Category: calibration

Keith S. Taber

---

---

I was intrigued by the title of an article I saw in a notification: "Psychological Skills in Relation to Academic Achievement through Swimming Context". In part, it was the 'swimming context' – despite never having been very athletic or sporty (which is not to say I did not enjoy sports, just that I was never particularly good at any), I have always been a regular and enthusiastic swimmer.  Not a good swimmer, mind (too splashy, too easily veering off-line) – but an enthusiastic one. But I was also intrigued by the triad of psychological skills, academic achievement, and swimming.

Perhaps I had visions of students' psychological skills being tested in relation to their academic achievement as they pounded up and down the pool. So, I was tempted to follow this up.

Investigating psychological skills and academic achievement

The abstract of the paper by Bayyat and colleagues reported three aims for their study:

The article was published in a journal called 'Psychology and Education', which, its publishers*  suggest is "a quality journal devoted to basic research, theory, and techniques and arts of practice in the general field of psychology and education".

A peer reviewed journal

The peer review policy reports this is a double-blind peer-reviewed journal. This means other academics have critiqued and evaluated a submission prior to its being accepted for publication. Peer review is a necessary (but not sufficient) condition for high quality research journals.

Journals with high standards use expert peer reviewers, and the editors use their reports to both reject low-quality submissions, and to seek to improve high-quality submissions by providing feedback to authors about points that are not clear, any missing information, incomplete chains of argumentation, and so forth. In the best journals editors only accept submissions after reviewers' criticisms have been addressed to the satisfaction of reviewers (or authors have made persuasive arguments for why some criticism does not need addressing).

(Read about peer review)

The authors here report that

"The statistical analysis results revealed an average level of psychological skills, significant differences in psychological skills level in favor of female students, A students and JU[**], and significant positive relation between psychological skills and academic achievement."

Bayyat et al., 2021: 4535

Rewriting slightly it seems that, according to this study:

• the students in the study had average levels of psychological skills;
• the female students have higher levels of psychological skills than their male peers;
• and that there was some kind of positive correlation between psychological
• skills and academic achievement;

Anyone reading a research paper critically asks themselves questions such as

• 'what do they mean by that?';
• 'how did they measure that?;
• 'how did they reach that conclusion?'; and
• 'who does this apply to?'

Females are better – but can we generalise?

In this study it was reported that

"By comparing psychological skills between male and female participants, results revealed significant differences in favor [sic] of female participants"

"All psychological skills' dimensions of female participants were significant in favor [sic] of females compared to their male peers. They were more focused, concentrated, confident, motivated to achieve their goals, and sought to manage their stress."

Bayyat et al., 2021: 4541, 4545

---

---

A pedant (such as the present author) might wonder if "psychological skills' dimensions of female participants" [cf. psychological skills' dimensions of male participants?] would not be inherently likely to be in favour of females , but it is clear from the paper that this is intended to refer to the finding that females (as a group) got significantly higher ratings than males (as a group) on the measures of 'psychological skills'.

If we for the moment (but please read on below…) accept these findings as valid, an obvious question is the extent to which these results might generalise beyond the study. That is, to what extent would these findings being true for the participants of this study imply the same thing would be found more widely (e.g., among all students in Jordanian Universities? among all university students? Among all adult Jordanians? among all humans?)

Two key concepts here are the population and the sample. The population is the group that we wish our study to be about (e.g., chemistry teachers in English schools, 11-year olds in New South Wales…), and the sample is the group who actually provide data. In order to generalise to the population from the sample it is important that the sample is large enough and representative of the population (which of course may be quite difficult to ascertain).

(Read about sampling in research)

(Read about generalisation)

In this study the reader is told that "The population of this study was undergraduate male and female students attending both intermediate and advanced swimming courses" (Bayyat et al., 2021: 4536). Taken at face value this might raise the question of why a sample was drawn exclusively from Jordan – unless of course this is the only national context where students attend intermediate or advanced swimming courses. *** However, it was immediately clarified that "They consisted of (n= 314) students enrolled at the schools of Sport Sciences at three state universities". That is, the population was actually undergraduate male and female students from schools of Sport Sciences at three Jordanian state universities attending both intermediate and advanced swimming courses.

"The Participants were an opportunity sample of 260 students" (Bayyat et al., 2021: 4536). So in terms of sample size, 260, the sample made up most of the population – almost 83%. This is in contrast to many educational studies where the samples may necessarily only reflect a small proportion of the population. In general, representatives of a sample is more important than size as skew in the sample undermines statistical generalisations (whereas size, for a representative sample, influences the magnitude of the likely error ****) – but a reader is likely to feel that when over four-fifths of the population were sampled it is less critical that a convenience sample was used.

This still does not ensure us that the results can be generalised to the population (students from schools of Sport Sciences at three Jordanian state universities attending 'both' intermediate and advanced swimming courses), but psychologically it seems quite convincing.

Ontology: What are we dealing with?

The study is only useful if it is about something that readers think is important – and it is clear what it is about. The authors tells us their study is about

• Psychological Skills
• Academic Achievement

which would seem to be things educators should be interested in. We do need to know however how the authors understand these constructs: what do they mean by 'a Psychological Skill' and 'Academic achievement'? Most people would probably think they have a pretty good idea what these terms might mean, but that is no assurance at all that different people would agree on this.

So, in reading this paper it is important to know what the authors themselves mean by these terms – so a reader can check they understand these terms in a sufficiently similar way.

What is academic achievement?

The authors suggest that

"academic achievement reflects the learner's accomplishment of specific goals by the end of an educational experience in a determined amount of time"

Bayyat et al., 2021: 4535

Bayyat et al., 2021: 4535

This seems to be  the extent of the general characterisation of this construct in the paper *****.

What are psychological skills?

The authors tell readers that

"Psychological skills (PS) are a group of skills and abilities that enhances peoples' performance and achievement…[It has been] suggested that PS includes a whole set of trainable skills including emotional control and self-confidence"

Bayyat et al., 2021: 4535

Bayyat et al., 2021: 4535

For the purposes of this particular study, they

"identified the psychological skills related to the swimming context such as; leadership, emotional stability, sport achievement motivation, self-confidence, stress management, and attention"

Bayyat et al., 2021: 4536

Bayyat et al., 2021: 4536

So the relevant skills are considered to be:

• leadership
• emotional stability
• sport achievement motivation
• self-confidence
• stress management
• attention

I suspect that there would not be complete consensus among psychologists or people working in education over whether all of these constructs actually are 'skills'. Someone who did not consider these (or some of these) characteristics as skills would need to read the authors' claims arising from the study about 'psychological skills' accordingly (i.e., perhaps as being about something other than skills) but as the authors have been clear about their use of the term, this should not confuse or mislead readers.

Epistemology: How do we know?

Having established what is meant by 'psychological skills' and 'academic achievement' a reader would want to know how these were measured in the present study – do the authors use techniques that allow them to obtain valid and reliable measures of 'psychological skills' and 'academic achievement'?

How is academic achievement measured?

The authors inform readers that

"To calculate students' academic achievement, the instructors of the swimming courses conducted a valid and reliable assessment as a pre-midterm, midterm, and final exam throughout the semester…The assessment included performance tests and theoretical tests (paper and pencil tests) for each level"

Bayyat et al., 2021: 4538

Bayyat et al., 2021: 4538

Although the authors claim their assessment are valid and reliable, a careful reader will note that the methodology here does not match the definition of "accomplishment of specific goals by the end of an educational experience" (emphasis added)- as only the final examinations took place at the end of the programme. On that point, then, there is a lack of internal consistency in the study. This might not matter to a reader who did not think academic achievement needed to be measured at the end of a course of study.

Information on the "Academic achievement assessment tool", comprising six examinations (pre-midterm, midterm, and final examinations at each of the intermediate and advanced levels) is included as an appendix – good practice that allows a reader to interrogate the instrument.

Although this appendix is somewhat vague on precise details, it offers a surprise to someone (i.e., me) with a traditional notion of what is meant by 'academic achievement' – as both theory and practical aspects are included. Indeed, most  of the marks seem to be given for practical swimming proficiency. So, the 'Intermediate swimming Pre-midterm exam' has a maximum of 20 marks available – with breast stroke leg technique and arm technique each scored out of ten marks.

The 'Advanced swimming midterm exam' is marked out of 30, with 10 marks each available for the 200m crawl (female), individual medley (female) and life guarding techniques. This seems to suggest that 20 of the 30 marks available can only be obtained by being female, but this point does not seem to be clarified. Presumably (?) male students had a different task that the authors considered equivalent.

How are psychological skills measured?

In order to measure psychological skills the authors proceeded to "to develop and validate a questionnaire" (p.4536). Designing a new instrument is a complex and challenging affair. The authors report how they

"generated a 40 items-questionnaire reflecting the psychological skills previously mentioned [leadership, emotional stability, sport achievement motivation, self-confidence, stress management, and attention] by applying both deductive and inductive methods. The items were clear, understandable, reflect the real-life experience of the study population, and not too long in structure."

Bayyat et al., 2021: 4538

So, items were written which it was thought would reflect the focal skills of interest. (Unfortunately there are no details of what the authors mean by "applying both deductive and inductive methods" to generate the items.) Validity was assured by asking a panel of people considered to have expertise to critique the items:

"the scale was reviewed and assessed by eight qualified expert judges from different related fields (sport psychology, swimming, teaching methodology, scientific research methodology, and kinesiology). They were asked to give their opinion of content representation of the suggested PS [psychological skills], their relatedness, clarity, and structure of items. According to the judges' reviews, we omitted both leadership and emotional stability domains, in addition to several items throughout the questionnaire. Other items were rephrased, and some items were added. Again, the scale was reviewed by four judges, who agreed on 80% of the items."

So, construct validity was a kind of face validity, in that people considered to be experts thought the final set of items would elicit the constructs intended, but there was no attempt to see if responses correlated in any way with any actual measurements of the 'skills'.

Readers of the paper wondering if they should be convinced by the study would need to judge if the expert panel had the right specialisms to evaluate scale items for 'psychological skills',and might find some of the areas of expertise (i.e.,

• sport psychology
• swimming
• teaching methodology
• scientific research methodology
• kinesiology)

more relevant than others:

Self-reports

If respondents responded honestly, their responses would have reflected their own estimates of their 'skills' – at least to the extent that their interpretation of the items matched that of the experts. (That is, there was no attempt to investigate how members of the population of interest would understand what was meant by the items.)

Here are some examples of the items in the instrument:

Construct ('psychological skill')	Example item
self-confidence	I manage my time effectively while in class
sports motivation achievement	I do my best to control everything related to swimming lessons.
attention	I can pay attention and focus on different places in the pool while carrying out swimming tasks
stress-management	I am not afraid to perform any difficult swimming skill, no matter what

Analysis of data

The authors report various analyses of their data, that lead to the conclusions they reach. If a critical reader was convinced about matters so far, they would still need to beleive that the analyses undertaken were

• appropriate, and
• completed competently, and
• correctly interpreted.

Drawing conclusions

However, as a reader I personally would have too many quibbles with the conceptualisation and design of instrumentation to consider the analysis in much detail.

To my mind, at least, the measure of 'academic achievement' seems to be largely an evaluation of swimming skills. They are obviously important in a swimming course, but I do not consider this a valid measure of academic achievement. That is not a question of suggesting academic achievement is better or more important than practical or athletic achievements, but it is surely something different (akin to me claiming to have excellent sporting achievement on the basis of holding a PhD in education).

The measure of psychological skills does not convince me either. I am not sure some of the focal constructs can really be called 'skills' (self-confidence? motivation?), but even if they were, there is no attempt to directly measure skill. At best, the questionnaire offers self-reports of how students perceive (or perhaps wish to be seen as perceiving) their characteristics.

It is quite common in research to see the term 'questionnaire' used for an instrument that is intended to test knowledge or skill – but questionnaires are not the right kind of instrument for that job.

(Read about questionnaires)

Significant positive relation between psychological skills and academic achievement?

So, I do not think this methodology would allow anyone to find a "significant positive relation between psychological skills and academic achievement" – only a relationship between students self-ratings on some psychological characteristics and swimming achievement. (That may reflect an interesting research question, and could perhaps be a suitable basis for a study, but is not what this study claims to be about.)

Significant differences in psychological skills level in favor of female students?

In a similar way, although it is interesting that females tended to score higher on the questionnaire scales, this shows they had higher self-ratings on average, and tells us nothing about their actual skills.

It may be that the students have great insight into these constructs and their own characteristics and so make very accurate ratings on these scales – but with absolutely no evidential basis for thinking this there are no grounds for making such a claim.

An alternative interpretation of the results is that on average the male students under-rate their 'skills' compared to their female peers. That is the 'skills' could be much the same across gender, but there might be a gender-based difference in perception. (I am not suggesting that is the case, but the evidence presented in the paper can be explained just as well by that possibility.)

An average level of psychological skills?

Finally, we might ask what is meant by

"The statistical analysis results revealed an average level of psychological skills…"

"Results of this study revealed that the participants acquired all four psychological skills at a moderate level."

Bayyat et al., 2021: 4535, 4545

Even leaving aside that what is being measured is something other than psychological skills, it is hard to see how these statements can be justified. This was the first administration of a new instrument being applied to a sample of a very specific population.

---

---

The paper reports standard deviations for the ratings on the items in the questionnaire, so – as would be expected – there were distributions of results: spreads with different students giving different ratings. Within the sample tested, some of the students will have given higher than median ratings on an item, some will have given lower than median ratings – although on average the ratings for that particular item would have been – average for this sample (that is, by definition!) So, assuming this claim (of average/moderate levels of psychological skills) was not meant as a tautology, the authors might seem to be suggesting that the ratings given on this administration of the instrument align with what would be typically obtained, that is from across other administrations.

That is, the authors seem to be suggesting that the ratings given on this administration of the instrument align with what they expect would be typically obtained from across other administrations. Of course they have absolutely no way of knowing that is the case without collecting data from samples of other populations.

What the authors actually seem to be basing these claims (of average/moderate levels of psychological skills) on is that the average responses on these scales did not give a very high or very low rating in terms of the raw scale. Yet, with absolutely no reference data for how other groups of people might respond on the same instrument that offers little useful information. At best, it suggests something welcome about the instrument itself (ideally one would wish items to elicit a spread of responses, rather than having most responses rated very high or very low) but nothing about the students sampled.

On this point the authors seem to be treating the scale as calibrated in terms of some nominal standard (e.g. 'a rating of 3-4 would be the norm'), when there is no inherent interpretation of particular ratings of items in such a scale that can just be assumed – rather this would be a matter that would need to be explored empirically.

The research paper as an argument

The research paper is a very specific genre of writing. It is an argument for new knowledge claims. The conclusions of the paper rest on a chain of argument that starts with the conceptualisation of the study and moves through research design, data collection, analysis, and interpretation. As a reader, any link in the argument chain that is not convincing potentially invalidates the knowledge claim(s) being made. Thus the standards expected for research papers are very high.

---

---

In sum then, this was an intriguing study, but did not convince me (even if it apparently convinced the peer reviewers and editor of Psychology and Education). I am not sure it was really about psychological skills or, academic achievement

…but at least it was clearly set in the context of swimming.

---

Work cited:

Bayyat, M. M., Orabi, S. M., Al-Tarawneh, A. D., Alleimon, S. M., & Abaza, S. N. (2021). Psychological Skills in Relation to Academic Achievement through Swimming Context. Psychology and Education, 58(5), 4535-4551.

Taber, K. S. (2019). Experimental research into teaching innovations: responding to methodological and ethical challenges. Studies in Science Education, 55(1), 69-119. doi:10.1080/03057267.2019.1658058 [Download manuscript version]

---

* Despite searching the different sections of the journal site, I was unable to find who publishes the journal. However, searching outside the site I found a record of the publisher of this journal being 'Auricle Technologies, Pvt., Ltd'.

** It transpired later in the paper that 'JU' referred to students at the University of Jordan: one of three universities involved in the study.

*** I think literally this means those who participated in the study were students attending both an intermediate swimming course and an advanced swimming course – but I read this to mean those who participated in the study were students attending either an intermediate or advanced swimming course. This latter interpretation is consistent with information given elsewhere in the paper: "All schools of sports sciences at the universities of Jordan offer mandatory, reliable, and valid swimming programs. Students enroll in one of three swimming courses consequently: the basic, intermediate, and advanced levels". (Bayyat et al., 2021: 4535, emphasis added)

**** That is, if the sample is unrepresentative of the population, there is no way to know how biased the sample might be. However, if there is a representative sample, then although there will still likely be some error (the results for the sample will not be precisely what the results across the whole population would be) it is possible to calculate the likely size of this error (e.g., say ±3%) which will be smaller when a higher proportion of the population are sampled.

***** It is possible some text that was intended to be at this point has gone missing during production – as, oddly, the following sentence is

facilisi, veritus invidunt ei mea (Times New Roman, 10)

Bayyat et al., 2021: 4535

which seems to be an accidental retention of text from the journal's paper template.