Weird science


So, you are in Year 9 and it’s time to pick your options. (We’ll leave aside whether funding constraints leave you with many options). Which science route are you going to take? Triple science or combined science? Will your GCSE grades be roughly the same whichever route you opt for?

In this blogpost we look at the outcomes of pupils who pursued different routes in science at Key Stage 4. Bear in mind that we’re looking back at what happened in legacy GCSEs in 2016. Things might turn out differently with the reformed GCSEs that current Year 10 pupils will take next year. And then again they might not.

How many pupils take triple sciences?

In total, 89% of pupils who reached the end of Key Stage 4 at state-funded mainstream schools in 2016 were entered for two or more science GCSEs. This was up from 63% in 2011. Most of the increase was due to schools switching back to GCSEs from equivalent qualifications (e.g. BTECs) in response to changes to the accountability regime.

A quarter of pupils were entered for triple science in 2016[1]. As you would expect, those with higher levels of prior attainment were more likely to be entered for it.

Percentage of pupils entered for two or more science GCSEs by route, 2016

Contextual factors play a part in entry rates for triple science. Disadvantaged pupils, particularly those with higher levels of prior attainment, were less likely to be entered than other pupils with similar prior attainment. Some (but not all) of this gap is explained by disadvantaged pupils being more likely to attend schools where triple science is not offered. (The Wellcome Trust Science Education Tracker [PDF] investigates differential access to triple science among different groups of pupils.)

Percentage of pupils entered for triple science by disadvantage, 2016

What is also clear is that pupils entered for triple science tend to achieve higher grades in science than pupils with similar prior attainment who pursue other science routes at Key Stage 4. (For the most part this means core and additional science.)

As the chart below shows, pupils entered for triple science tend to achieve around three-quarters of a grade higher in each of their best two science GCSEs – the number of science qualifications counted when English Baccalaureate (EBacc) achievements are being considered – compared to pupils with similar prior attainment pursuing other science routes.

The pattern still holds if we make a rudimentary adjustment for context by not including Pupil Premium pupils and those with a first language other than English.

EBacc science points score by route, pupils entered for two or more science GCSEs, 2016

(Note that I have censored the triple science line in the chart above due to the low numbers of entries among pupils with a KS2 mean fine grade below Level 4).

What is causing this?

Most triple science entrants will have an advantage in the EBacc science measure in that the worst of their three grades can be discarded. But there are other issues at play.

Firstly, as we wrote here, core science and (particularly) additional science appear to be graded more severely than biology, chemistry and physics.

This seems to happen despite the efforts of awarding bodies to maintain standards between subjects with overlapping content.

However, that blogpost just looked at variation in pupils’ GCSE grades. It did not consider the possibility that pupils’ subject choices might be in part determined by their level of attainment at the time of choosing their options. The chart below suggests that this might be a factor.

It shows that pupils who took triple science tended to achieve higher grades in maths at Key Stage 4 than pupils with similar prior attainment who pursued other science options. Again, this pattern holds if a simple adjustment for pupil context is made. And a similar looking chart is produced if Attainment 8 is used as the outcome measure.

EBacc maths points score by science route, pupils entered for two or more science GCSEs, 2016

Is it something to do with school effectiveness?

There are some schools where no pupils (or very few) take triple science, and some where all pupils (or nearly all) do.

If the latter group tend to be more effective than the former then this might explain the higher attainment (taking into account prior attainment) among triple science entrants shown in the previous charts.

In the next chart we remove any schools at which fewer than 20 pupils took the triple science route, and any at which fewer than 20 took other routes to two science GCSEs. The same picture holds as before – as it does if we do the same for maths. So this does not look, superficially at least, like a school effectiveness issue.

EBacc science points score by route, pupils entered for two or more science GCSEs, schools offering both triple science and other double science routes, 2016

So is it the case that pupils who take triple science tend to be more motivated? Or does triple science somehow boost attainment in other subjects? Or a bit of both? Or something else entirely? Vote now!

Want to stay up-to-date with the latest research from Education Datalab? Sign up to our mailing list to get notifications about new blogposts, or to receive our half-termly newsletter.

[1] In the old days, triple science meant all three of biology, chemistry and physics. However, the definition was changed a couple of years ago, and it now means any three from biology, chemistry, physics or computing.Three per cent of pupils entered core, additional and further additional science in 2016. To be clear, I am not including this under the heading of triple science in this post.
By | 2017-03-23T10:33:01+00:00 March 20th, 2017|Exams and assessment|

About the Author:

Dave Thomson is Chief Statistician at FFT with over fifteen years’ experience working with educational attainment data to raise attainment in local government, higher education and the commercial sector. His current research interests include linking education and workplace datasets to improve estimates of adult attainment and study the impact of education on employment and benefits outcomes.


  1. Stephen Down March 23, 2017 at 11:15 am - Reply

    It seems pretty obvious that kids who enjoy science (and do well in science) are going to be the kids most likely to choose triple science at GCSE, whereas those who are less keen will stick to double science. This is backed up by the correlation between kids taking triple science and getting better maths results, as there is a strong overlap between those subjects.

    • Simon Laycock March 23, 2017 at 11:34 am - Reply

      Exactly what I thought.

  2. Helen Hart March 23, 2017 at 12:49 pm - Reply

    What about the effect of taking exams early. For the 2016 GCSE cohort at my daughter’s school, the Double Science candidates (including my daughter) took Core Science at the end of Y10 in 2015, whereas the Triple scientists took all science exams at the end of Y11. Surely this will affect overall results as many of the Core Science candidates in 2016 were actually Y10s rather than Y11s.

    • Dave Thomson March 23, 2017 at 1:38 pm - Reply

      Thanks Helen- that’s a really good point. Almost 80% of pupils who entered core science did so in Year 10. I’d be interested in more views from teachers about whether they think this approach a) risks a lower grade for core but b) potentially improves performance in additional science.

  3. Peter Atherton March 23, 2017 at 3:39 pm - Reply

    A lot of students might start off doing triple science but then drop off into a core and additional route. This is particularly a consideration schools might have been thinking about since Progress 8 has been counting the three best Ebacc grades in it’s calculations.

    A student who is studying triple sciences and say Geography and is projected C grades may at some stage drop onto core and additional science in order to have an increased chance of getting higher grades.

    A possible side effect of this is that it further streamlines the triple science pathway and enables greater focus on attaining the top grades with remaining students.

    • Dave Thomson March 23, 2017 at 3:51 pm - Reply

      Thanks Peter. A useful comment in many respects. Given the numbers entering core science in Y10 presumably the decision to switch from triple to core and additional took place during Y10. Will be interesting to see whether this continues under reformed GCSEs.

      • Peter Atherton March 23, 2017 at 4:11 pm - Reply

        I’m shocked by that figure to be honest but very few Y10s will be doing Core Science this year.

        • Dave Thomson March 23, 2017 at 4:18 pm

          Yes you’re right. 2017 core science entries won’t count in 2018. It does make me wonder about how comparable outcomes will work for reformed GCSEs in combined science given the sheer volume of Y10 entries in core science.

        • Alex Weatherall March 23, 2017 at 8:01 pm

          Unless I misunderstand, no students in Y10 will be doing Core Science this year as they are now doing either Combined or Single Science 9-1 GCSEs and won’t be sitting the exams until Y11 as the first exams are scheduled for 2018.

  4. Jill Noakes March 23, 2017 at 6:07 pm - Reply

    This summary of research at KCL is worth a look They found that in many cases pupils don’t really have a “choice” at all even where triple science is offered. Not being selected for triple labels kids as “not the clever sciencey ones” which can be demotivating for some.

  5. Steve March 23, 2017 at 7:48 pm - Reply

    A science hod at a school I know suggested that he had raised results at his previous school by having students sit all the units necessary for triple science but entering them for core, additional and further additional science gcses. He tells me that they ended up with 2 high (core and add) and 1 lo wish (further add) rather than 3 middling gcses because the hardest units (lower unit marks) were clumped into the further additional gcse rather than spread across 3 gcses.

    • John March 23, 2017 at 8:23 pm - Reply

      I must admit, we have had a similar experience to this (though we only did Further Additional in that first year that it was available). We’re crying out for this exact type of analysis of science results. We’re in very muddy water in terms of trying to hold science depts to account; drawing comparisons with other local schools and indeed nationally. The entry patterns are so different from school to school that it all becomes meaningless.

      I was a little surprised at the trend for higher outcomes in Triple across the ability range. We’ve recently reduced from 2 sets to 1 set taking all three because analysis of the UMS scores for ‘set 2’ showed that they were consistently scoring below the other units and dragging down what would have been slightly higher grades had they been following the Core/Add route. I’d suspected higher grade boundaries due to the able cohort generally getting more marks on the papers. – A ‘bright’ set 2 student can perform well alongside those lower ability students entered for Core/Add but might not necessarily pull their weight against the select group doing Triple??

      I realise that on a sample of one school this doesn’t preclude us being rubbish at teaching Unit 3. We generally perform well though.

      I’m sure a better statistician can tell me what I’m missing… !

      • Dave Thomson March 24, 2017 at 7:24 am - Reply

        I don’t think you’re missing anything John. What we can’t tell from the data is how outcomes might have differed if those entered for triple science had taken core and additional (and vice versa). It looks to be the case that the triple science group a) made better progress from KS2 to KS3 and/ or b) are more predisposed/ motivated to study science.

    • Dave Thomson March 24, 2017 at 7:15 am - Reply

      Fascinating. It’s certainly the case that grades in further additional tend to be a bit lower than those in core and additional (among pupils entered for all 3). Something else to take a look at- even if just for posterity.

  6. Kay Achenbach March 23, 2017 at 9:32 pm - Reply

    Hello! I’m a science teacher. What an interesting topic! Thanks for sharing it! I think the reason you see this difference is because the triple cohort in most schools is selected based on their performance in KS3. Even if two students arrive with, say, a KS2 level of 5.5, one may work hard in KS3, or mature faster, or have a more supportive home environment, or have better attendance, etc., which results in higher attainment in internal assessments in KS3. That student will be selected for triple, whereas the lazier / more disadvantaged / less mature / poorly attending student will be channeled towards the dual award. If your lines on the graph represent a mean with a cloud of data points above and below the line, in my school, we basically skim the highest performers off the top our Y9 science exam graph that would be the equivalent to your 3rd graph (before life-after-levels, our cutoff was Level 7), and we channel them into triple–so if you give them another two years and then calculated a mean based on just the selected high performers, it would be a much higher mean than that of the ones who weren’t selected, assuming the same factors that affected their KS3 performance affects their KS4 performance as well (which feels like a reasonable assumption based on the students I know–and would apply to their maths as well, and, I would predict, their other Ebacc subjects). Just out of curiosity, how would your 4th graph look if you used other subjects like English or History?

    • Dave Thomson March 24, 2017 at 7:27 am - Reply

      Thanks very much Kay. It’s good to have a science teacher’s perspective. There’s certainly something about learner progress from KS2 to KS3 and their predisposition to studying science underpinning the results. And yes, those who take triple science also tend to perform better in English and humanities than pupils who achieved similar results at KS2 but who took core + additional science.

      • Kay Achenbach March 24, 2017 at 6:17 pm - Reply

        My gut feeling is that most of what you’re seeing here is that top sets do better than lower sets, and most triple science students are in top sets. I don’t know of many mainstream state secondary schools that allow triple science to be a choice–in our local authority I know of only one that allows triple as an optionopen to anyone. It would be interesting if you could look at the spread of science results in triple and dual award compared to the spread of results in English and maths within individual schools, as this could be a good way to separate out the schools that put their top sets into triple and everyone else into core & additional. If you see that a school’s triple results are nearly 100% A* – C but their English and maths are not, then they will have a biased sample of triple students because they’ve already chosen the exam route by KS3 performance. If the spread in triple more closely resembles “academic” options subjects like history or geography, then that school probably allows triple as an option. I think if you don’t separate schools out like this, your graph is most likely just saying “top set kids with a given KS2 level do better than lower set kids with the same KS2”, which is a complete non-surprise for a science teacher in a school that sets–after all, that’s why they were allowed to take triple in the first place. But if you looked at just the “science as an option” schools, then the comparison between triple and core is more interesting, especially for the C/D grade students–because we often have students who love science but aren’t allowed to take triple for fear they’d get a worse grade (a good example of accountability measures driving students’ options, which is a sad thing to see for a science teacher like me who just wants students to enjoy the subject and do their best). I’d like to know if the lower-ability kid would most likely have done just as well if they were allowed to pick triple, once you’ve eliminated the top-set bias.

        Another interesting thought is, could you use the spread in triple science to identify top set students in schools where the spread of results indicate setting in science? It could give you a way to compare top sets and bottom sets in other subjects–which is information you wouldn’t necessarily be able to extract from, say, maths results (assuming you can use science setting as a proxy for all setting–which won’t be 100% accurate for all students, but there will be a big correlation).

        • Dave Thomson March 25, 2017 at 10:52 am

          Agreed. All the charts show is that there is selection (either by pupils or, it seems, by schools on behalf of pupils) into triple science and the criteria underpinning the selection process is driving the difference in results between triple and core/ additional. If the allocation had been random, the difference in results would be meaningful.

          I like the idea of trying to identify schools which allow triple as an option, perhaps comparing them to otherwise similar schools that don’t. The doubt I have is that this may something about the quality of the science departments in both groups. But it would be worth at least having a look at. Having end of KS3 data would be really helpful.

  7. Jon Whellan March 24, 2017 at 8:48 am - Reply

    I’d be interested to see how English prior attainment correlates. I would look at KS2 Eng as a firmer predictor of attainment in KS3 rather than maths because of the linguistic demands of decoding text in science. There seems to have been a shift in language complexity for the new GCSEs (e.g. in questions) which is also becoming evident now in KS3, so the effect may be exaggerated. Interestingly, my school managed to put out around a third of the cohort in Triple science last year, all by choice rather than being guided in anyway. There results were good and we now have a sizeable chunk of them doing sciences at A level and performing very well. However, since in science departments there is a zero sum situation with the students, the core and additional groups were reduced in terms of strong science students and they underperformed.

    Other thoughts:

    Are ‘stronger’ teachers getting Triple classes?
    Is there a sense among students that Triple science is somehow different because there is a often (?) a minimum entry requirement nor seen in other subjects?
    I strongly suspect there is a ‘critical mass’ situation that plays out subjects and schools, where once some unknown proportion of students dominates in terms of behaviour, attitude or whatever metric you choose then that becomes the overall culture of the class or school. With the selection for motivation and aptitude that occurs in Triple Science it might not be surprising if classes tend towards a positive learning atmosphere.

  8. Liam Collins March 24, 2017 at 2:43 pm - Reply

    The KS2 score they are using is a combination of English and Maths attainment.

    Student A and Student B spring to mind. A was upper ability but did double science because she was arty and not mathematical. B, liking all things science and maths, chose triple. They would have had a similar starting point but B is more likely to do better in maths. You could say, those who liked/ were better at maths at primary school were more likely to do triple and succeed in maths. Those who liked/ did better in English are more likely to choose an option other than triple.

    We could assume that, for triple scientists, a greater proportion of their KS2 score comes from their ability in maths. For those who do double science, I would wager a greater proportion of their KS2 score comes from English.

    A more worthwhile exercise would be to repeat the analysis just looking at KS2 maths as a starting point rather than English and Maths average. I would imagine the correlation would be stronger when looking at maths prior attainment rather than English and Maths combined.

    (i’d love to claim this analysis but it was our amazing AHT Theo Richards)

    • Dave Thomson March 25, 2017 at 11:08 am - Reply

      Thanks Liam (and Theo). Actually, the group entered for triple science achieve higher results in GCSE English than the group entered for core and additional. It’s probably nothing more than those entering triple science tending to be higher attaining (at KS3) and being more motivated than other pupils with similar KS2 results. In other words, they would still have achieved better results if they’d taken core + additional. Probably.

  9. Bruce Nightingale March 25, 2017 at 6:02 am - Reply

    You can see ‘STEM: Progression from GCSE to A Level’ data analysis by Andrew Powell – The analysis highlights that most students who studied for one science A’level had studied triple science at gcse and achieved level 6/7 at KS3 science.

    The RSA research highlights regional differences in accessing triple science –

    I have an interest in computer science and see many parallels between access to the subject and access to physics and subsequently progression paths. I wonder how many schools offer triple science gcse and gcse computing. I wonder what the progression rates are for pupils moving onto A’level computing from triple science gcse. I wonder how many girls study A’ level computing study at sixth form college as opposed to school sixth form?

    • Dave Thomson March 25, 2017 at 10:56 am - Reply

      Thanks Bruce. At least 5 pupils entered triple science and computing at over 1400 schools in 2016. We’ve got more work planned on transition to STEM subjects post-16 over the next year.

  10. PiqueABoo March 25, 2017 at 5:27 pm - Reply

    Why not simply: is it the case that pupils who take triple science tend to be more intelligent? It would be interesting to see how this looks with a more IQ-ish baseline than SATs.

    Outcomes are clearly highly dependent on a mix of both intelligence and industrious nature, but I suspect motivation might be more significant here given the local comp’s apparently quite common scenario i.e. school-side staffing difficulties that are largely absorbed by lower-priority KS3, together with the option of triple depending on KS3 performance.

    For instance my Y9 daughter is relatively clever, but her quite respectable science scores are partly because she just goes off to the net to teach herself topics on ‘revision’ lists for science tests that were not taught, or were quite poorly taught, in far too many dodgy KS3 supply/cover lessons.

  11. Andrew Mander April 21, 2017 at 8:12 am - Reply

    I think the key difference between the cohorts is going to be their attainment in Science in Yr 9 – students who have made the most progress in KS3 will be far more likely to take Triple science (assuming it is an option). At any school I have worked with, Triple has only been offered to a subset of students (based on their prior attainment) with only the top achievers having access to it (as I have worked in schools where no extra teaching time has been given to Triple students)

Leave A Comment