New : Teachers and performance related pay - what's the evidence?

Last week Schools Week published a piece by Lee Miller – Deputy Chief Executive of the Thinking Schools Academy Trust (TSAT) titled Performance related pay will solve teacher retention crisis which describes a new teacher pay structure for the TSAT – which includes

• From September 2018 a minimum starting salary of £25,000 (£2,000 above the national average)
• All progression at the upper pay levels is based solely on performance (excellent teaching)
• Teachers at the top of the scale – who exceed performance targets receive a 3% non-consolidated bonus

Having read the article – my first response was to ask: What research evidence is there to support the pay structure which has been put forward?

Pietro Marenco (@PMarencoHR and scienceatwork.com) very kindly tweeted an extract from (Weibel et al., 2009) who undertook an meta-analysis of the effect of pay for performance on performance and who state:

Our meta-analysis clearly demonstrates that the task type moderates the effect of pay for performance on performance. Pay for performance has a strong, positive effect on performance in the case of noninteresting tasks. Pay for performance, however, tends to have a negative effect on performance in the case of interesting tasks. The vignette study reveals (a) why pay for performance sometimes undermines performance and (b) how pay for per- formance produces hidden costs, which also need to be accounted for. 

Pay for performance causes a cognitive shift, that is, it strengthens extrinsic motivation for behavior (causes a price effect) and at the same time weakens intrinsic motivation for behavior (causes a crowding-out effect). Depending on the strength of these two opposing effects, pay for performance either hurts or promotes personal efforts: The more intrinsic motivation was there at the beginning, the more of it can be destroyed. 

Hidden costs arise even if the price effect is stronger than the crowding-effect. The loss of intrinsically motivated behavior has always to be compensated by external rewards.  P18

Which tends to suggest that in the context of teaching – a complex task if ever there was one (Shulman, 2004) - performance related may have a negative effect on performance and may ‘crowd-out’ intrinsic motivation (and associated discretionary effort.

If we turn now to research evidence which relates specifically to the education sector, we can look at the (EEF, 2017) summary on performance pay which states:

The results of rigorous evaluations, such as those with experimental trials or with well-controlled groups, suggest that the average impact of performance pay schemes has been just above zero. Some approaches appear to show more promise, such as bonuses or enhanced pay to attract teachers to challenging schools, or loss aversion, where the award has to be paid back if student results fall below a certain level. 

And which goes onto to state:

The evidence is not conclusive. Although there has been extensive research into performance pay, much of this is either from correlational studies linking national pay levels with general national attainment or from naturally occurring experiments. More recent randomized trials have had mixed results. Overall, it is hard to make causal claims about the efficacy of performance pay on the basis of the existing evidence

In addition, a systematic review by (Bajorek and Bevan, 2015)  of performance-related-pay in the UK public sector who found.

…. some evidence that PRP schemes can be effective across the three domains of the public sector for which there was evidence available (health, education and the civil service), but findings within and between the sectors are mixed, with scheme effectiveness often dependent on scheme design and organisational context 

Of particular interest to those who manage the teaching workforce is that there may well be - gender and age differences in the response to PRP

(Leigh, 2012)) and (Jones, 2013) report evidence to indicate that male teachers respond more positively to and support PRP schemes than their female counterparts, and Jones (2013) also highlighted that women are more likely to reduce their hours under PRP than men. Evidence also suggests that teachers with more experience display negative reactions to PRP in comparison to early career teachers (Jones, 2013; Leigh, 2013), although it is  PRP in the UK public sector  unclear whether this results from hostility to changes in the system, or previous negative experiences to PRP.(p91)

Finally, we could examine the work of (Lynch et al., 2016) on teacher retention who found

Factors that are significantly associated with intent to stay in the profession could be labelled as ‘protective factors’, worthy of attention among school leaders and policymakers. 
Unsurprisingly, we found that by far the strongest predictor is ‘job satisfaction’. Among the other factors, the strongest predictors are: 
• being proud to work at the school 
• having adequate resources 
• being well supported and valued by school management 
• having an effective governing body 
• appropriate pay for level of responsibility. P17

(Since I posted this blog @Jonathan_Haslam of the Institute of Effective Education has pointed me in the direction of this IEE Best Evidence in Brief on performance pay)

In summary

On the one hand, it would appear that the research suggests that:

• Performance related pay is not suited to complex tasks such as teaching 
• Performance related pay may reduce intrinsic motivation
• The impact of performance related on pupils’ results is just above zero
• Male teachers tend to respond more positively to PRP than their female counterparts
• The introduction of PRP may lead to female teachers reducing the number of hours taught
• More experienced teachers are more likely to display negative reactions to PRP compared to early career teachers 
• Job satisfaction is by far the biggest predictor of teachers’ intention to stay in the profession

On the other hand

• There is a problem of recruitment and retention
• The research evidence could be viewed as being inconclusive
• Pay does appear to influence teacher retention rates
• The senior team of a MAT clearly believe that they have the expertise to make it work
• There would appear to be stakeholder support from the relevant teaching union

So where does this leave us?

Evidence-based practice involves the conscientious, explicit and judicious use of the best available evidence – research, organisational data, practitioner expertise and stakeholder  in order to make decisions which will hopefully lead to a favourable outcome.

Given the nature of the evidence – the decision as to whether to proceed with PRP within a MAT is a judgment call.  That said, in making that judgement one hopes that the decision-makers have undertaken an appropriate process of due diligence and have asked the following questions.

• Is the issue an important problem for which a remedy is sought and that can be locally implemented?
• How substantial are the desirable anticipated effects?
• How substantial are the undesirable anticipated effects?
• How robust and secure are the different sources - research, practitioner expertise, stakeholder views and school data - of evidence?   
• Does the balance of the desirable and undesirable effects favour the proposal or a realistic alternative?
• How large are the resource requirements – attention, time, money, professional learning?
• What impact does the decision have on educational equity?  Will it help close gaps in attainment?
• Are there important ethical issues which need to be taken into account? 
• Are there key stakeholders – teachers, parents, trustees, who would not accept the distribution of the benefits, harms and costs?
• Would the intervention adversely affect the autonomy of teacher, department, school or MAT?
• Are there important barriers that are likely to limit the feasibility of implementing the intervention (option) or require consideration when implementing it?
• Is the intervention or strategy sustainable?

And finally

How strong is the recommendation ranging from a strong recommendation – where benefits clearly outweigh costs or vice versa, with consistent supporting evidence from research evidence, practitioner expertise, school data and stakeholders – without major limitations.  Or is it a weak recommendation with uncertainty in estimates of benefits and costs Some supporting evidence from research evidence practitioner expertise – though with major limitations.

References

BAJOREK, Z. M. & BEVAN, S. M. 2015. Performance-related-pay in the UK public sector: A review of the recent evidence on effectiveness and value for money. Journal of Organizational Effectiveness: People and Performance, 2, 94-109.
EEF 2017. Teaching and Learning Toolkit : Performance Pay. London: Education Endowment Foundation.
JONES, M. D. 2013. Teacher behavior under performance pay incentives. Economics of Education Review, 37, 148-164.
LEIGH, A. 2012. The economics and politics of teacher merit pay. CESifo Economic Studies, 59, 1-33.
LYNCH, S., WORTH, J., BAMFORD, S. & WESPIESER, K. 2016. Engaging Teachers: NFER Analysis of Teacher Retention. Slough: NFER.
SHULMAN, L. 2004. The Wisdom of Practice-Collected Essays of Lee Shulman: Vol 1, ??????, Jossey-Bass.
WEIBEL, A., ROST, K. & OSTERLOH, M. 2009. Pay for performance in the public sector—Benefits and (hidden) costs. Journal of Public Administration Research and Theory, 20, 387-412.

Guest post by Sarah Selezynov - Japanese lesson study: a question of culture

Given the publication of the EEF's evaluation report on Lesson Study, it seems sensible to take a slightly broader perspective on Lesson Study, so this week's post is by Sarah Selezynov - Programme Leader - Bespoke Leadership Programmes; IOE - Learning & Leadership; UCL Institute of Education. Sarah has extensive knowledge of Lesson Study and is currently organising two lesson study events in December - with Professor Akihiko Takahashi - to explore the Japanese approach to problem solving in mathematics, lesson study as a tool to improve teaching and learning, and the role of the koshi.

(see http://www.ioe-rdnetwork.com/uploads/2/1/6/3/21631832/takahashi_events_flyer2.pdf for more details)

As a school leader who is interested in Japanese lesson study (JLS), you are probably reading the debate on this blog with interest – Should I or shouldn’t I?  Will it make a difference to my pupils in my school?  How can I be sure that the time and effort my school invests in this will pay dividends for pupil learning?  And you are right to be cautious.

And yet, I qualify this warning by saying that I believe that JLS has great potential for teachers and pupils.  JLS aligns with the wider research base on effective teacher professional development: it focuses on learning and not performance, begins with an end goal, engages teachers in and with research over an extended time frame, in collaborative groups.  And our research has shown strong evidence of improved teacher practice and student learning (Godfrey et al, forthcoming). 

So why the warning?  Because borrowing an education policy from another country and expecting it to simply work here as it does there, doesn’t really work – it rarely has.  Pasi Sahlberg (who is not against global borrowing per se) describes how a ‘network of interrelated factors – educational, political and cultural - …function differently in different situations’ (2011: 6) meaning that we cannot be sure that any one educational approach will function in the same way when it is translated from one country to the next.

So what do we need to consider when attempting to use JLS as an approach to teacher professional development in Britain?  First and foremost, we need to understand the cultural differences between Japan and Great Britain and how this might affect teachers’ responses to JLS. 

Hofstede (2010) categorises cultural differences along five dimensions, using a 0-100 ranking.  And on three of these rankings, Great Britain has a very different score to Japan:

1. Uncertainty avoidance (Gap: Japan 11, Great Britain 68.5)

‘The extent to which the members of a culture feel threatened by ambiguous or unknown situations' (2010: 191).

The lengthy, meticulous and detailed planning of JLS, including exploring known evidence through kyozai kenkyu, and the significant time spent predicting student responses, are all attempts to avoid any unanticipated events in the research lesson.  British teachers are very likely to be much less averse to taking risks in lessons and to therefore plan in less detail and not see the need for kyozai kenkyu.  

Uncertainty avoidance cultures also feel a greater need for protocols and rules, which may explain the formal and rigid processes of JLS.  It is highly likely that English teachers would not see the need for this level of formality and would want to deviate from LS protocols.  

Uncertainty avoidance leads to a greater tendency to believe in and revere expertise: hence the valued role of the koshi or ‘expert other’ in JLS.  British teachers engaged in LS are likely to value practice expertise as much as academic expertise, and less likely to see the need for a koshi. 

2. Individualism versus collectivism (Gap: Japan 36, Great Britain 3)

Individualistic societies are those where 'ties between individuals are loose: everyone is expected to look after him- or herself and his or her immediate family' (2010: 92).  In collectivist societies, 'people …. are integrated into strong, cohesive in-groups, which throughout people's lifetime continue to protect them in exchange for unquestioning loyalty' (2010: 92).

Great Britain is a highly individualistic society: occupational mobility is higher, teachers are managed as individuals and feedback on performance is given directly.   This difference is reflected in the English performance management system for teachers, the hiring and firing based on performance judgements, and performance feedback being given directly to the teacher after a lesson observation.  Japan is a collectivist society: occupational mobility is lower, teachers are managed collectively and it would not be productive to the group to give direct feedback to an individual.   JLS has evolved as a way of giving feedback on performance through the group, with the lesson plan as a collaborative product.  We might predict that teachers in Great Britain would shy away from the live observation element of JLS, fearing a judgement on their individual professional performance, which may affect job security.   Other collaborative aspects may also be challenging to implement in Great Britain, such as committing extended amounts of time to collaborative lesson planning process and working towards a whole school shared research theme.

3. Long term orientation (Gap: Japan 3, Great Britain 40.5)

‘The fostering of virtues oriented toward future rewards - in particular perseverance' (2010: 239). 

The importance of perseverance and effort is clearly seen in JLS, where a research theme will be pursued by a school for two or three years.  This contrasts with the British short-termist attitude which is likely to influence policy decisions about the time teachers are asked to commit to investigating a research theme. 

 In summary, some key cultural differences between Great Britain and Japan are likely to mean teachers struggle with several distinguishing features of lesson study as a research process, namely:

• Focusing on a shared research theme over a longer period of time;
• Spending time on collaborative lesson planning, including exploring relevant material around their research theme;
• Being observed by colleagues as they gather evidence in the research lesson;
• Seeking outside expertise to develop and enhance their research ideas.

In our work with schools, we have managed to support teachers to engage with models of JLS that feature all of the above elements and these teachers and leaders have spoken highly of lesson study.  However, we have also encountered schools who say they are doing ‘lesson study’ but do not work on a shared research theme, nor plan collaboratively, nor act as silent observers in the lesson observation, and do not look to outside expertise to enhance their learning.  

What does this mean for you as a school leader?  If you are already using JLS, make sure that teachers are not just paying lip service to its features but adhering strictly to the features that distinguish lesson study as a research process.  If you are seeking to introduce JLS, anticipate the above cultural resistance.  Make sure teachers really understand why JLS is designed in the way it is and what you will lose if you leave out any of its critical research features.

References

Godfrey, D., Seleznyov, S., Wollaston, N., and Barrera-Pedamonte, F. (forthcoming). Target oriented lesson study (TOLS) Combining lesson study with an integrated impact evaluation model.
Hofstede, G., Hofstede, G.J. and Minkov, M., 2010. Cultures and organizations: Software of the mind (Vol. 3). London: McGraw-Hill.

Sahlberg, P., 2011. Finnish lessons. Teachers College Press.