correlation. Imagine, for example, if half the correlations in the meta-analysis consisted of relationships between learning pathologies and achievement and the average correlation of these relationships was -.4. Imagine the other half as correlations between deep processing and outcome and the average correlation of these relationships was +.4. Thus, combining these two groups would lead to a correlation approaching zero. Instead, a more correct description would reverse the sign of the learning pathologies and thus lead to a mean correlation about .4 -- indicating more correctly the strength of the relationship between study skills and outcomes. In the results section, the signs were changed to be all in the positive direction (indicating that the study skills relate positively with the learning outcome) unless otherwise stated.

Results

There were 653 effect-sizes identified from 52 studies, published between 1968 and 1993. The majority of the studies (75%) were written specifically to assess the correlations between some study skills and outcome, and the others included the correlates as part of a larger study not relating specifically to study skills. The quality of articles was coded as low (7%), medium (39%), or high (55%) using the method previously described. The majority of articles were published in journals (88%), with only 6% in books and 7% in unpublished sources (e.g. conference papers).

Most study participants were upper secondary students (23%) and university students or adults (59%). The others were primary students (3%) or junior high school students (16%). The majority of students were described as mixed ability (41%). The typical study included 700 participants, of which 47% were females and 53% were males.

The average correlation between a study skill (with the sign of the learning pathologies reversed) and an outcome was r=.14 ([z.sub.r]=.141) and the standard error of the correlation was .007. Figure 1 presents a stem and leaf diagram of all 653 correlations. To provide the most encompassing information in the stem and leaf diagram, the correlations between learning pathologies and outcomes are presented in italics. The average correlation between study skills and outcomes, with the 52 articles as the unit of analysis, was .230, and the average correlation across all effects weighted by sample size was .205. Overall, between 2 and 5% of variance was shared between study skills and learning outcomes. The effects of moderators on correlations is more important than an overall effect, however, and we present this information in Tables 1 to 5.

[Figure 1 ILLUSTRATION OMITTED]

Table 1 Correlations for the outcome measure moderated by quality of the study

Quality of study   Count   Mean   Std error

Low                  43    .25      .020
Medium              253    .18      .006
High                357    .11      .003

Table 2 Correlations for the outcome measure moderated by age of students

Age level          Count   Mean   Std error

Primary              18    .178     .053
Middle              103    .133     .021
Secondary           149    .172     .013
University/Adult    381    .132     .012

Table 3 Correlations between various types of study skills and learning outcome

Study Skill                      Count    Mean   Std error

Achieve approach                  32      .140     .015
Achieve motivation                18      .047     .045
Achieve strategy                  95      .163     .019
 Organisation                     22      .115     .038
 Searching                         9      .136     .023
 Note taking                      40      .216     .033
 Exam technique                   14      .099     .032
 Other achieve strategy           10      .166     .052
Deep approach                     38      .148     .016
Deep motive                       72      .171     .027
 Attitude/self-efficacy           37      .215     .045
 Internal locus of control        20      .147     .039
 Intrinsic motivation             15      .096     .037
Deep strategy                     97      .177     .017
 Depth & complexity               56      .158     .020
 Deep & elaborative processing    33      .222     .033
 Other deep strategy               8      .123     .055
Surface approach                  48     -.128     .017
Surface strategies                101     .023     .023
 Memorisation                     23      .066     .048
 Reproducing                      46      .022     .032
 Inflexibility                    13     -.041     .061
 Other surface strategies         19      .016     .061
Self-regulation                   29      .055     .061
Pathology/globetrotting           34     -.301     .041
Time on task                      36      .119     .036
Total study skills                53      .324     .044

Table 4 Correlations between study skill and various outcome measures

Study skill       Count   Mean   Std error

Ability             79    .094     .015
Achievement        462    .157     .011
Memory               4    .048     .162
Subject-based       48    .078     .037
Mathematics         10    .277     .060
Science              6    .103     .017
Language skills     21    .212     .039
Self                11    .138     .010
Attitude             4    .202     .085
Study skills         8    .022     .062

Table 5 Correlation between different types of study skills and various outcome measures

                  Achieve      Deep     Learning       Self
Outcome           approach   approach   pathology   regulation

Ability       n       22         23           2           8
              r     .093       .122       -.055       -.106
Achievement   n      106        140          24          17
              r     .157       .185       -.337        .031
Subject-based n       15         29           6           4
              r     .122       .170       -.227        .270
Att/SS/Self   n        2         15           2           0
              r     .141       .097       -.152           *
Totals        n      145        207          34          29
              r     .142       .170       -.302        .055

                  Study    Surface
Outcome           skills   approach   Total(a)

Ability       n       6        19         80
              r    .323     -.031       .095
Achievement   n      61       116        464
              r    .259     -.031       .154
Subject-based n      19        12         85
              r    .139      .047       .136
Att/SS/Self   n       3         2         24
              r    .370     -.091       .139
Totals        n      89       149        653
              r    .241     -.026       .144

(a) Total column is calculated with signs reversed for learning pathologies.

Table 1 presents the correlations (with the signs for learning pathologies reversed). The higher quality studies had similar correlations to the medium quality studies, and both of these were slightly lower than the lower quality studies.

Age was best categorised into four levels, which corresponded to levels of schooling. Table 2 presents this information, and there appear to be no differences in the relationship of study skills to outcomes across levels of schooling.

The study skills were coded into 12 categories (Table 3) and the correlations between the learning pathologies and outcomes were left in their original form (i.e. as negative correlations). Having a pathology is clearly related to lower outcomes. Surface approaches are negatively related to outcomes, although many surface strategies such as inflexibility and reproducing are unrelated to outcomes. Thus, most of the well-known surface strategies are not helpful in enhancing achievement.

Deep approaches (motives and strategies) are positively related to outcomes. The highest correlations with outcomes are for self-efficacy, and deep and elaborative processing (e.g. critical evaluation, conceptual organisation, comparing and contrasting information being studied, using mental imagery, reorganising material to make it personally relevant). Achieving approaches and strategies such as note taking, searching, examination techniques, and organisation were positively related to outcomes, with the highest correlation being for note taking. In general, strategies are more related to outcomes than are the motives.

Having many study skills, as assessed by total study skills scores, is positively related to outcomes. Merely increasing time on task or memorising information, however, are strategies not highly related to learning outcomes. Self-regulation methods were also unrelated to outcomes, although several of the highest correlations ([z.sub.r]=.618 and [z.sub.r]=.633) were found in the Zimmerman and Martinez-Pons (1986) study in which the measure was a total score for a range of 14 strategies of self-regulation, thereby highlighting the importance of versatility in the use of strategies for learning.

The outcomes were either achievement-related (98%) or affect-related (2%). They were further broken down into 10 categories (Table 4, note the signs for learning pathologies have been reversed in this table). The relationships of study skills with achievement, mathematics, science, language skills, and self-related outcomes were similar. There were close to zero relationships between the study skills methods and enhancing study skills as an outcome, and with memory and subject-based outcomes.

When the various outcomes are further broken down into study skills approaches (Table 5), it can be noted that versatility (labelled `Study skills' in the table) has higher relationships with ability, achievement and affective outcomes than do the other categories of study skills. Self-regulation is more related to subject-based than achievement-based outcomes (which tended to be GPA), although the small sample size adds caution to this conclusion. The surface approaches are not related to any outcome, whereas achieving and deep approaches are positively related to all outcomes. Learning pathologies were negatively correlated across most outcomes, reinforcing the view that these types of study skills are not beneficial, and worth changing. Deep and achieving approaches were positively correlated across most types of outcomes.

Conclusions

The overall correlation between study skills and various learning outcomes of .14, which translates into an effect-size of .14, is quite small given other educational effects on achievement. For instance, Hattie (1992) outlined a measurement procedure for ascertaining the typical effect of most innovations in education. He ascertained that the typical effect-size of most interventions in education is .40, and there are few innovations with average effect-sizes less than .14 (see Hattie, Biggs, & Purdie, 1996, Table 3, for a summary of typical effects of innovations in education). Thus, on average, the correlations between a student's study skills and achievement rank among the lowest of many other relationships or effects on student achievement. Even the very best comparisons of study skills on achievement rarely exceed an effect-size of .3, and thus the relationship between effective study skills and achievement can be considered minimal.

In a previous study (Hattie, Biggs, & Purdie, 1996), we argued that the effects of students undertaking study skills programs is more substantial (average effect = .45) and pointed to the importance of situated cognition, whereby training other than for simple mnemonic performance should be in context, tasks should be used within the same domain as the target content, and a high degree of learner activity and metacognitive awareness should be promoted. The present study highlights the importance of not merely assessing a student's study skills. Students who possess study skills that are very surface oriented or globetrotting and pathological are in very much need of study skills programs. Such strategies are negatively related to achievement. Those students with deeper strategies tend to have more positive relations, although given that these are correlations, it is not possible to induce causation. It is suggested, however, that it may be more effective to change the study skills and thus affect achievement, than to enhance achievement and expect consequential changes to study skills. There is no doubt that the latter situation occurs, but this is more often not in the power of the teacher.

Despite the generally low correlations, there is evidence from the present study that students who use a range of study skills (i.e. are more versatile) achieve greater success on a range of performance measures than students who focus on only one skill. Versatility in the use of study skills was more highly correlated with all types of learning outcomes than was any single study skill. Of course, it is desirable that students possess a repertoire of desirable study skills, but they also must know when to use study skill x, and when to use study skill Y. For example, despite an instructor's rhetoric exhorting students to use deep learning strategies, a final examination may require more surface learning. Knowing when to study using surface methods, and when to use deep methods is a skill -- and it can be a learned skill. Repeatedly, research points to the importance of selecting the right set of study skills to use for a particular purpose in a clearly defined context (e.g. Biggs, 1993; Brown, Collins, & Duguid, 1989). Effective strategies in one domain may be weak strategies in another (Alexander & Judy, 1988). In this meta-analysis, the importance of context in student choice of study skill is exemplified in the study by Zimmerman and Martinez-Pons (1990). Not only were students' total strategy scores strongly related to their performance in mathematics (r = .55) and English (r =.56); assessment of strategy use was based on students' self-descriptions of what they did to perform a range of academic tasks in naturalistic settings. So, although the results of this study emphasise the need to nurture within students an aptitude for flexibility/versatility -- to be skilled in many different learning behaviours -- this finding must be coupled with the findings of research in situated cognition which demonstrate that strategy use is of greater value when it is context specific.

Results from several of the studies in this meta-analysis indicate that versatility may also be an important notion in some cross-cultural settings. Correlations between achievement and total study skills scores were higher than correlations between achievement and individual study skills in three studies in which participants were specifically designated as non-Anglo (Haynes, Comer, & Hamilton-Lee, 1988; Moreno & Di Vesta, 1991; Patel, 1985).

Lack of consistency in terminology and conceptual identification in study skills research often makes it difficult to compare the results of studies in this area. Nevertheless there is a fairly well-established consensus of opinion that academic performance is likely to be better when students use strategies that involve them with more than the superficial elements of a learning activity; that is, when they use strategies that require complex cognitive engagement with the learning material. Thus, in this meta-analysis, superior academic performance was correlated with such study skills as relating new information to prior knowledge (Nolen & Haladyna, 1990), evaluating, analysing, organising, comparing and contrasting information (Gadzella, Ginther, & Williamson, 1987), wide reading and relating ideas (Biggs & Braun, 1972), analysis, elaboration, and synthesis (Bartling, 1988), and identifying main ideas, supporting details, and interrelationships (Miller, Finley, & McKinley, 1990).

Of the other individual study skills, the association between notetaking and learning outcome was noticeably higher. Although notetaking was categorised as an achieving strategy (in accordance with the Biggs, 1987, Learning Process Questionnaire), closer inspection of those studies in which notetaking was correlated with a measure of student achievement showed that higher correlations were obtained when notetaking involved the identification and manipulation of the most important ideas (e.g. Kiewra & Benton, 1982) rather than the mere recording of information read in texts or heard in lectures. This finding serves to emphasise the notion that it is not the study skill or strategy itself that is either intrinsically `good' or `bad'. Of greater importance is the manner in which a student applies the strategy for the improvement of learning.

The finding that surface strategies such as the memorisation and reproducing of information were not related to learning outcomes supports the findings of other research on this topic. However, in this regard, we should not overlook the developmental perspective on strategy use that is evident in some of the research on student learning strategies (e.g. Brown & Day, 1983). In support of such research, one of the studies in our sample (Nolen & Haladyna, 1990), for instance, reported that belief in the utility of surface-level strategies (e.g. rote memorisation of facts) was related to achievement in science for younger but not for older students. Nolen and Haladyna suggested that developmental differences in strategy use may have accounted for this difference. Another explanation may lie in the nature of the science curriculum itself -- in the beginning stages, students must acquire a set of basic tools (terminology, formulae, etc.) which they can subsequently use to understand and synthesise more conceptual scientific material. Lerner's (1971) `just-world' hypothesis (people get what they deserve and deserve what they get) does not appear to be supported by the results of this study. That is, students who work hard (i.e. study for long hours) do not appear to be rewarded (with higher grades). However such a conclusion is based on the assumption that study time is the sole indicator of effortful learning. Effort is probably better represented by the quality of time spent. The total amount of time students devote to study activities may not be nearly as important as is their allocation of study time to specific study activities and the skill with which they execute these study activities (Wilhite, 1990). Increasing study time may serve only to perpetuate the use of inappropriate strategies and test students' tolerance to the limit. As Delucchi, Rohwer, and Thomas (1987) point out, additional time spent using a generative strategy such as elaborating newly encountered information can only be expected to be positively related to course achievement to the extent that students can successfully elaborate. It is an oversimplification to consider study time in isolation from how this time is allocated to different study activities.

In sum, the results of this meta-analysis do not support the notion that encouraging students to devote more time to studying will necessarily increase their academic achievement. Rather than simply investing more time in study, students are better advised to put effort into becoming competent users of a range of study skills, particularly those that are of a deep or elaborative nature, and to make wise judgements about which study skills are best suited to their purposes.

Keywords

correlation
learning strategies
meta-analysis
outcomes of education
study skills
time on task

References

Alexander, P. A. & Judy, J. E. (1988). The interaction of domain-specific and strategic knowledge in academic performance. Review of Educational Research, 58, 375-404.

Bartling, C. A. (1988). Longitudinal changes in the study habits of successful college students. Educational and Psychological Measurement, 48, 527-535.

Biggs, J. B. (1987). Student approaches to learning and studying. Hawthorn, Vic.: Australian Council for Educational Research.

Biggs, J. B. (1993). What do inventories of students' learning processes really measure? A theoretical review and clarification. British Journal of Educational Psychology, 63, 3-19.

Biggs, J. B. & Braun, P. H. (1972). Models of evaluation and their relation to student characteristics. Journal of Educational Measurement, 9, 303-309.

Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring. Educational Researcher, 13, 4-16.

Brown, J. S., Collins, A., & Duguid, P. (1989). Situated cognition and the culture of learning. Educational Researcher, 18, 32-42.

Brown, A. L. & Day, J. (1983). Macrorules for summarizing texts: The development of expertise. Journal of Verbal Learning and Verbal Behavior, 22, 1-14.

Cantwell, R., & Beamish, P. (1994, November). Executive strategy control in secondary and tertiary populations: Contrasting understandings of self-regulation. Paper presented at the annual national conference of the Australian Association for Research in Education, University of Newcastle, New South Wales.

Carroll, J. B. (1963). A model for school learning. Teachers' College Record, 64, 723-733.

Delucchi, J. J., Rohwer, W. D., & Thomas, J. W. (1987). Study time allocation as a function of grade level and course characteristics. Contemporary Educational Psychology, 12, 365-380.

Derry, S. J. (1988). Putting learning strategies to work. Educational Leadership, 46, 4-10.

Gadzella, B. M., Ginther, D. W., & Williamson, J. D. (1987). Study skills, learning processes and academic achievement. Psychological Reports, 61,173-174.

Hattie, J. (1992). Measuring the effects of schooling. Australian Journal of Education, 36, 5-13.

Hattie, J., Biggs, J. & Purdie, N. (1996). Effects of learning skills interventions on student learning: A meta-analysis. Review of Educational Research, 66, 99-136.

Haynes, N. M., Comer, J. P., & Hamilton-Lee, M. (1988). Gender and achievement status differences on learning factors among black high school students. Journal of Educational Research, 81, 233-237.

Hedges, L. V. & Olkin, I. (1985). Statistical methods for meta-analysis. New York: Academic.

Hill, L. (1990). Effort and reward in college: A replication of some puzzling findings. Journal of Social Behaviour and Personality, 5, 151-161.

Karweit, N. (1984). Time-on-task reconsidered: Synthesis of research on time And learning. Educational Leadership, 41(8), 32-35.

Keith, T. Z., Reimers, T. M., Fehrmann, P. G., Pottebaum, S. M., & Aubey, L. W. (1986). Parental involvement, homework, and TV time: Direct and indirect effects on high school achievement. Journal of Educational Psychology, 78, 373-380.

Kiewra, K. A., & Benton, S. L. (1988). The relationship between information-processing ability and notetaking. Contemporary Educational Psychology, 13, 33-44.

Krantz, S. E. (1983). Cognitive appraisals and problem-directed coping: A prospective study of stress. Journal of Personality and Social Psychology, 44, 638-643.

Lerner, M. J. (1971). Observer's evaluation of a victim: Justice, guilt, and Veridical perception. Journal of Personality and Social Psychology, 20, 127-135.

Miller, C. D., Finley, J., & McKinley, D. L. (1990). Learning approaches and motives: Male and female differences and implications for learning assistance programs. Journal of College Student Development, 31,147-154.

Moreno, V. & Di Vesta, F. J. (1991). Cross-cultural comparison of study habits. Journal of Educational Psychology, 83, 231-239.

Nist, S. L., Simpson, M. L., & Hogrebe, M. C. (1985). The relationship between the use of study strategies and test performance. Journal of Reading Behaviour, 17, 15-28.

Nolen, S. B. & Haladyna, T. M. (1990). Motivation and studying in high school science. Journal of Research in Science Teaching, 27, 115-126.

Patel, Y. P. (1985). The impact of study habits of intellectually backward pupils upon their academic achievement. Indian Psychological Review, 28, 7-13.

Surber, J. R. (1992). The effect of test expectation, subject matter, and passage length on study tactics and retention. Reading Research and Instruction, 31, 32-40.

Walberg, H. J. (1988). Synthesis of research on time and learning. Educational Leadership, 45, 76-85.

Walberg, H. J., Fraser, B. J., & Welch, W. W. (1986). A test of a model of educational productivity among senior high school students. Journal of Educational Research, 79, 133-139.

Wilhite, S. C. (1990). Self-efficacy, locus of control, self-assessment of memory ability, and study activities as predictors of college course achievement. Journal of Educational Psychology, 82, 696-700.

Zimmerman, B. J. & Martinez-Pons, M. (1986). Development of a structured interview for assessing student use of self-regulated learning strategies. American Educational Research Journal, 23, 614-628.

Zimmerman, B. J. & Martinez-Pons, M. (1990). Student differences in self-regulated learning: Relating grade, sex, and giftedness to self-efficacy and strategy use. Journal of Educational Psychology, 82, 51-59.

Dr Nola Purdie is a Senior Lecturer in the School of Learning and Development, Faculty of Education, Queensland University of Technology, Kelvin Grove Campus, Victoria Park Road, Kelvin Grove, Queensland 4059. Professor John Hattie is the Head of the School of Education, University of Auckland, Private Bag 92019, Auckland 1, New Zealand.