Category Archives: Self Concept

Self-esteem and Academic Achievement

If sources are to be trusted, there is a mental health crisis in our schools at present. I’m not sure of how accurate these claims are or if we are indeed witnessing the ‘emotionalisation’ of education. Nick Haslam, a psychologist at the University of Melbourne does suggest that psychological concepts are being expanded so that they encompass broader areas; certain behaviours are becoming ‘pathologised’: Fear of maths becomes ‘math anxiety’ and fear of taking tests becomes ‘test anxiety’, while the NUS has banned clapping in an attempt to prevent trauma (clapping anxiety?). Of course, such anxieties are real to the individual and can be reduced through support, but whether such ‘concept creep’ is desirable is yet to be seen.

There also exists the misguided notion that raising self-esteem in young people can somehow protect them from further psychological distress and raise academic attainment. But high self-esteem can have a dark side. Roy Baumeister found that the most aggressive people tend towards high levels of self-esteem, suggesting that violence becomes more likely when other people and situations contradict a persons highly favourable view of themselves. Very high levels of self-esteem might also lead to arrogance and fragile self-concept. This results in such individuals being easily threatened and more likely to use violence in order to protect their fragile and inflated sense of self-worth. People high in self-esteem often have a mistaken impression of themselves and are more likely to claim to be more likeable and attractive, to have better relationships and to make better impressions on people than those with low self-esteem. Objective measures however, appear to cast doubt on these beliefs, leading Baumeister to conclude that ‘narcissists can be charming as first, but tend to alienate others eventually’.

There are certainly benefits to having high levels of self-esteem. Self-esteem is strongly correlated with happiness, although without a clearly established causation. What this means is that there is no way to confirm that the high levels of self-esteem cause people to be happier. It could be that happy people develop higher levels of self-esteem or that other factors indirectly lead to higher levels of happiness. It is clear, however, that low self-esteem is related to depression and stress, but the direction remains unclear. High levels of self-esteem have also been shown to foster experimentation among children, leading to an increase in risky behaviours such as smoking, drinking, drug taking and underage sex. However, high self-esteem does reduce the risk of bulimia in females.

Self-esteem doesn’t impact academic performance.

While some studies have identified a link between high self-esteem and academic attainment, correlations have been modest at best. High self-esteem, therefore, appears to have little impact on academic outcomes but findings do suggest that high levels of self-esteem emerge from good school performance. Furthermore, efforts to boost the self-esteem of school pupils have not resulted in improvements in academic performance and, in some cases, have been found to be counterproductive. Similar results have been found in adults, in that job performance is sometimes related to self-esteem but such correlations vary widely and, again, direction of causality hasn’t been reliably established. Certainly, high levels of occupational success may boost self-esteem.

It would appear, therefore, that while self-esteem is related to some positive outcomes, there is little to suggest that students with high self-esteem do any better academically than those with low self-esteem. This is not to say that schools should abandon such notions, but that interventions should be directed towards well-being rather than that formal academic outcomes. Harter suggests that scholastic competence can be seen as a specific type of self-esteem that might be better placed as a means of investigating the role of self-theories on academic outcomes. A more developed concept similar to that proposed by Harter is that of ‘academic self-concept’, of which I have written previously.


Questioning the stability of academic buoyancy.

Back in October I conducted a small-scale exploratory study into three constructs (academic self-concept, academic buoyancy and implicit theories of intelligence). You can read the details here. A few weeks ago I asked the same students toStressed-Student complete the questionnaires again to confirm that these constructs remain stable over time. I was particularly interested in academic buoyancy (day-to-day resilience) due to the forthcoming AS exams. What I wanted to confirm was that those students who considered themselves resilient at time 1 (October 2013) still considered themselves resilient at time 2 (May 2014). This would be measured using the Academic Buoyancy Scale (Martin and Marsh, 2007), a four-item measure of academic buoyancy (AB) that has proved reliable over time and within different settings.

Let’s get some of the problems with the ‘study’ out of the way now.

At time 1, the sample consisted of 41 year 12 students. At time 2, and due to a number of factors (including subject/school drop-out and a lower volunteer rate) this had dropped to 27. The final sample is therefore very low and is far from representative.

The sample is small and unrepresentative – predominantly white, middle-class and with a higher percentage of female participants.

However, as this was an exploratory study, I was looking for general patterns needed to establish possible further avenues of investigation.

Ethical Issues.

The study was conducted in line with ethical procedures of the University of York. Participants were volunteers and gave informed written consent (all participants were over the age of 16). They had the right to withdraw from the study as any time (including the withdrawal of their data).

What did the data show?

Data analysis was conducted using the R statistical package. The results of the t-test found a significant difference between AB at time 1 and AB at time 2 (p<0.01). Further analysis found an effect size of 0.675. If we apply Cohen’s (1988) conventions for effect size, we also find a highly significant difference between time 1 and time 2 (so we can be pretty confident that timing was a major factor).

What does all this mean?

Results would suggest that AB isn’t stable and is mitigated by other factors. The timing of the second data collection activity (a week before the start of AS exams) could play a role in the difference between the two sets of scores, begging the question “Do students feel less confident about their abilities at different times?” Outcome measures (in the form of AS results) can be examined in August and could (but only ‘could’) yield more information.

Where now?

The plan now is to use experience-sampling methods (ESM) to collect data on a number of factors ‘as they happen’. The problem with much of the research into academic buoyancy is that participants are asked to complete measures in isolation (i.e. “I am good at dealing with setbacks”). ESM allows for participants to think about these measures in a more realistic and moment-by-moment way via electronic ‘prompts’ sent to mobile devices. ESM tends to result in large data sets, dependent upon the number of prompts and length of the study, so sample sizes can be smaller (and, for practical reasons, need to be). An additional possibility would be to supplement the ESM data with a end of day/end of week questionnaire to investigate the difference between immediate and retrospective self-assessments.

What’s the point?

Emotion appears to impact on learning. Research has suggested that factors such as self-concept, boredom, anxiety and resilience can have both positive and negative effects on academic outcomes, as well as cognitive functions like attention. Understanding the nature of these factors could help to develop interventions to stabilise some of them. Emotion impacts on cognition, for example, stress can heighten recall to a point but too much anxiety leads to inaccurate recall. The so-called Yerkes-Dodson suggests that performance increases as physiological and mental arousal increases to an optimum level, at which point cognitive functions begin to decline. Although the Yerkes-Dodson law in somewhat dated, more recent research appears to support its validity.

In a system where more and more of our young people are suffering from heightened levels of anxiety (the reason for which is highly debatable), examining their daily classroom lives can be provide rich data into how, when and why they do and do not learn.

What do we really mean by resilience?

Character, resilience, buoyancy, grit – these concepts have been floating about a lot lately.

In February the All-party Parliamentary Group on Social Mobility published the document ‘Character and Resilience Manifesto’ while in the same month Tristram Hunt told conference delegates at the AQA Creative Education conference that character and resilience can and should be taught in schools (a point he returned to recently in an address to the Institute for Effective Education in York). In March Liz Truss suggested that mindfulness lessons should be introduced to improve resilience in schools, adding weight to a growing agenda on the importance of non-cognitive skills.

If I’ve learnt anything over the past few months it’s been that educationalists are concerned that our obsession with exam success is neglecting the importance of these non-cognitive skills. The politicians, I cynically assume, are just along for the ride. I’ve blogged about resilience before, only I don’t tend to call it resilience because resilience as it stands in the research literature isn’t always the same resilience as the one I’m interested in. More recently, the term ‘grit’ has also entered the non-cognitive skill lexicon, unfortunately it’s often seen as synonymous with resilience – which it isn’t. Ultimately I often end up discussing the wrong thing with people because what I mean by resilience and (occasionally) ‘grit’ just isn’t the same thing. And herein lies the problem.

Resilience, according to the literature, is the ability to overcome major adversity. Much of the research has focussed on the way in which certain at-risk groups cope with major negative outcomes. This isn’t necessarily what others think it is – what they mean is day-to-day resilience, what psychologists Herb Marsh and Andrew Martin call ‘buoyancy’. Academic buoyancy refers to individuals’ ability to cope with all those everyday challenges (like a bad mark on a test or an impending exam) which a more characteristic of the lives of students than those covered by resilience. Buoyancy assumes minor negative outcomes, the accumulation of which can lead to major academic underachievement. Like resilience, buoyancy is a dynamic process – it isn’t stable and is state dependent (in the same way academic self-concept is state dependent). What we mean by this is that a student might be confident in, say, maths where they display a positive self-concept and high buoyancy but less confident in English where their self-concept is negative and their buoyancy low. Self-esteem represents a global measure and doesn’t necessarily relate to either academic or non-academic self-concept – increasing a child’s self-esteem won’t make them more confident in English if their English academic self-concept remains negative.

So what about ‘grit’? Duckworth defines ‘grit’ as “persuasiveness and passion for long-term goals” which is distinctly different from buoyancy and resilience (although all three are related). Grit does appear to predict resilience, which in turn predicts buoyancy but grit remains a rather vague term with a research base much lower than either resilience or buoyancy. Hunt used the terms ‘grit’ and ‘resilience’ interchangeably at the IEE, assuming that the former was simply an American manifestation of the latter. Angela Duckworth (the American psychologist who coined the term), however, suggests that ‘grit’ is a trait (it remains relatively stable over time) – resilience and buoyancy (as already mentioned) are dynamic processes – they aren’t stable.

So, with what appears to be a number of misunderstandings and misconceptions, how does all this play out on the ground? Several schools have certainly embraced the ‘idea’ of resilience but have interpreted it in different ways. The Knowledge is Power Programme (KIPP), in the United States, is certainly one of the more successful ones. KIPP has identified 24 characteristics that schools should try and develop, while Bedford Academy in the UK has narrowed these down to ‘the magnificent seven’ (grit, zest, optimism, social intelligence, gratitude, curiosity and self-control) while Rossett School in North Yorkshire, UK concentrates on the ‘3R’s’ (Resilience, Reflectiveness and Respect). Marsh and Martin identify the 5C’s (or motivational predictors) of academic buoyancy (Confidence, Coordination, Commitment, Composure and Control), suggesting another set of characteristics that could form part of a resilience or ‘character’ intervention.

Measuring day-to-day resilience presents yet another problem. KIPP, Bedford Academy and Rossett all issue pupils with a score (dependent on the system used) as part of the reporting system. In the case of Rossett, this appears to be subject specific – correctly identifying resilience as state (not trait) based. However, research into resiliency and buoyancy tends to use self-completion psychometric tools and it would be interesting to discover if these pupils would rate their resiliency as similar to the teacher rating (and which one we should choose as ‘accurate’). Unfortunately, any systematic investigation into the effectiveness of these systems is proving hard to track down and outcomes measures don’t seem to exist.

I’m certainly in favour of programmes that help to improve resilience and improve character. The problems lie both in a common terminology and accurate measurement, in terms of progress and outcomes. If we can’t trust the tools and we can’t measure the outcome it all turns into a bit of a farce.

I want to tell you a story…

jackanoryI’m rapidly realising that research with young people is more fraught with problems than I originally surmised. Ironically, the very area of my research should have highlighted this possibility, in that young people will go to often extraordinary lengths to safe-guard their self esteem and provide an answer that they believe the researcher will judge at ‘right’. Unfortunately, as researchers, we are rarely interested in the right answer – usually because there isn’t one.

As a result, it can become difficult for us to confirm that any answer provided by young people is honestly given as opposed to an attempt to ‘get it right’. Furthermore, understanding the reasoning behind their answer could actually provide a more accurate indication of factors such as self-esteem and resilience than the very measure itself. This calls into question the reliability of measuring, say, academic resilience using psychometric methods such as scales.

What if we could make these questions more objective and still use the answers as a means to providing subjective measures?

One area I have decided to explore is the use of scenarios (or vignettes). The basic premise is fairly simple: Rather than presenting participants with questionnaires that prompt a subjective response  using a likert-type scale like this:




…the questions are presented once the participant has read a fictional vignette.

For example:

Frank is studying A-level maths and his teacher has just given him the results of his latest test. Frank’s target grade is B but the grade he has achieved for the test is a D.

On a scale of 1 to 10, decide how this bad result would have affected Frank’s confidence in maths (where 1 is ‘a great deal’ and 10 is ‘not much at all’).

Vignette methodology is widely used in the social sciences, including sociology and education. Vignettes provide a way for the participant to distance himself or herself from the question by ensuring that on the surface the questions aren’t necessarily about them. The assumption from the researchers point of view, however, is that the participant will provide an answer that is more in keeping with their views but limits the desire to safe-guard self-esteem.

Conducting educational research – are teens more problematic?

I recently conducted a small-scale exploratory study with some year 12 students. I wasn’t totally sure of what I really wanted to accomplish with the study or what to expect but I hoped that I could at least try to establish relationships between a number of different measures.

There were two main points I was particularly interested in:

1. If students could choose their own target grade, how close would this be to their actual target grade?

2. How do the constructs of Academic Self Concept (ASC), Academic Buoyancy (AB) and Implicit theories of Intelligence (Mindset) relate both to each other and the accuracy of target grade predictions?

The three constructs were measured using established and validated likert-scale questionnaires and were administered to 41 year 12 students following the completion of consent procedures.

There was an anticipation that the three constructs would correlate significantly and that those who scored highly on all three constructs would be more accurate with their target prediction. However, while there was a small positive correlation between ASC and Mindset (i.e. those participants who scored highly on the measure of ASC were more likely to have ‘growth mindsets’), none of the other measures appeared to correlate. Furthermore, the majority of target predictions formed a cluster around the “C” grade.

As I’ve already mentioned, this was very much an exploratory study using a small unrepresentative sample. Nevertheless, the results did get me thinking about a number of problems related to carrying out studies with teenagers and their self-concepts, and while this might not be directly related to the task in hand (and certainly won’t explain the results) I began to look back on a previous, half-abandoned project on teenage self concept.

As most parents of teenagers will confirm, adolescence brings with it a period of heightened self-consciousness unseen in younger children. Teenagers become overly, at times almost pathologically, preoccupied with the way in which others interpret their actions and behaviour. In developmental psychology terms this is known as egocentrism and is considered a normal healthy stage of development.

The American developmental psychologist and writer David Elkind expanded on this idea in the 1960’s in a manner that allows us to at least attempt an explanation for the seemingly egocentric nature of teenagers. Once children are able to distinguish between their own thoughts and the thoughts of others (often referred to as mentalizing) they become preoccupied with the fear that the thoughts of other people are focussed on their own behaviour and appearance – what Eldkind calls the imaginary audience. The imaginary audience is seen by the teenager as scrutinising their every move, being critical of their behaviour – the way they dress or their haircut and so on. This fear leads to increase in self-consciousness in terms of the ways in which others (especially their peers) see them. Of course the view held by the teenager is more likely to be imagined than having any existence within reality, nevertheless, it is real for them and no amount of persuasion will convince them otherwise.  A related concept, according to Elkind, is that of the personal fable. Personal Fable describes the teenagers obsession with their own uniqueness and the belief that they are in some way special, leading to the view that ‘nobody understands me’. While Elkinds’ theory has been expanded and developed since the 1960’s we can already see how the way in which teenagers behave is closely linked to the way in which they are developing socially and emotionally.

More recently, researchers have suggested that the these twin concepts of imaginary audience and personal fable represent the teenagers search for identity in the midst of a rapidly changing set of social circumstances – greater independence, the move from primary to secondary education and the increase in pressure from parents and peers. When the young teenager begins to development close relationships outside the family she finds herself torn between her new social life and the desire to remain connected to the family, leading perhaps to anxiety and fixation on one to the detriment of the other. While we can attempt to support the teenager and suggest that she divide her time equally between the two, there is a real possibility that her egocentrism and centration prevents her from dealing with, what is to her, a major all-consuming dilemma. In order to resolve this dilemma our teen is likely to engage in daydreaming as way of placing herself within a number of different social contexts and interactions involving various groups of others, simultaneously she may also employ the concept of personal fable (the emphasis on her uniqueness) as means to separate herself from her family. Indeed this complex and often agonising search for identity may lead to periods where the teenager separates himself from both his family and his own social groups, and while such behaviour may seem unusual, it is rarely detrimental.

So what impact (if any) does this have on the researcher who has chosen teenagers at their focus? My concern is that the objectivity of any self-completed questionnaire is compromised even if basic criteria regarding anonymity and confidentiality are adhered to. There might be a danger that teenagers will attempt to safeguard their self-esteem even with the guarantee that no other person will have access to their personal results. High achieving teenagers might be reluctant, therefore, to accurately reflect their ability for fear of being seen as arrogant while others might not want to underestimate their ability for fear of being seen as stupid. As many researchers in the areas of ASC and related concepts rely heavily on the use of such measures, do we need to question the validity of any results.

8 Habits of Highly Successful Learners

The study of student motivation and engagement has uncovered a great deal about the particular habits and traits of those learners who are more likely to succeed. Listed here are the 8 of these habits which appear to be the best predictors.

1. Effort: If you don’t put the work in, you’re not going to get much out. Those learners who literally put in the hours beyond the classroom, reap higher rewards.


2. Engagement: Staying focused and on-task. Not getting distracted by things that are going on around you of by disruptive peers.

engage3. Skill development: Just like playing the guitar or becoming an expert batsman, you need to develop your skills as a learner. This might include note-taking, essay writing or honing your listening skills.

4. Participation: Being involved in what’s going on in class. Ask and answer questions and help others who might be struggling. Being fully involved in what’s going on is an important part of the learning process.

5. Attendance: Turning up for lessons – there is a causal relationship between attendance and exam success.

6. Self-concept: How you see yourself as a learner. A positive self-concept is positively correlated with achievement. However, self-concept needs to be realistic and goals achievable – over-estimation can be an equally damaging as under-estimation.

7. Persistence: Not giving up, even when it gets difficult! We quite often use the terms resilience and buoyancy  for this ability to bounce back after failure.

8. Enjoyment of learning: Not seeing learning as something you a forced to do, but rather something that can be enjoyable and personally fulfilling. Although some will use the term ‘Enjoyment of School’ I would argue that school isn’t necessarily a prerequisite for success (after all, many successful learners don’t go to school). I’m reminded of a year 12 pupil who (after completing her AS exams)  once commented that she couldn’t wait to start the A2 work, ‘I know it sounds sad,” she said, ‘I just love learning’.


5 Reasons Why Students Fail

There are many reasons for academic failure, however the underlying cause appears to be related to self-esteem or ego-protection. Here I’ve identified five possible reasons why our students might underachieve.

1. They would rather be thought of as lazy than stupid

Safeguarding self-esteem is often the most important priority for many adolescents. Students hate to be viewed as less intelligent than their classmates and will often go to extreme lengths to ensure that they are not seen as stupid. One way of ensuring that this doesn’t happen is to be seen as lazy by not studying, not preparing for exams or excessive procrastination. Then, if they fail, they can simply say it’s because they didn’t work hard enough, rather than believing that others think they’re less intelligent.

2. They see intelligence as fixed

Harvard psychologist Carol Dweck suggests that people generally fall into two groups when it comes to thinking about intelligence (so called implicit theories of intelligence). The first group see intelligence as a fixed entity, that is, they view intelligence as wholly innate, hard-wired and impervious to change. Dweck calls this a ‘fixed mindset’ because those who hold such a view are prevented from succeeding by their very narrow and restricted view of their capabilities. When fixed ‘mindsetters’ experience failure or get stuck on a particular problem, they blame their lack of intelligence for their inability to progress and simply give up. The second group view intelligence as flexible, malleable and incremental. Success to these so-called ‘growth mindsetters’ is about hard work, learning from failure and not being restricted by a view that intelligence is innate and unchangeable.

3. They set their goals too high

It might seem odd that students could set their goals too high but for goals to be achievable they need to be realistic. Students might be tempted to set their goals too high (such as aiming for an A when all indicators suggest that they are a solid C student) as a way of defending themselves. When they don’t achieve the standard they set for themselves they can explain it in terms of setting their standards too high – so-called ‘defensive optimism’. While optimism of this kind is counter-productive, ‘realistic optimism’ can be beneficial. By examining the evidence (such as past marks on essays, tests and so on) the student can approach goal setting from an informed perspective. So, rather than a C student aiming for an A, they might decide that a B is more realistic and in-line with past achievement.

4. They set their goals too low

While some students aim too high, others are likely to aim too low. Again, this is a form of ego-defence (or a way of protecting their self-esteem). Defensive pessimism leads the student to set their goals lower than the evidence suggests, so a grade B student might say, “I’ll be lucky if I get a C”, or even “I’ll be lucky if I pass.” By doing this the student is almost guaranteeing a positive outcome – if they under-perform they can claim that they were right all along; if they get their predicted grade B (or exceed this) they can revel in the success.

5. They don’t follow advice

It’s not that students always think that their way of doing something is better than yours – it’s just that many students are always safeguarding themselves from possible failure. By not following advice they manage to ensure that it was the strategies they employed rather than anything to do with their intelligence that led to failure (see also my previous post on self-handicapping). For example, they might insist on listening to music while preparing for an exam even though they know that all the evidence suggests that most music is a distraction. A more obvious attempt at safeguarding self-esteem is to refrain from self-testing. Even though research has found that self-testing is one the most effective forms of preparation, many students avoid it because they don’t like to experience the feelings that arise when they get things wrong – self-testing, therefore, acts as a direct assault on self-esteem.