Have a look at this link and search the studies:

"We identified 19 studies that met our standards. Most of these addressed reduced class size programs in grades K-3. Indeed, most programs in the past 20 years have targeted these early grades, in part because earlier research (see, for example, Glass and Smith 1978)...

Smaller classes in the early grades (K-3) can boost student academic achievement;

A class size of no more than 18 students per teacher is required to produce the greatest benefits;

A program spanning grades K-3 will produce more benefits than a program that reaches students in only one or two of the primary grades;

Minority and low-income students show even greater gains when placed in small classes in the primary grades;

The experience and preparation of teachers is a critical factor in the success or failure of class size reduction programs;

Reducing class size will have little effect without enough classrooms and well-qualified teachers; and

Supports, such as professional development for teachers and a rigorous curriculum, enhance the effect of reduced class size on academic achievement."

http://www.centerforpubliceducation.org/Main-Menu/Organizing-a-school/Class-size-and-student-achievement-At-a-glance/Class-size-and-student-achievement-Research-review.html

Hi Gerd, that's a very good question! It is right to question the assumption that 'small classes are better' regardless of contextual and other factors. Have you read David Pedder's work on class size?

• PEDDER, D. (2006). Are Small Classes Better? Understanding Relationships between Class Size, Classroom Processes and Pupils' Learning. Oxford Review of Education. 32, 213-234.

To quote from the abstract: 'complex classroom processes...mediate class size effects on pupils' learning' and 'Recent research evidence from secondary school classrooms calls into question simple one-way relationships between class size and pupils' learning.'

Best wishes

it seems an obvious question, however it is not so easy to find empirical results. I did not find it in my phd, not with correlations neither regression analysis. In my case, teachers with smaller sizes did not do it better than teachers with bigger size, the control of other variables like teacher training, type of students, etc.. were afecting the results.

On Hattie's review of the literature, class size does not significantly improve learning. He then asks why class size does not improve learning more, and when we should do it and when not.

http://www.sciencedirect.com/science/article/pii/S0883035506000607

For me, the words "empirical" and "why" don't always fit together. We can find reasonable empirical evidence that one variable correlates more or less strongly with another, and we can hypothesize why and study the problem from multiple angles.

But empirical research can be very consistent without spelling out a causal nexus. 

What Miranda said gives a clue to the problem with seeking emprirical data to support this hypothesis. Teachers have developed teachingstrategies that are designed to work around larger classes. They rarely involve much meaningful differentiation for different needs, so reducing class sizes has little effect except for a little more time to help students. Miranda's point that there was evidence in lower grades is for the same reason. In the lower grades differentiation is necessary because differences in student skill level is are so large in lower grades, so reduction in class size makes a measurable difference in outcomes.

Summary:

Differences in outcomes will be limited in classes where the teacher is using whole class/undfferentiated teaching practices as the gain is incremental.

Differences wil be more observable if the class size was reduced where the class teacher is trying to differentiate their teaching as the gain is proportional.

Check out this link:  http://www.ericdigests.org/pre-922/size.htm. There are lots of variables in play (of course that's a problem we social scientists always have), but there's evidence that points to a rule of thumb that "smaller is better." For reasons good and bad, rules can be broken. 

Alex Pentland's popularization of his research, Social Physics: How Social Networks Make Us Smarter, collected data to show that the equality of participation and the diversity of ideas generated in a group are two factors that have the most to do with improved intelligence of the group (measured in terms of group problem-solving). There might be a kernel of an answer to your question in this finding. Not all small classrooms facilitate learning, but when they do, it might be explained, in part, by these factors.

I teach class sizes of 15 my whole school, and a senior class of 7. You have to adjust your teaching to suit small classes. Check Hattie. A big class approach does not translate to smaller classes. I have written about it.

https://independent.academia.edu/PeterFarrell/Curriculum-&-Pedagogy

You need qualitative research to answer that. I am a researcher now but have been a primary school teacher a long time and from personal experience I can say that I would tire out after going through the 20th copy book, or simply not find time to go over the pupil's process in math. There just is not enough time. Do the math and see as well.

While it is true that Hattie's data (from a meta analysis) shows little gain with class size, I wonder sometimes about whether the reduced class size was accompanied by a change in teaching style. If not, then the results are as onemight expect. I think that reduction of class size has a second tier effect ie if you change the class size, this allows a more personalised teaching which leads to better outcomes, and I don't think this has been examined carefully enough.

Dear colleagues, first many thanks to You all  for your helpful comments.

I agree to Mark that we don´t know enough about the practices of teachers in small classes.

Hattie writes in his mataanalysis:

"One reason for these small effect sizes relates to teachers of smaller classes adopting the same methods as they were using in larger classes and thus not optimizing the opportunities presented by having fewer students. It is difficult, however, to find studies that investigate or that demonstrate whether the nature of classroom experiences are different in the smaller than the larger classes."  and

"A typical response to this lower than expected effect of reducing class sizes is to note that many of the more powerful influences identified in this book could be more effective if the class sizes was lower. … As noted above, this lack of outcome differences is most likely because teachers do not change their current teaching strategies."

David Patter comes to similiar conclusions in his article "Are small classes better? Understanding relationships between class size, classroom processes and pupils’ learning" (Oxford Review of Education,Vol. 32, No. 2, May 2006, pp. 213–234).

My actual position is: there are  positive effects of small class sizes on learning outcomes, but only if teachers use the chances of small classes in their teaching practises (see my article "Kleinere Klassen haben (doch) Einfluss auf den Lernerfolg - Eine kontroverse Diskussion zwischen Öffentlichkeit und Bildungsforschung")

Similarly, I agree with the questions and elaborations by both Mark and Gerd in relation to John Hattie's effect size for small classes not confirming that smaller classes have a positive effect on student learning. I believe that teaching is a wicked problem, and there is no silver bullet or silicone bullet (e.g. some look for the technology to magically improve student learning without considering the complexity of teaching and learning).

Therefore, some caution is needed in understanding smaller classes more deeply in terms of the what, the how and the why of implementing smaller classes. For example smaller classes might enable improved and more timely feedback to individual students, more personalised, individualised learning plans being designed and implemented for students with learning difficulties, and more effective engagement with parents and carers. There is research which suggests that the research is inconclusive, but that classes might also take into account the nature of the learners, teaching preparation, teaching approaches, purpose of the classes, etc; e.g.

- special needs (Zarghami & Schnellert. Class Size Reduction: No Silver Bullet for Special Education Students' Achievement. International Journal of Special Education, v19 n1 p89-96 2004 - see http://eric.ed.gov/?id=EJ852046)

- early childhood (Project Star - http://www.ascd.org/publications/educational-leadership/feb96/vol53/num05/Students-Achieve-More-in-Smaller-Classes.aspx)

- adult learners (e.g. many 'Master Classes' deliberately keep the numbers very small to be able to call it a 'Master Class' and achieve deeper learning; many Conservatoires and design studios adopt 1:1 teaching)

Mosteller, F. (1995). The Tennessee study of class size in the early school grades. The future of children, 113-127.

Nye, B., Hedges, L. V., & Konstantopoulos, S. (2000). The effects of small classes on academic achievement: The results of the Tennessee class size experiment. American Educational Research Journal, 37(1), 123-151.

From my own experience I say that with 6 students the teaching is most effective. I know all by name, I know exactly what they understand and what they produce. I know their strength and their weakness. 

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3817785/

https://www.heacademy.ac.uk/system/files/resources/small_group_teaching_1.pdf

Studies that show no improvement in smaller class sizes have been used politically to 'prove' that smaller class sizes make no difference (and hence not worth spending money on). In each of these studies I have seen no suggestion that teaching strategies were changed. It seems to me that the essence of smaller class sizes is that they allow teachers to use more personalised pedagogy and focus on important psychological needs related to motivation.

Dear Mark,

thank you for your answer.

In my meta-analyses I found similiar results. Teachers often do not use the possibilities and chances of smaller learning groups, their teaching practice in smaller groups does not differ from their practices in bigger groups

Best regards from Germany

Gerd  

Dear Gerd, 

That is difficult to understand. I hated to have 180 students in a group and loved to have my small efficient groups.

Alles Gute aus Finnland! 

Dear Beatrice Marianne,

many thanks for your interesting answers and docs.

Ich schicke dir einen  link einer Analyse von mir zu dem Thema, da ich sehe, dass du deutsch sprichst.

Alles Gute aus Bayern

Gerd 

https://www.researchgate.net/publication/279715315_Kleinere_Klassen_haben_doch_Einfluss_auf_den_Lernerfolg-_Eine_kontroverse_Diskussion_zwischen_Offentlichkeit_und_Bildungsforschung

Article Kleinere Klassen haben (doch) Einfluss auf den Lernerfolg- E...

These links may help you

http://www.centerforpubliceducation.org/Main-Menu/Organizing-a-school/Class-size-and-student-achievement-At-a-glance/Class-size-and-student-achievement-Research-review.html

http://www.centerforpubliceducation.org/Main-Menu/Organizing-a-school/Class-size-and-student-achievement-At-a-glance

http://www.centerforpubliceducation.org/Main-Menu/Organizing-a-school/Class-size-and-student-achievement-At-a-glance/Class-size-and-student-achievement-References.html

http://www.centerforpubliceducation.org/Main-Menu/Organizing-a-school/Class-size-and-student-achievement-At-a-glance/Class-size-and-student-achievement-Resources-.html

https://www.ilr.cornell.edu/sites/ilr.cornell.edu/files/WP136.pdf

http://www.cbc.ca/news/canada/manitoba/small-class-sizes-experts-1.4027094

https://www.classsizematters.org/research-and-links/

https://chaschool.org/benefits-small-class-size/

https://www.brookings.edu/research/class-size-what-research-says-and-what-it-means-for-state-policy/

http://www.ncte.org/positions/statements/why-class-size-matters

Have you seen the study at: http://www.wallacefoundation.org/knowledge-center/Documents/How-Leadership-Influences-Student-Learning.pdf

"Among the explanations for small class effects are improved teacher morale, more time spent by teachers on individual instruction and less on classroom management, along with fewer disruptions and fewer discipline problems. Other explanations for small class size effects include greater engagement by students in instruction, more opportunities for better teaching to take place, reduced grade retention, reduced dropout rates in secondary schools and

increased aspirations among students to attend college."

http://www.wallacefoundation.org/knowledge-center/Documents/How-Leadership-Influences-Student-Learning.pdf