I think you should be more explicit here

``I think you should be more explicit here in step two. . . ''

INTRODUCTION TO RESEARCH DESIGN Stephen Gorard s. gorard@bham. ac. uk http: //www. tlrp. org/capacity/ rcbn. html http: //trials-pp. co. uk http: //streamingportal. multis tream. co. uk/ngflcymru/eng_ webstream_pres 02. htm

The argument for research design A key ethical concern for those funding, conducting or using publicly-funded social science research must be the quality of the research, and the security of the conclusions drawn This approach to research ethics calls for appropriate scepticism to be used more widely – the warranting principle Traditional research methods training and ‘experts’ are not enhancing quality in this warranting way. Working towards an ideal design can be an important part of any research enterprise, even where that design is not possible It is helpful to consider the research enterprise as a cycle of complementary phases and activities, because this illustrates how all methods can have an appropriate place in the full cycle of research

Reay, D. and Lucey, H. (2000) Children, school choice and social differences, Educational Studies, 26, 1, 83 -100 Interviews involved 15 parents of primary school children (p. 85). ‘a majority of the working-class parents concurred with their child… a significant deviation from this class trend was mothers of mainly black working-class boys’ (p. 90).

Reay, D. and Lucey, H. (2000) Children, school choice and social differences, Educational Studies, 26, 1, 83 -100 Interviews involved 15 parents of primary school children (p. 85). ‘a majority of the working-class parents concurred with their child… a significant deviation from this class trend was mothers of mainly black working-class boys’ (p. 90). Class? Sex of parent? Ethnicity? Sex of child? working-class and at least one other male/female black and at least one other male/female If we assume equal distribution of categories…. ‘majority of the working-class parents’ could be 4 cases, ‘mothers of mainly black working-class boys’ could not even be one case.

International comparisons indicate that intergenerational mobility in Britain is of the same order of magnitude as in the US, but that these countries are substantially less mobile than Canada and the Nordic countries. Intergenerational mobility fell markedly over time in Britain, with there being less mobility for a cohort of people born in 1970 compared to a cohort born in 1958. Source: Blanden, Machin and Gregg (2005) Intergenerational mobility in Europe and North America, London: Centre for Economic Performance, p. 2 'The comprehensive system was brought in to try to improve social mobility, but the opposite has happened' Sir Peter Lampl (http: //www. suttontrust. com/press 068. asp). ‘According to the LSE, the extra places made available by the expansion of universities in the Eighties and Nineties have largely been filled by students from better-off backgrounds’ (http: //www. suttontrust. com/press 068. asp).

Internationally comparable estimates of intergenerational mobility Country Sons born Sons earning measure Measure of parental status Partial correlation Britain 1970 Age 30 Average parental 1980+1986 . 271 Norway 1958 Average 1992+1999 Father 1974 . 139 Denmark 1958 -1960 Average 1998+2000 Father 1980 . 143 Sweden 1962 Average 1996+1999 Father 1975 . 143 Finland 1958 -1960 Average 1995+2000 Father 1975 . 147 Source: Blanden, Machin and Gregg (2005) Intergenerational mobility in Europe and North America, London: Centre for Economic Performance, p. 6

Table 3: Transition Matrix for Britain, Sons Born in 1958 Sons’ earnings quartile aged 33 in 1991 Bottom 2 nd 3 rd Top Bottom parent quartile (when son aged 16). 31 . 28 . 23 . 17 Table 4: Transition Matrix for Britain, Sons Born in 1970 Sons’ earnings quartile aged 30 in 2000 Bottom 2 nd 3 rd Top Bottom parent quartile (when son aged 16). 38 . 25 . 21 . 16 Source: Blanden, Machin and Gregg (2005) Intergenerational mobility in Europe and North America, London: Centre for Economic Performance, p. 8

The format of a warrant as a syllogism Observation/Data: This is an A Conclusion: Therefore, this is B

The format of a warrant as a syllogism Explicit Warrant: All As are B Observation/Data: This is an A Conclusion: Therefore, this is B

An effect of education? p. 238 – ‘The data are based on the entire population of schools… Consequently, the results can be considered as representative of the whole population of pupils who take both KS 3 and GCSE exams’. Multi-level model of the effect of single-sex teaching. p. 246 – ‘The effect of school type is highly significant’ Malacova, E. (2007) Effects of single-sex education on progress in GCSE, Oxford Review of Education, 33, 2, 233 -259

Contradictory trends Between 1994 and 2004, reports of studies using a researcher manipulated intervention declined from 45% to 33% in the US. But causal statements in non-intervention studies grew from 34% to 43%. In particular, complex statistical approaches such as HLM (MLM) and structural equation modelling were routinely misunderstood by researchers as testing causation. Robinson, D. , Levin, J. , Thomas, G. , Pituch, K. and Vaughn, S. (2007) The incidence of ‘causal’ statements in teaching-and-learning research journals, American Educational Research Journal, 44, 2, 400 -413

Impact of feedback? Using national data and multi-level modelling to compare the progress of schools given performance feedback by NFER. p. 102 – ‘Although no actual samples have been drawn… Statistical checks were carried out and no significant difference between the groups was found’ [i. e. in existing background measures]. p. 103 – ‘Schools who participated in PASS showed a significant difference (p<0. 05) in attainment compared to those who received feedback as part of another project. Hammond, P. and Yeshanew, T. (2007) The impact of feedback on school performance, Educational Studies, 33, 2, 99 -113

A school effect? Looks at successive years of positive value-added scores in one England district, 1993 -2002. An improving school is one where VA increases in successive years p. 271 - ‘the pupil intake and time trend explanatory variables included in the fixed part of the value-added model (Model A) were statistically significant (at 0. 05 level)’. p. 261 – ‘it appears that only one in 16 schools managed to improve continuously for more than four years at some point over the decade in terms of value-added’. Thomas, S. , Peng, WJ. And Gray, J. (2007) Modelling patterns of improvement over time: value-added trends in English secondary school performance across ten cohorts, Oxford Review of Education, 33, 3, 261 -295

Result of academisation? Df. ES - We expect that all Academies will make steady upward progress… Good teaching, excellent facilities and motivated pupils will deliver real improvements in educational standards. BBC - The government has released GCSE figures from three of its new flagship Academies in England. All the schools, which were set up in deprived areas, showed remarkable improvements in results. Bexley Business Academy 1997 1998 1999 2000 2001 2002 2003 Percentage 53 eligible for FSM 49 52 50 49 46 42 GCSE points per candidate 23 20 20 22 23 -

Achievement and attainment tables Back to the Department for Children, Schools and Families LA : Nottingham, City of Greenwood Dale School Sneinton Boulevard Nottinghamshire NG 2 4 GL 0115 9103200 MIXED Key Stage 2 to 4 Contextual Value Added Measure Key Stage 2 to 4 contextual value added measure Upper limit of CVA confidence interval Lower limit of CVA confidence interval Coverage indicator - % of students included in Key Stage 2 to 4 CVA 1068. 6 1077. 2 1059. 9 93%

Sources of errors in CVA Missing data - examples over 15% of KS 4 PLASC missing one of FSM, sex, ethnicity over 5% of KS 4 NPD have no equivalent for KS 2 over 10% of PLASC records unmatched in NPD Incorrect data – examples coding (neither male nor female) entry/transcription storage (dropout) storage (binary conversion) Measurement error – examples IDACI and the ecological fallacy Validity of assessment, marking and moderation Definitions such as SEN, ethnicity, or first language

Propagation of errors in CVA Imagine a pupil with a predicted KS 4 of 100 and an actual score of 110 Imagine that all scores are only 90% correct – a conservative view The predicted score really lies between 90 and 110 The actual score really lies between 99 and 121 CVA calculates a residual of +10 The real residual is between -11 and +31 We do not even know if the pupil has done better or worse than expected

The link between KS 2 points and GCSE benchmark, secondary schools in England, 2004 Correlation of +0. 87 Source: Gorard, S. (2006) Value-added is of little value, Journal of Educational Policy, 21, 2, 233 -241

The link between GCSE benchmark and Df. ES KS 2 to GCSE value-added, secondary schools in England, 2004 Correlation of +0. 84 Source: Gorard, S. (2006) Value-added is of little value, Journal of Educational Policy, 21, 2, 233 -241

Guess what? Secondary data analysis Feminist theory and approaches The role of theory Ethnogaphy Working overseas Adult learning Discourse analysis ………….

Outline of a full cycle of research