Home

UK Innovation Survey - User Guide, 1st edition

1. CIS3_ARD dta merged CIS3 and ARD 2000 data The DTI sent the original data in SPSS format We transferred it using Stat_Transfer into a new STATA dataset called CIS3 dta We then cleaned this dataset using the file cis3_clean do The cleaned dataset is called cis3_clean_O1 dta At the same time we create the ARD population in 2000 using version 1 1 with the file called do_ard_2000 do and generate the dataset called ard_2000 dta We match the cleaned CIS3 dataset cis3_clean_0O1 dta to the 2000 ARD dataset ard_2000 dta using the do file merge_cis3_ard_2000 do The matched dataset is called CIS3_ARD dta 5 Detailed description of programs The original SPSS file received by DTI has been transferred to STATA using STAT Transfer Care is needed so that the reporting unit and the enterprise identifiers are transferred as double variables 5 1 CIS3_clean do Part 1 correcting problems identified by DTI CIS3_clean do checks the original CIS3 and implements some of the cleaning steps that the DTI documents report as being implemented but do not appear in the file we received Namely e We replace numerical values 1 and 2 for capex98 with missing values e The financial variables turn98 turn00 export98 export00 capex98 capex00 xinterm xextram xmachm xknowm xdesignm xtrainm xmarketm xtotalm have to be divided by 1 000 for the reporting units in the DTI document we find another reporting unit for which this correction needs t
2. Services and Process Innovation ccsscsssssssssseeseessereees 22 510 New or significantly improved QOOdS cccccccccccceeceeeseeeeceeeeeeeeeeeeeeeeeeeeees 22 520 New or significantly improved services ccccececeeceeeeeeeeeeeeeaeeeteeeeeeeeeees 22 601 Services developed by business enterprise ccc ccceccceeeceeeceeeeeeeeeeess 22 602 Services developed by business with another business ccccceeeeeeees 23 603 Services developed by other DUSINGSSES ccccccccccccceeeeceeeeeeeeeeeeseeeseeeess 23 610 Goods developed by business enterprise ccccccccccccccecececeseeeeeeeeeeeeeeeeess 23 620 Goods developed by business with another DUSINESS cccccececeeeeeeees 24 630 Goods developed by other DUSINESSES ccccccccccccceeeeeeeeeceeeeeeeeeeeeseeeeeeeees 24 710 Goods and services New to Market cccccccceeeeeeeeeeeeeecneeeeeeeesssneneaeeees 24 720 Goods and services New to business ccccccccccceceeeeeeseeeececeeceeeeeeeseeeeeeeess 25 810 Turnover split Goods or Services new to market cccccccccceceecceeeeeeeeeeees 25 820 Turnover split Goods or Services new to the business ccccceeeceeeeeeees 25 830 Turnover split Goods or Services significantly improved but not new 26 840 Turnover split Goods or Services unchanged or marginally modified 26 900 New
3. 53 48 43 46 45 1 929 271 2 200 829 118 947 438 44 438 51 2 008 383 2 391 918 123 1 041 46 32 44 60 64 1 327 446 1 773 601 172 773 45 39 44 65 67 823 312 1 135 331 74 405 40 24 36 70 74 2 751 572 3 323 1 194 192 1 386 43 34 42 Total 15 412 4 190 19 602 6 784 1 388 8 172 44 33 42 Source CIS3 Response rates were reasonably consistent across industrial sectors ranging from 36 to 46 However in nearly all sectors SMEs were more likely to respond than large firms A note on weighting Sample based business survey data are commonly grossed up to make inferences about the whole business population and in order to do this a set of weights is applied to the raw data There are two potential sets of weights that can be used namely business weights and employment weights to adjust estimates to account for firms not included in the sample In the past CIS data have typically been weighted using business weights which are calculated as the inverse sampling fraction of the number of firms in each cell of the stratification Using business weights each firm in the population carries an equal weight This can provide basic measures of population performance but does not correct for differing firm sizes Using employment weights which are calculated using the inverse sampling fraction of employees in each cell of the stratification allows data to be grossed up and corrected for firm size So the g
4. SIC 55 These sectors were not included in the 2001 survey CIS3 The 2005 survey sample was drawn from the ONS Inter Departmental Business Register IDBR in December 2004 Response and Weighting The questionnaires from the initial survey were distributed on March 31 2005 Valid responses were received from 16 446 enterprises to give a response rate of 58 per cent national STATISTICS Virtual Micro Data Laboratory Data Brief Spring 2007 Community Innovation Survey Tomas Hellebrandt The Community Innovation Survey CIS is a survey conducted every 4 years by EU member states to measure progress in the area of innovation The CIS complements other indicators of innovativeness by providing a regular snapshot of innovation inputs and outputs and the constraints faced by businesses in their innovation efforts This data brief provides an overview of the UK CIS and describes how the CIS data has been used for research in conjunction with other ONS business data sets held within the Virtual Microdata Laboratory The brief concludes with an analysis of how various firm characteristics influence innovative activity 1 Overview of the CIS The CIS is based on a core questionnaire developed by the European Commission EuroStat and member States to which the DTI adds questions for the purpose of the UK CIS The survey structure has changed over time In general the survey covers product process and wider innovation including expendit
5. Transport amp storage 8 490 8 110 380 64 1 Post amp courier activities 580 555 25 64 2 Telecommunications 615 555 55 65 67 Financial intermediation 3 965 3 620 345 70 Real estate 4 745 4 605 140 71 Renting 2 085 2 020 60 72 Computer amp related activities 5 290 5 145 145 73 1 R amp D natural sciences amp engineering 540 480 60 73 2 R amp D social sciences amp humanities 70 65 0 74 2 Architectural amp engineering activities 4 260 4 150 110 74 3 Technical testing and analysis 350 330 15 rest of 74 Other business activities exc SIC 74 2 amp 74 3 25 865 24 735 1 130 Total 170 735 164 250 6 485 Recommendation Sector stratification as in table 5 with sample selection based on optimal allocation for SMEs It might also be sensible to take a census of SMEs in certain sectors where the population is particularly small for example in SIC 40 41 and 73 2 and also to cap the sample in sectors where the population is particularly large such as SIC 50 51 52 55 and rest of 74 Sample Design Scenario Analysis Four scenarios for sample designs are now considered In each case it is assumed that the minimum cell size in the sample is 5 and that a census is taken for firms with 250 employees The pros and cons of each design are considered below each scenario Scenario I Optimal Regional Allocation This scenario considers the sample required to achieve the regional precision levels presented i
6. UK CIS4 Introduction The Community Innovation Survey CIS is traditionally based on a stratified random sample drawn from the ONS Inter Departmental Business Register IDBR A considered sample design is essential to ensure that the data collected are as precise as possible and representative of the population of interest Key users of disaggregated data include regional analysts in the Regional Development Agencies RDAs and the Devolved Administrations DAs as well as analysts of industrial statistics This paper considers appropriate methodologies goes on to investigate the structure of the population counts in the IDBR and finally makes a proposal for a design for the sample The CIS4 population CIS4 will again be based on a stratified random sample using the same stratification variables as in CIS3 namely sector region and sizeband Sector coverage Coverage of the following sectors in the target population is required under an EU regulation on innovation statistics SIC 10 14 Mining and quarrying SIC 15 37 Manufacturing SIC 40 41 Electricity gas and water supply SIC51 Wholesale trade SIC 60 64 Transport storage and communication SIC 65 67 Financial intermediation SIC 72 Computer and related activities SIC 74 2 Architectural and engineering activities SIC 74 3 Technical testing and analysis All of the above sectors were also covered in CIS3 in the UK In addition to these it
7. cases the variables RDCONT and RDOCCAS are equal to 1 even firms does not do any internal R amp D according to question 9 1 In 7 cases there are no R amp D personnel no expenditure in R amp D but R amp D is done continuously in 1 case the R amp D is done extramurally e SUPPORT there are various inconsistencies but perhaps as noted by DTI is a problem with the survey itself and it is not possible to correct for those e Wider innovation we replace missings with zero following the DTI method when an answer for another question is present After the cleaning the dataset created is ready to be matched at the reporting unit level with any other dataset 5 2 Matching the CIS3 with the ARD 2000 We match the cleaned CIS3 dataset to the 2000 ARD dataset at the reporting unit level using the do file merge_cis3_ard_2000 do The matched dataset is called CIS3_ARD dta This contains matched observation and observations only present in CIS3 Note that there are discrepancies in the sic92 classfication as clear when comparing cis3_sic92 and sic92 The dataset contains only few variables from the ARD but it can be merged back to the ARD data to add information from the survey 6 Reviewer s comments Reviewed by lt name gt lt date gt 6 1 Information 6 2 Technical accuracy 6 3 Semantic issues 6 4 Other comments 7 User comments 7 1 Comment by lt name gt lt date gt
8. changes Similar tests were carried out using turnover over employment and innovation expenditure over employment Those showing blatantly unfeasible data were cross checked against the actual images of the CIS forms and amended accordingly The chief cause of this was unitary errors s instead of 000 s e Some forms displayed logical inconsistencies Q some enterprises claimed they received no financial assistance however also claimed that they took part in schemes which entailed financial assistance This is a possible error with the questionnaire a A number of firms also claimed that they engaged in innovation activity both continuously and occasionally e In the question on the percentage of employees holding degrees in 1 science and 2 other subjects were the respondent has filled in one box but not the other we have entered a zero in the empty box Data Cleaning Applied to CIS3 Data 20 05 02 A number of respondents to the CIS3 questionnaire have entered their financial details incorrectly in that where we ask for units in thousands of pounds they have answered in pounds It would be impossible to detect all of these but I have amended some of the blatantly obvious cases Within the dataset there are two variables turn00 turnover in 2000 and idbrturn the enterprises turnover according to the inter departmental business register in no specific year Where turn00 is over 750 times greater than idbrturn I have divided all th
9. details of cleaning see the audit documents and other files held within the documentation drive X The CIS 3 had previously been linked to the 2000 ARD by reporting unit reference creating the dataset cis3_ard_2000 dta It is held within the CIS_data archive zip within the source drive This linkage includes a number of assumptions about behaviour and may not be appropriate for all users While users may use this file it may be worthwhile to create your own linked panel to the latest version of ARD2 The default linking mechanism is simply to use the Reporting Unit reference dlink_ref2 common to both files and most other BDL files Known data issues There are a large number of concerns about the data primarily as the result of obvious inconsistencies As a result a lot of marker variables were created during the cleaning process These are summarised in the variable reference xls held within the documentation drive X CIS for more details please refer to the audit document The dataset has not been thoroughly investigated yet nor have the inconsistencies been corrected Some of these are probably due to the design of the survey some to misunderstandings by the respondents Some are probably clerical errors eg 0 6 of innovation expenditure breakdowns in the CIS3 do not add up to the reported total A specific issue is that the CIS2 contains three duplicated RU references For each of the duplicated references there is no clear
10. divided by 1 000 for the reporting units in the DTI document We find another reporting unit for which this correction needs to be made and we add this to the list given to us by the DTI We replace numerical values 1 and 2 for capex98 with missing values We checked for duplicates There are none This do file creates indicator variables for observations that present inconsistencies Firstly one of the main concerns is that reporting units are not reporting at the reporting unit level i e they might be reporting at the local unit or at the enterprise level This concern is caused by major differences between turnover and employment figures both in 1998 and 2000 reported by the reporting units versus the figures from the IDBR register at the time the reporting unit was selected for the survey We create an indicator variable for these observations The dummy variable is called checkturn98 checkturn00 and it is equal to one every time the turnover reported by the reporting units is greater than 2 IDBR_turnover and less than 0 5 IDBR_turnover Similarly a variable called checkemp98 checkemp00 was created for the employment variable Secondly we looked at presumably incredible values for turnover per employee figures We generated a dummy variable checkte98 checkte00 equal to one if the difference between the turnover per employee as reported by the reporting unit is greater than 2 times or less than half the turnover per employee
11. forthcoming Canadian Journal of Economics Crespi Criscuolo Haskel and Slaughter 2007 Productivity Growth Knowledge Flows and Spillovers mimeo CeRiBA Swann 2002 Innovative Business and the Science and Technology Base Report for DTI file L UK_CIS_ 20Innovation 20Survey CIS_Info Note 200n 20the 20CIS4_CIS2007 20panel 20data txt Please note the variables included in the CIS4 amp CIS2007 panel data are the same as those found in the CIS2007 Also a time variable is included in the panel data This is labelled CIS_Version and equals 4 to indicate data from the CIS4 and 2007 to indicate data from the CIS2007 The data should therefore be sorted by ruref and CIS_Version Because not all of the questions in the CIS2007 are included in the CIS4 there will be missing values for some variables which were not included in the CIS4 Please note that approximately 7000 of the approx 15000 companies surveyed in each of the surveys have been matched to create the panel file L UK_CIS_ 20Innovation 20Survey CIS_Info Note 20on 20the 20CIS4_CIS2007 20panel 20data txt 2 8 2011 10 25 31 CLEANING THE CIS2 Question 1b and 2b e Replace prod_1 prod_2 prod_3 procecss_1 process_2 and process3 with missing if prod_inn proc_inn 0 nov_inn prod_inn is missing and nov_inn 1 these are service sector firms turn_nov in 63 cases nov_inn 1 but turn_nov 62 of these cases are in services only 1 in manufacturing PROBLEM WHAT DO WE D
12. in CIS2 and ruref in CIS3 and CIS4 There are 789 reporting units which appear in CIS2 and CIS3 and 959 reporting units which appear in CIS3 and CIS4 One hundred and one reporting units appear in all three CIS surveys Linking of information on the same enterprise between surveys provides the opportunity to explore research questions that otherwise would not have been possible The largest and most comprehensive ONS business survey is the Annual Business Inquiry This survey includes information on turnover costs employment and investment Due to the size and content of this survey the ABI generally forms the spine against which most linking activity takes place Responses to the CIS can therefore be linked to information collected on these organisations collected from the ABI Within the VML information from the ABI is held in the Annual Respondents Database ARD To reduce compliance costs the ABI is not a census of all businesses with smaller reporting units being sampled Within the ARD there are therefore two types of enterprise Information collected directly from the survey returns of the ABI are held on the selected files of the ARD Information on those organisations included within the ABI survey universe but which are not included within the actual survey during a given year are held on the non selected files By including information from the non selected ARD files the coverage of the ARD is broadened considerab
13. s eseseeeeeees ereer rrree errereen 41 1905 Innovation constraints Lack of qualified personnel ceceseeeeeeeeees 41 1906 Innovation constraining factors Lack of information on technology 41 1907 Innovation constraints Lack of information on markets cccccccceeeeeeees 42 1908 Innovation constraints Market dominated by established businesses 42 1909 Innovation constraints Uncertain demand for innovative goods or services 42 1910 Innovation constraints UK Government regulations cccceeeeeseeeeees 43 1911 Innovation constraints EU regulations ccccecceceeeeeeeeeeeeeeeeetesneneeeeeees 43 2130 Innovation protection apply for a patent ccccececceeeeeeeeeeeeeeeetetteneeeees 43 2110 Innovation protection register an industrial design ceeeeeeeeeeeees 43 2120 Innovation protection register a trademark ccccccececeeseeeeeeeeseeneeeeeees 44 2150 Innovation protection produce materials eligible for copyright 44 2210 Financial support UK local or regional authorities ccceeeeeeeeereeees 44 2220 Financial support UK central government ccccceceeeeeteeeeeeeeeeteneeeeees 44 2240 Financial support European Union institutions or programmes 45 SECTION E General Economic Information 2 ccccesesseeceeeeeeee
14. the three CIS surveys Figure 4 suggests that innovative activity has increased in manufacturing between CIS3 and CIS4 whereas it has declined in construction utilities and the distribution and service sectors Electricity gas and water supply is notable for the large fall in reported innovation over time Figure 5 suggests that whilst innovative activity was relatively 3 This variable is limited to a dummy designating whether or not the firm employs at least one employee with a degree Respondents were asked to report the proportion of employees who have a degree but many appear to have reported the number of employees instead The reported figures could therefore not be used to construct a more detailed measure of the education attainment of employees Innovation active firms in CIS3 and CIS4 are defined as those engaging in product process or wider innovation or innovation projects which have been abandoned or ongoing or one or more innovative activities in the Innovative activities and expenditures part of the survey The structure of the CIS2 questionnaire is quite different to the latter CIS surveys In CIS2 innovation active firms are defined as those that have undertaken service innovation or innovation projects which have been unsuccessful terminated delayed or not yet completed questions 3 and 13 or one or more organisational changes or new management techniques in question 4 or one or more innovative activities in questio
15. 172 firms 3 605 are in services and 4 567 in production Table 1 CIS3 reporting unit profiles CIS3 1 Number of Reporting Units 8 172 2 Number of Reporting Units in Services 3 605 3 Number of Reporting Units in Production 4 567 Source Authors calculations 4 Program datafile structure The DTI sent the original data in SPSS format We transferred it using Stat_Transfer into a new STATA dataset called CIS3 dta We then cleaned this dataset using the file cis3_clean do The cleaned dataset is called cis3_clean_01 dta At the same time we create the ARD population in 2000 using version 1 1 with the file called do_ard_2000 do and generate the dataset called ard_2000 dta We match the cleaned CIS3 dataset cis3_clean_01 dta to the 2000 ARD dataset ard_2000 dta using the do file merge_cis3_ard_2000 do The matched dataset is called CIS3_ARD dta 5 Detailed description of programs The original SPSS file received by DTI has been transferred to STATA using STAT Transfer Care is needed so that the reporting unit and the enterprise identifiers are transferred as double variables 5 1 CIS3_clean CIS3_clean checks the original CIS3 and implements some of the cleaning steps that the DTI documents reports as being implemented but do not appear in the file we received Namely the financial variables turn98 turn00 export98 export00 capex98 capex00 xinterm xextram xmachm xknowm xdesignm xtrainm xmarketm xtotalm have to be
16. 2 356 93 110 559 74 3 60 45 15 120 rest of 74 1910 827 1130 3867 Total 16457 5032 6495 27984 Pros Provides optimal national level estimates Consistent with CIS3 methodology Cons Regional imbalance with larger regions having significantly larger samples than smaller regions and hence more precise estimates Scenario 3 Floor Allocation This scenario was carried out in the following stages e The sample for Wales Scotland and Northern Ireland was as in scenario 1 e For the English regions a floor sample size of 1 500 was initially set in order to provide a minimum level of regional comparability e The remainder of the sample approximately 8 000 firms was distributed across England using Neyman allocation Table 8 Sample design under scenario 3 Floor Allocation by region 10 49 50 249 250 Total North East 1382 298 225 1905 North West 1359 446 705 2510 Yorks and Humber 1355 448 530 2333 East Midlands 1471 396 445 2312 West Midlands 1335 493 580 2408 East 1477 437 515 2429 London 1064 473 1190 2727 South East 1311 468 965 2744 South West 1589 405 445 2439 Wales 1271 504 240 2015 Scotland 1162 449 485 2096 Northern Ireland 1852 221 175 2248 UK 16628 5038 6495 28161 Pros Floor Allocation by industr 10 49 50 249 250 Total 10 14 100 70 45 215 15 22 1 125 484 720 2 329 23 29 1 463
17. 626 790 2 879 30 33 486 214 275 975 34 35 247 155 240 642 36 37 499 155 95 749 40 41 40 25 30 95 45 1 716 314 350 2 380 90 51 2 333 571 595 3 499 52 1 533 204 540 2 277 55 2 082 337 350 2 769 60 63 982 231 380 1 593 64 1 109 60 25 194 64 2 103 70 55 228 65 67 504 143 345 992 70 340 121 140 601 71 186 100 60 346 72 347 143 145 635 73 1 105 90 60 255 73 2 55 0 0 55 74 2 352 120 110 582 74 3 106 65 15 186 rest of 74 1 815 740 1 130 3 685 Total 16 628 5 038 6 495 28 161 Strikes a balance between regional and industrial national accuracy Data requirements in the devolved administrations maintained Cons Small firms in SIC 50 51 wholesale and automotive retail 55 hotels and restaurants and rest of 74 other business activities excl technical consultancy and testing have exceptionally high samples accounting for nearly a quarter of the overall sample 10 Scenario 4 Hybrid Allocation This scenario was identical to scenario 3 but with the following additional stage e The sample of small firms in SIC 50 51 55 and rest of 74 was capped to 1 000 This freed up around 3 000 firms which were distributed across the whole of the UK using Neyman allocation Table 9 Sample design under scenario 4 Hybrid Allocation by region 10 49 50 249 250 Total North East 1093 352 225 1640 North West 1370 528 705 2598 Yorks and Humber 1293 527 530 2350 East Mi
18. Business changes None of the above ceceeecececeeeeeeeeeneeeeeeeetsenenseeeeees 14 1 Business objective Profit margin ON SAaleS eccceceeceeeeeeeeeeeeeeeeeeseseeneeeeeees 14 2 Business objective Growth in sales turnover ccccceeeeeeeeeeeeeeeeeeeseeneeeeees 14 3 Business objective Growth in Exports c cccceceeeeceeeeceeeeeeeeteeaeaeeeteesetesnenees 15 4 Business objective Market share in UK 0 cccccceceeeececeeeeeeeeneeaeeeeeeteseeneeees 15 Section B Innovation activity isisisi iiini iinit 15 1310 Innovation related activity Internal R amp D ccecccceeeeeseeeeeeeeeeeeseeneeeeeees 15 1320 Innovation related activity External R amp D ccccccceeeeeeeeeeeeeeeeeeesneneeeeees 16 1331 Innovation related activity Acquisition of advanced machinery 16 1332 Innovation related activity Acquisition of computer hardware 16 1333 Innovation activity Acquisition of computer software cccceeeeeeees 16 1340 Innovation related activity Acquisition of external knowledge 17 1350 Innovation related activity Training ccccccceeceeceeeeeeeeeceeeeeeeeessneneeeeees 17 1360 Innovation related activity Design cccccecceeeeeceeeeeeeceneeeeeeeetesssneeeeees 17 1371 Innovation related activity Changes to GeSIQN cccceececeeeeeeeeeeet
19. O WITH THIS FIRM It filled in the description of the innovation dti3tx but it did not fill in any of the turnover questions turn_nov turnnew turnimp turnung ORGANISATIONAL CHANGE e In 2 cases everything is equal to zero including intronon which should be 1 e If intronon 1 then we should put all the others equal to 0 when missings FILTER QUESTIONS e Dti innl if prod_inn or proc_inn or inun 1 then necessarily dti_innl 1 therefore we replace dti_inn1 1 if dti_innl amp proc_inn 1 or productivity_inn 1 or inun 1 e If dti_inn2 is not missing then dti_inn1 has to be zero viceversa replace missing in dti_inn if dti_innl e Lots of problems with consistency here 3 firms say they do R amp D continuously and 5 that they do it occasionally but they do not report R amp D expenditure or R amp D personnel e In general it should be the case that if firms answer no to 6b they do not answer the rest of the survey excluding question 14 onwards but this is not always the case THE QUESTION ARISE what to do with these firms Should we put their answers to missings TURNOVER DISTRIBUTION e Replace missings when we can calculate the percentage as residual from the other two categories INNOVATION EXPENDITURES e Replace 0 when it is missing if at least one of the other is not missing e Create percentages of expenditures turnover 1996 as reported in CIS2 and generate a marker out_pc_ var for those observations that have
20. UK Data Archive Study Number 6699 UK Innovation Survey Secure Data Service Access CIS Quick Guide for Users Dataset Community Innovation Survey CIS Dates available 1994 1996 CIS2 1998 2000 CIS3 2002 2004 CIS4 2005 2007 CIS 2007 mini Survey questions Innovation Collected by CIS2 CIS4 DTI CIS2007 DIUS Link fields IDBR reference Legal restrictions Voluntary survey not covered by STA covered by NS Code of Practice Quick summary The Community Innovation Survey CIS is a voluntary postal survey carried out by ONS on behalf of the DTI now DIUS Eurostat proposes an initial questionnaire and the DIUS adds questions ONS randomly selects a stratified sample of firms with more than 10 employees drawn from the Inter Departmental Business Register IDBR by SIC92 2 digit class and 8 employment size bands The survey covers both the production manufacturing mining electricity gas and water construction and the service sectors The retail sector has been excluded from the survey as this sector has been a poor responder in previous surveys and generally has shown very little innovation An enterprise is defined as being innovation active within the period 1998 2000 if it has e Introduced a new or significantly improved good service or process e Engaged in innovation projects that are not yet complete e Engaged in longer term innovation activities such as basic R amp D or technology watch e Had inno
21. a percentage equal to or higher than 100 R amp D personnel e Create percentages of r amp D personnel employment 1996 as reported in CIS2 and generate a marker out_pc_rdper for the 1 observation that has a percentage equal to or higher than 100 For now we mark observations firms for which pc_rdper gt 100 for the analysis we are likely to drop them e Create marker weird_rdper 1 if rdper gt 0 amp intra_rd 0 or intra_rd weird_rdper 2 if rdper 0 amp intra_rd 1 R amp D ACTIVITY e try to correct for inconsistencies if you have not spent anything in R amp D and do not have any R amp D personnel how can you do R amp D We create a marker for these cases weird_rdcon 1 if intra_rd 0 amp extra_rd Olextra_rd Weird_rdcon 2 if rdcon lt 3 amp intra_rd 0 amp extra_rd Olextra_rd amp rdper gt 0 amp rdper Ray explanation for this Question On Q 8 of CIS2 firms answer they are doing NO intramural R amp D but do employ personnel and do R amp D continuously Equally there are firms who report they are doing intramural R amp D but they do NOT employ personnel and do NOT do R amp D continuously Answer is that both groups are indeed engaged in intra mural R amp D The first group have mostly just miscoded their answer to the Xinter question though some may have used their R amp D staff for other purposes in the accounting year in questions must be a small The second are undertaking R amp D in the survey year but wi
22. acted from the IDBR at the time of the main sample extraction and this will accompany the main reporting unit level dataset Note that it may also be possible to obtain information about the enterprise groups for the selected reporting units and all reporting units within them Sample selection methodology At the CIS4 project board meeting of 2 July 2004 it was agreed that the sample would be selected using a Neyman allocation based on firms innovation active rates in CIS3 This will help ensure that optimal precision will be obtained in CIS4 assuming that the observed patterns are broadly similar over time Sample size The ONS has agreed to a size limit of 28 000 firms for the UK sample The working assumption is that the final sample will be very close to 28 000 Population counts In order to consider an appropriate sample design it is first useful to look at the population counts in a recent IDBR extract taken on 8 December The CIS4 population is heavily skewed towards the smaller firms with roughly two thirds of firms based in the service sector Of note roughly 40 of all firms in the identified population are small firms 10 49 employees in four service sectors namely SIC 50 51 amp 52 wholesale and retail 55 hotels and restaurants and rest of 74 other business activities There are also considerable differences in the sizes of the regions with the likes of London and the South East being nearly five tim
23. cannot answer a particular question and leaves it blank This type of response is coded as 8 and labelled not answered The numbers of businesses that left a particular question blank on some occasions is considerable and for this reason not answered is not treated as an invalid or missing response Innovation Concepts Innovation for the purpose of this survey is defined as new or significantly improved goods or services and or the processes used to produce or supply these Product innovation bringing to the market or into use by business new and improved products including both tangible goods and the provision of services The degree of innovativeness is shown by the distinction between products new just to the business or which are also new to the market Process innovation significant changes in the way that goods or services are produced or provided again differentiating between processes new to the business only or also new to the industry Innovation related activities categories of innovation directed investment such as R amp D capital goods and software acquisition design activity for implementing current innovations or directed to future product or process changes New to market the introduction of a new good or service to the market before competitors New to this business introduction of a new good or service that was essentially the same as a good or service already available fro
24. cument created 17 June 2003 Supercedes lt nothing gt 2 Associated documents Changes applied to the raw CIS3 data Explains cleaning by DTI official version CIS3 main questionnaire v2 nov version LABELS pdf Questionnaires and variable labels CIS3_clean do Stata do file for cleaning Do_ard_2000 do Stata do file for preparing ARD 2000 Merge_cis3_ard_2000 do Merge clean CIS3 with ARD 2000 3 General description The Community Innovation Survey CIS is a voluntary postal survey carried out by ONS on behalf of the DTI Eurostat proposes an initial questionnaire and the DTI adds questions ONS randomly selects a stratified sample of firms with more than 10 employees drawn from the Inter Departmental Business Register IDBR by SIC92 2 digit class and 8 employment size bands The IDBR excludes agriculture fishing and forestry public administration and defence education health and social work The survey covers both the production manufacturing mining electricity gas and water construction and the service sectors The Third Community Innovation Survey CIS 3 was in the field twice The first wave sampled 13 340 enterprises the second top up covered 6 285 to make the sample representative at the regional level The CIS 3 covers the period 1998 2000 Of the total 19 625 enterprises to which the survey was sent 8 172 responded Table 1 row 1 achieving a response rate of 42 Of these 8
25. cuolo Haskel and Slaughter 2007 The linked CIS2 CIS3 ARD panel is used to explore the role of knowledge flows and TFP growth Results suggest that the main sources of knowledge are competitors suppliers plants that belong to the same group and universities They find a statistically significant association between TFP growth and above firm average information flows from other firms in the enterprise group competitors and suppliers The effects are economically significant as well with such information flows explaining in a growth accounting sense about 50 of TFP growth They conclude that the main free information flow spillover is from competitors and that multi national presence may be a proximate source of this spillover 4 Which firms innovate Analysis of CIS4 The section describes the incidence of innovative firms in the CIS4 according to a number of firm and market characteristics Innovative activities are divided into three types product process and wider innovation Overall 29 of firms in CIS 4 report product innovation 20 report process innovation and 39 report wider innovation Figure 1 shows that innovation activity is most prevalent in manufacturing In construction mining and quarrying and electricity gas and water supply innovative activity has a very different composition to that in other production sectors dominated by wider innovation and process innovation outweighing product innovation in the lat
26. d compared it with the total expenditure in innovation as reported in the survey by the reporting units xtotalm In 3 359 cases they were equal in 7 cases the sum we generated is less than the total reported by the firms in 13 cases it is greater e In 39 cases the total innovation expenditure is greater than total turnover in 2000 These observations have been marked with an indicator variable outpc e There are some weird cases in the rdpers persons involved in R amp D activities within the enterprise in 2000 We flag those with the marker outrdpers Outrdpers equals one if e in5 cases R amp D personnel is greater than the number of total employees and in 4 cases R amp D personnel is a positive number and employment in 2000 is equal to 0 in 7 cases they amay be explained by the firm having incurred a structural change Finally in 13 cases R amp D personnel is a positive number and employment in 2000 is missing e Internal R amp D We flag inconsistencies between the various indicators of the R amp D activities within the firm These are xinter xinterm xextra xextram rdpers rdcont rdoccas For the observations that present inconsistencies we construct an indicator variable called outrd Outrd takes value one in the following cases e R amp D expenditure 0 even if in Question 9 1 no intramural R amp D and or no intramural R amp D expenditure e Question 10 2 how did your enterprise engage in R amp D during the three year period In 8
27. d_coop _ var 1 WE REPLACE COOPERATION 1 in these cases e FACTORS HAMPERING INNOVATION e Replace 0 when it is missing if at least one of the other is not missing e GOVERNMENT SUPPORT GOVERNMENT PROGRAMS e PROBLEM what do we do with missings in govsuppt and govprog Do we replace them with Os or do we leave them missings I have checked and there are about 70 missings in manufacturing and only 32 are justified by the filter questions e Replace 0 when it is missing if at least one of the other is not missing CLEANING FORREGRESSIONS e Generate dummy for prod_1 prod_2 etc 0 when prod_inn 0 POST MERGING CLEANING e We check level of reporting unit using turnover and employment from ARD and also location through region variable e We create a dummy for firms which report across groups USE of NON RESPONDENTS DATA e Check characteristics of non respondents e Buid weights to reweight the data basic idea selectivity on observables Changes Applied to the Raw CIS3 Data Authour Brian Stockdale DTI e When the data was originally received from ONS missing data was denoted with a 1 This was changed to SPSS s system missing as others in would ruin results e Where one part of a question had a response all the other unanswered parts were assumed to have an answer of well This entailed changing non response to nil response in some cases e Calculated percentage changes in employment turnover to look for any unfeasible
28. dlands 1351 471 445 2257 West Midlands 1323 570 580 2458 East 1405 518 515 2418 London 1241 572 1190 2988 South East 1428 563 965 2956 South West 1456 474 445 2360 Wales 1012 558 240 1790 Scotland 1033 532 485 2060 Northern Ireland 1499 266 175 1920 UK 15504 5931 6495 27930 Pros Hybrid Allocation by industry 10 49 50 249 250 Total 10 14 15 22 23 29 30 33 34 35 36 37 40 41 45 50 51 52 55 60 63 64 1 64 2 65 67 7O 71 72 73 1 73 2 74 2 74 3 rest of 74 Total 145 80 45 1329 572 720 1722 736 790 580 277 275 308 207 240 589 205 95 40 25 30 2024 384 350 1000 571 595 1785 265 540 1001 337 350 1156 294 380 169 65 25 158 85 55 601 211 345 417 176 140 246 150 60 433 194 145 155 110 60 60 0 0 425 177 110 161 70 15 1000 740 1 130 15 504 5 9316 495 270 2 621 3 248 1 132 755 889 95 2 758 2 166 2 590 1 688 1 830 259 298 1 157 733 456 772 325 60 712 246 2 870 27 930 Asin scenario 3 but also ensures that small firms in SIC 50 51 55 and rest of 74 are not given unnecessarily large sample sizes Cons Rather complicated Reduces the sample size in Wales and Northern Ireland by approximately 10 11 Conclusion It is recommended that the methodology described in scenario 4 be used for the UK CIS4 sample design This will ensure that a decent level of accuracy can be achieved at the regional ind
29. e financial data figures by a thousand This has removed many of the unfeasible situations where a painting and decorating or plumbing firm was earning around 50m per man This method is by no means perfect and still leaves several firms that are very likely to have this error However we cannot hope to track down all such errors and would like to avoid over amending the raw data If in the course of your research you find any major errors or ideas to improve the data quality that you feel we may want to know about please contact us For those of you who already had the data before the date above I enclose an annexe containing the reporting unit ru references of the cases amended and the variables which have been altered in those cases Cleaning linking audit document 1 Basic information Dataset CIS3 Major version 0 Minor version 0 Document created 17 June 2003 by Chiara Criscuolo Supercedes lt nothing gt 2 Associated documents Changes applied to the raw CIS3 Explains cleaning by DTI official data doc version CIS3 main questionnaire v2 nov Questionnaires and variable labels version LABELS pdf CIS3_clean_1_0 do Stata do file for cleaning do_ard_2000 do Stata do file for preparing ARD 2000 merge_cis3_ard_2000 do Merge clean CIS3 with ARD 2000 3 General description The Community Innovation Survey CIS is a voluntary postal survey carried out by ONS on behalf of the DTI Eurostat propos
30. eceiving government support raises design expenditure by about 3 of mean expenditure for those firms undertaking expenditure Information Technology Organisational Change and Productivity Crespi Criscuolo and Haskel 2006 This research uses the CIS3 and ARD to examine the relationships between productivity growth IT investment and organisational change Consistent with the small number of other micro studies the researchers find that a IT appears to have high returns in a growth accounting sense when organisational change is omitted when organisation change is included the IT returns are greatly reduced b IT and organisational change interact in their effect on productivity growth c non IT investment and organisational change do not interact in their effect on productivity growth Productivity Exporting and the Learning by Exporting Hypothesis Crespi Criscuolo and Haskel 2006 This research uses a matched CIS2 CIS3 panel to examine the proposition that exporting firms learn from their clients and this learning raises their productivity The research finds that a firms who exported in the past are more likely to report that they learnt from buyers relative to learning from other sources and that b firms who had learned from buyers more than they learnt from other sources exhibited higher productivity growth supporting the learning by exporting hypothesis Productivity Growth Knowledge Flows and Spillovers Crespi Cris
31. eeeeeeeeeeeees 45 2410 Business turnover 2006 ccccccceceecececeeeeceeccaeeeeeeeeseeccceeeeeeeetsseeneaeeees 45 2420 Business turmOVen 2008 iiiisisciiec cite laveletcicsnieriiestsdsieciies ARE ENEE AS 45 2510 Average employees 2006 cccccceeeccececeeeeceeceneeeeeeeesecneneeeeeeeesssnsneaeeees 45 2520 Average employees 2008 ccccccececcececeeeeseeceneeeeeeeeseencaeeeeeeeeseeneneaeeees 46 2610 Skills Science or engineering SUbDjeCtS cee ceeeececeeeeeeteeceeeeeeeeeetetteaeees 46 2620 Skills Other SUD SCUS sessisccccavavedsascesenecdeeees mianzeriecesedencvsuea delenseecaacevenestvececdcess 46 Derived Variable iiisicisscccegeasscccctasks cecdonecs sane sanssscceces se ccadesia seaueiceedeeeseies iiaiai 46 PRODINOV whether a product innovator 0 c cccccceceeccceeeeeeeeeeecceeeeeeeeesesneneeeeees 46 PROCINOV whether a process innovator cccccceccececeeeeeeeeeeeeeeeeteeeesesneneaeeees 47 ACTIVITIES whether engaged in ANY innovation related activities 0 47 INNOACT whether innovation active UK definition cccccceeeeeeeeeeeeeeeeeetees 47 W_INNOV whether a Wider INNOVALON sidccisihisiediscieste einen 47 B_INNOV whether a broader innovator ccccccceceecceeeeeeeeeeeeceneeeeeeeeessneneeaeeees 47 COOPERATE whether business cooperated on innovation c ccceceeeeeeeeeees 47 Wee light is
32. eeeeees 18 1372 Innovation related activity Market research cccccceecesceeeeeeeteseeneeeeeees 18 1373 Innovation related activity Marketing Methods ecceeeeeeeeeeeseeeeeeeeees 18 1374 Innovation related activity Launch advertising ceccceeeeeeeeeeeeeeeeees 18 1410 Innovation related activity expenditure Internal R amp D cceeeeeeeeeeees 19 1420 Innovation related activity expenditure External R amp D ccceeeeeeeees 19 1430 Innovation related activity expenditure Acquisition of machinery equipment and SONWANE 2th ats cee ser vale nne leu atl ASEET sets sna ewtserish nade wnat ESEE ES 19 1440 Innovation related activity expenditure External knowledge 0008 20 1450 Innovation related activity expenditure Training cccceceeeeeeeeeeeeeeeeees 20 1460 Innovation related activity expenditure DeSIQN cceeeceeeeeeesteeteeeeees 20 1470 Innovation related activity expenditure Market introduction of innovations 21 2310 Business strategy and practices Corporate strategy ccccccccceeeceeeeeeees 21 2320 Business strategy and practices New management techniques 21 2330 Business strategy and practices Organisation structure cccceceeees 21 2340 Business strategy and practices Marketing cccccccccccccccececeeeeeeeeeeeeeeees 22 Section C Goods
33. es an initial questionnaire and the DTI adds questions ONS randomly selects a stratified sample of firms with more than 10 employees drawn from the Inter Departmental Business Register IDBR by SIC92 2 digit class and 8 employment size bands The survey covers both the production manufacturing mining electricity gas and water construction and the service sectors The Third Community Innovation Survey CIS 3 was in the field twice The first wave sampled 13 340 enterprises the second top up covered 6 285 to make the sample representative at the regional level The CIS 3 covers the period 1998 2000 Of the total 19 625 enterprises to which the survey was sent 8 172 responded Table 1 row 1 achieving a response rate of 42 Table 1 CIS3 reporting unit profiles CIS3 1 Number of Reporting Units 8 172 2 Number of Reporting Units in Services 3 605 3 Number of Reporting Units in Production 4 567 Source Authors calculations 4 Program datafile structure cis3 whole data 10April2002 read only sav Raw SPSS data CIS3 dta Raw data in STATA format cis3_clean do STATA cleaning program for CIS data cis3_clean_O1 dta cleaned data in STATA format dat2000xxx dta nul2000xxx dta cleaned ARD data for 2000 all sectors do_ard_2000 do program to put all ARD 2000 sectors together ard_2000 dta ARD RU population 2000 cis3_clean_O1 dta cleaned CIS3 data in STATA format ge merge_cis3_ard_2000 do
34. es larger than regions such as Northern Ireland and the North East Table 2 Number of firms in UK CIS4 population by region Employees size bands Region 10 49 50 249 250 499 500 999 1000 250 Total North East 4 525 885 125 65 35 225 5 635 North West 15 120 2 905 390 160 155 705 18 735 Yorkshire and The Humber 11 445 2 235 270 150 110 530 14 205 East Midlands 10 195 1 960 235 120 90 445 12 600 West Midlands 12 585 2 400 295 160 125 580 15 560 East 12 995 2 295 280 130 105 515 15 805 London 19 915 3 935 535 330 325 1 190 25 035 South East 19 660 3 565 465 225 275 965 24 195 South West 11 850 2 020 240 110 95 445 14 315 Wales 4 890 1 020 140 70 30 240 6 145 Scotland 10 070 1 965 255 135 95 485 12 520 Northern Ireland 4 930 880 100 45 30 175 5 985 UK total 138 190 26 060 3 325 1 690 1 470 6 485 170 735 Response rates In CIS3 response rates varied according to the size of the firm Table 3 CIS3 UK response rates Numbers of businesses in Number of businesses sample with response Response rates Wide SIC group SME Large Total SME Large Total SME Large Total 10 14 275 49 324 111 16 127 40 338 39 15 22 1 641 702 2 343 768 237 1 005 47 34 43 23 29 1 959 752 2 711 904 217 1 121 46 29 41 30 33 954 340 1 294 422 105 527 44 31 41 34 35 735 204 939 285 59 344 39 29 37 36 37 941 113 1 054 388 55 443 41 49 42 40 41 69 46 115 33 20
35. et 5 2 Matching the CIS3 with the ARD 2000 We match the cleaned CIS3 dataset to the 2000 ARD dataset at the reporting unit level using the do file merge_cis3_ard_2000 do The matched dataset is called CIS3_ARD dta This contains matched observation and observations only present in CIS3 Note that there are discrepancies in the sic92 classification as clear when comparing cis3_sic92 and sic92 The dataset contains only few variables from the ARD but it can be merged back to the ARD data to add information from the survey 6 Reviewer s comments Reviewed by Naveed Khawaja August 2003 6 1 Information CIS is quite complex and users need to have a good look at the survey form before attempting to use the data 6 2 Technical accuracy Seems like all the obvious checks on the data have been done plus other checks 6 3 Semantic issues None 6 4 Other comments None 6 5 Other reviews Reviewed by Felix Ritchie October 2003 Found 2 errors in code corrected while still at approve stage No impact on published results just indicator variables Otherwise concur with NK s comments and warnings about being wary of using the data Restructured CIS_Clean_1 0 do original file was called CIS_clean do to make it readable No effect on datasets 7 User comments 7 1 Comment by lt name gt lt date gt Cleaning linking audit document 1 Basic information Dataset CIS3 Major version 0 Minor version 0 Do
36. figure calculated using the IDBR values We then went on to check other variables in the dataset We describe the details for each of them e Propsci propoth in some cases one is missing but the other is not In these cases according to DTI guidelines we replace missings with 0 e Prodnew prodimp produnc we check that the sum of the three is 100 by creating a new variable prodtot the sum of the three e Ifthe variable prodinov product innovation is not missing i e equals 0 1 and prodnov new goods services new to the market is missing the latter is replace with zero Similarly for process innovation e We replace missings in the share of turnover due to products new to the market in 2000 with zero if prodnov equals 0 e How are the new prodducts processes developed Prodwho and procwho in 15 cases prodinov equals 1 and prodwho is missing in 9 cases procinov equals and procwho equals missing We have not flagged these observations but it is worth keeping this into account for the analysis e Innovation activities not completed or abandoned Aband nyettime nyetlate noteven no case in which one part of the question had a response and another was unanswered In 4 710 cases this information is missing e Factors hampering innovation missing in 925 cases e Innovation related expenditure in 2000 e We generated the sum of all the expenditure reported in the survey xinterm xextram xmachm xknowm xdesignm xtrainm xmarketm an
37. ghtly larger sample is necessary for the larger regions to achieve a given level of precision It can be shown that the following regional distribution of the sample would be necessary to ensure 95 confidence intervals no wider than 3 Table 4 Required regional distribution to achieve 3 precision of innovation active rate estimation in CIS4 Required achieved Required sample size n total Sample N p q for 3 Response sample size in Region CIS4 CIS3 CIS3 precision rate CIS3 size CIS3 North East 5 635 0 52 0 48 896 40 2 230 1 105 North West 18 735 0 458 0 542 1 003 42 2 416 2 026 Yorks and Humber 14 205 0 455 0 545 985 42 2 320 1 773 East Midlands 12 600 0 5 0 5 984 43 2 310 1 641 West Midlands 15 560 0 531 0 469 995 43 2 306 1 696 East 15 805 0 465 0 535 995 43 2 337 1 761 London 25 035 0 478 0 522 1 022 38 2 693 2 567 South East 24 195 0 527 0 473 1 019 41 2 506 2 488 South West 14 315 0 483 0 517 992 40 2 503 1 567 Wales 6 145 0 495 0 505 909 45 2 015 840 Scotland 12 520 0 448 0 552 974 46 2 097 1 734 Northern Ireland 5 985 0 459 0 541 901 40 2 246 404 UK 170 735 11 675 27 978 19 602 Source CIS3 and DTI calculations The calculations in table 4 are based on the following assumptions e The proportion of firms in each region that are innovation active is the same as in CIS3 e Regional response rates in CIS4 are the same as in CIS3 In all regions an increased sa
38. l other characteristics constant For example does being part of an enterprise group increase the likelihood of an enterprise engaging in innovative activity or do these enterprises differ in other characteristics that may also make them more likely to innovate To do this we employ logistic regression analysis to estimate the additional and independent effect of a range of firm and market characteristics on the probability that the firm will engage in innovative activity We restrict the analysis to the most recent Community Innovation Survey In addition to the information contained within the CIS additional variables drawn from the 2004 Business Structure Database are merged onto the CIS data The Business Structure Database is a version of the Inter Departmental Business Register held within the VML Specifically we use the BSD to explore how enterprise structure and company age effect innovative activity Explanatory variables are included to control for organisational structure legal status size age export market main customers industry and region Selected results from the logistic regression are presented in Figures 6 and 7 It is noted that these results are derived from the same models presented in the Annex and are presented in separate charts purely for expositional convenience For each of the variable sets the results are expressed in terms of the percentage difference in the probability of engaging in innovative activities rela
39. ly However the range of data items held on the non selected files is more limited Table 1 shows the number of links that can be made between the CIS and the ARD differentiating between links with selected and non selected files It is noted that whilst the ABI is an annual survey the CIS covers a three year period However a number of questions in the survey refer only to the last of the three years covered by the survey This suggests that the most appropriate link is likely to be to the ABI in that year Table 1 therefore shows the number of links that can be expected when linking ClS2 to the 1996 ARD CIS3 to the 2000 ARD and CIS4 to the 2004 ARD It can be seen that for the CIS4 over 90 of enterprises can be linked to the ARD with approximately 38 being able to be linked to the detailed information contained within the selected files Table 1 CIS Survey Sample and links to the Annual respondents Database Links to the Annual Respondents Database CIS Sample Selected Files Non selected Total Links Files CIS2 2 342 248 109 357 CIS3 8 172 3 472 4 010 7 482 CIS4 16 445 6 179 8 710 14 889 Three replicated reference numbers have been removed from the original CIS sample 3 Previous Research Using the CIS As the sponsor of the UK Community Innovation Survey the DTI is one of the most active users of the survey for research purposes A summary of some of the DTI sponsored research projects u
40. m competitors Wider or Strategic innovation new and significantly improved forms of organisation business structures or practices aimed at improving internal efficiency or effectiveness of approaching markets and customers Question Variable look up table Question No OOAONDOABRWHND Variable name 210 240 410 440 1 4 1310 1374 1410 1470 2310 2340 510 520 601 630 710 720 810 840 900 1010 1030 1100 1510 1520 1501 1210 1280 1601 1690 1811 1874 2011 2030 1901 1911 2130 2150 2210 2240 2410 2420 2510 2520 2610 2620 Short description Geographic market Business changes Business objectives Innovation related activity Innovation related activity expenditure in 2008 Business strategy and practices New or significantly improved products Who developed products New to market new to business products New to market new to business products 2008 turnover New or significantly improved processes Who developed processes New to industry processes Incomplete or abandoned innovation activities Innovation marker Innovation factors Importance of information Innovation cooperation Reasons for not innovating Innovation constraints Innovation protection Financial support Business turnover Average employees Skills 3 The prefix is used for variables in the SPSS dataset For Stata variable names are prefixed with _ 10 Sample Design in the
41. m the IDBR register at the time the reporting unit was selected for the survey We create an indicator variable for these observations The dummy variable is called checkturn and it is equal to one every time the turnover reported by the reporting units is greater than 2 IDBR_turnover or less than 0 5 IDBR_turnover Similarly a variable called checkemp was created for the employment variable e Innovation related expenditure in 2000 e We generated the sum of all the expenditure reported in the survey xinterm xextram xmachm xknowm xdesignm xtrainm xmarketm and compared it with the total expenditure in innovation as reported in the survey by the reporting units xtotalm In 3 359 cases they were equal in 7 cases the sum we generated is less than the total reported by the firms in 13 cases it is greater e In 39 cases the total innovation expenditure is greater than total turnover in 2000 These observations have been marked with an indicator variable outpc e There are some weird cases in the rdpers persons involved in R amp D activities within the enterprise in 2000 We flag those with the marker outrdpers Outrdpers equals one if e R amp D personnel is greater than the number of total employees 9 cases e R amp D personnel is a positive number and employment in 2000 is missing 13 cases e Internal R amp D We flag inconsistencies between the various indicators of the R amp D activities within the firm These are xinter xinterm xe
42. mber East Midlands West Midlands East of England London South East South West Wales Scotland Northern Ireland 2 Division Coverage of the following sectors in the target population Div 10 14 Mining and quarrying Div 15 22 Manufacture of food clothing wood paper publishing and printing Div 23 29 Manufacture of fuels chemicals plastics metals amp minerals Div 30 33 Manufacture of electrical and optical equipment Div 34 35 Manufacture of transport equipment Div 36 37 Manufacture not elsewhere classified Div 40 41 Electricity gas and water supply Div 45 Construction Div 50 Sale maintenance and repair of motor vehicles P Div 51 Wholesale trade Div 52 Retail trade exc cars amp bikes and repair Div 55 Hotels amp restaurants Div 60 63 Transport amp storage Div 64 1 Post amp courier activities Div 64 2 Telecommunications Div 65 67 Financial intermediation Div 70 Real estate Div 71 Renting Div 72 Computer amp related activities Div 73 1 R amp D natural sciences amp engineering Div 73 2 R amp D social sciences amp humanities Div 74 2 Architectural amp engineering activities Div 74 3 Technical testing and analysis rest of 74 Other business activities exc SIC 74 2 amp 74 3 SIC 92 11 Motion picture and video production Note Denotes required under an EU regulation on i
43. mple size would be necessary to achieve the pre defined level of precision although in some of the smaller regions the size of the required increase is considerably greater than in the larger regions Industrial Stratification Having proposed a regional stratification the next consideration is how the sample should be distributed between SIC groups within each region It has already been proposed that optimal allocation should be used for the SME element of the population Given that we have shown that there was little variation in response rates across industries in CIS3 it is proposed that the following 23 strata be used in the sample stratification for each region for the SME group Table 5 Proposed industrial stratification for CIS4 Population sic Sector Name size SME Large 10 14 Mining and quarrying 385 340 45 15 22 Mfr of food clothing wood paper publishing and printing 12 235 11 520 720 23 29 Mfr of fuels chemicals plastics metals amp minerals 17 305 16 515 790 30 33 Mfr of electrical and optical equipment 4 105 3 830 275 34 35 Mfr of transport equipment 1 685 1 445 240 36 37 Mfr not elsewhere classified 3 045 2 945 95 40 41 Electricity gas and water supply 105 75 30 45 Construction 16 800 16 455 350 50 51 Wholesale amp commission trade inc cars bikes and fuel 24 830 24 240 595 52 Retail trade exc cars amp bikes and repair 14 085 13 550 540 55 Hotels amp restaurants 19 310 18 960 350 60 63
44. n 7 stable over the last 15 years amongst the smaller firms whereas it seems to have declined among the larger firms Figure 3 Innovation activity by market characteristics 70 60 50 40 30 20 10 Local Domestic Europe International Other Public Consumers businesses sector E product innovation process innovation wider innovation Figure 4 Changes in Innovation Activity by industry Electrical and optical equipment Transport equipment Manufacturing not elsewhere specified Fuels chemicals plastic metal amp minerals Food clothing wood paper publishing amp printing Mining and quarry ing Mining and quarrying Construction Electricity gas amp water supply Financial intermediation Real estate renting amp business activities Wholesale trade including cars amp bikes Transport storage amp communication Retail trade excluding cars amp bikes Hotels amp restaurants 100 m CIS2 CIS3 w CIS4 Figure 5 Changes in Innovation Activity by size band 100 90 80 70 60 50 40 30 20 10 10 to 49 50 to 249 250 and more The above descriptive analysis of the CIS shows that innovation activity varies considerably across industry size and other characteristics of the firm and market However such analysis does not identify the individual contribution of each of these characteristics on innovation holding al
45. n related activities such as R amp D acquisition of equipment training design etc Information on innovations in business strategies and practices Product innovation Process innovation Abandoned and incomplete innovation activities The context for innovation e g increase range of goods or services entering new markets e Cooperation agreements e The factors constraining innovation Some core questions are only required every fours year rather than every two these are indicated in the Variable details Each business in the sample is sent the questionnaire and a leaflet containing some results from the previous survey The survey is voluntary however businesses receive two postal reminders and can be contacted by telephone to complete the questionnaire or to validate responses Following cogitative testing the 2009 survey was sent out to businesses at the end of March 2009 and remained active until September that same year Response and weighting Valid responses were received from 14 281 enterprises to give a response rate of around 50 per cent Accordingly weighting is used to compensate for the businesses that did not respond to the survey and those not selected for the sample Weighting allocates a weight to each business ensuring that the respondents are representative of the target population as a whole in terms of region division and business size Two weights are available to users 1 Business weight
46. n table 4 Each region is allocated the required sample size using the Neyman allocation Table 6 Sample design under scenario 1 Optimal Regional Allocation by region Optimal Regional Allocation by industry 10 49 50 249 250 Total 10 49 50 249 250 Total North East 1684 323 225 2232 10 14 63 55 45 163 North West 1302 406 705 2413 15 22 1138 495 720 2353 Yorks and Humber 1366 425 530 2321 23 29 1495 641 790 2926 East Midlands 1491 375 445 2311 30 33 489 206 275 970 West Midlands 1261 464 580 2305 34 35 241 134 240 615 East 1420 402 515 2337 36 37 509 130 95 734 London 1065 438 1190 2693 40 41 40 25 30 95 South East 1151 392 965 2508 45 1753 321 350 2424 South West 1658 398 445 2501 50 51 2366 574 595 3535 Wales 1271 504 240 2015 52 1565 202 540 2307 Scotland 1162 449 485 2096 55 2136 339 350 2825 Northern Ireland 1852 221 175 2248 60 63 998 231 380 1609 UK 16683 4797 6495 27975 64 1 73 50 25 148 64 2 67 45 55 167 65 67 509 126 345 980 70 346 98 140 584 71 180 61 60 301 72 347 126 145 618 73 1 60 50 60 170 73 2 40 0 0 40 74 2 358 88 110 556 74 3 64 45 15 124 restof 74 1846 755 1130 3731 Total 16683 4797 6495 27975 Pros Should provides better levels of regional precision than CIS3 and hence facilitates better regional comparison Gives regional analysts more comprehensive datasets to work with Cons Does not provide optimal results from
47. neern 31 1260 Innovation factors Reducing costs per unit produced or provided 31 1270 Innovation factors Improving health and safety cceeeeeeeeeeeeeeeeeeeeees 31 1212 Innovation factors Reducing environmental impacts cccceeeeeeeeees 32 1213 Innovation factors Replacing outdated products or processes 32 1280 Innovation factors Meeting regulatory requirements c cceeeeereeees 32 1601 Importance of information within your business or enterprise group 33 1620 Importance of information suppliers or equipment materials services or SONWANE i kiinni ioraa ces i a aaa ian a eiside eerie 33 1630 Importance of information clients Or CUStOMETS ccccccecccececeeceeeeeeeeeeeess 33 1640 Importance of information competitors or other businesses in your industry 34 1650 Importance of information consultants commercial labs or private R amp D IASUITUTES E E sade vaneysababanazacscdcuusspnbnan E E 34 1660 Importance of information universities or other higher education institutions EE E T E E A ee EE E E E E A A T 34 1670 Importance of information government or public research institutes 35 1680 Importance of information conferences trade fairs exhibitions 35 1610 Importance of information professional and industry associations 35 1611 Importance of informati
48. nnovation statistics Denotes additional sectors covered in the 2001 UK survey onwards Denotes additional sectors covered in the 2005 UK survey onwards d Denotes additional sector covered in the 2007 UK survey onwards 3 Business size All enterprises with 10 or more employees are included in the target population e Small 10 49 employees e Medium 50 249 employees e Large 250 or more employees Additionally to ensure representativeness the following conditions were also introduced e A census for all large firms 250 employees is taken e A census of SMEs in SIC 40 41 and 73 2 where the population is particularly small is taken and e Acap on the number sampled from SIC 50 51 52 55 and rest of 74 where the population is particularly large is taken The questionnaire The questionnaire content is determined by Eurostat regulatory requirements and BIS via the UKIS project board Eurostat provide the core harmonised questionnaire to ensure European Union data requirements are met and provide the basis for comparisons with other countries BIS are responsible for identifying the need for new questions or changes to existing questions so that the UKIS continues to provide a means to measuring the level types and trends in innovation activity in the UK and provide the empirical evidence to support policy The core questionnaire covers a broad range of innovation related concepts including e Details of any innovatio
49. o be made and we add this to the list given to us by the DTI e We checked for duplicates There are none Part 2 corrections from other consistency checks e Propsci propoth in some cases one is missing but the other is not In these cases according to DTI guidelines we replace missings with 0 e Ifthe variable prodinov product innovation is not missing i e equals 0 1 and prodnov new goods services new to the market is missing the latter is replace with zero Similarly for procinov process innovation e We replace missings in the share of turnover Sharenov due to products new to the market in 2000 with zero if prodnov equals 0 e Wider innovation we replace missings with zero following the DTI method when an answer for another question is present orgstrat orgmngt orgorgan orgmkt part 3 generate useful variables e generate RU reference as a string for compatibility with other datasets e generate total production and expenditures as a of turnover in 2000 part 4 generate inconsistency markers CIS3_clean do creates indicator variables for observations that present inconsistencies e Firstly one of the main concerns is that reporting units are not reporting at the reporting unit level i e they might be reporting at the local unit or at the enterprise level This concern is caused by major differences between turnover and employment figures both in 1998 and 2000 reported by the reporting units versus the figures fro
50. on technical industry or service standards 36 1690 Importance of information scientific journals and trade technical publications EEE PE T TE S O E ecaivaWeumwnundd E EEEE A OE T AAT E 36 1811 1814 Innovation cooperation Other businesses within your enterprise group 1821 1824 Innovation cooperation Suppliers of equipment etc ee 37 1831 1834 Innovation cooperation Clients Or CUStOMETS ccccccccceccceeeeeeeeeeees 37 1841 1844 Innovation cooperation Competitors or other businesses 37 1851 1854 Innovation cooperation Consultants etc ccceceeeeeseeeeeeeeeteneeeeeees 38 1861 1864 Innovation cooperation Universities or higher education 38 1871 1874 Innovation cooperation Government or public research institutes 38 2011 Reasons for not innovating no need due to previous innovations 39 2020 Reasons for not innovating no need due to market conditions 39 2030 Reasons for not innovating other factors constraining innovation 39 1901 Innovation constraints Excessive perceived economic risks 0 008 40 1902 Innovation constraints Direct innovation costs too high cceeee 40 1903 Innovation constraints Cost of finance 20 2 2 2 cece eeececeeeeeeeeeeeeeeeeeeteeenneeeeeees 40 1904 Innovation constraints Availability of finance
51. or significantly improved processes cccceeceeeeeeeeeeeeeeeeeeesneneeeeees 26 1010 Processes developed by business enterprise cccccccccccccceeeceeeeeeeeeeees 27 1020 Processes developed by business with another buSiness cccceeee 27 1030 Processes developed by other businesses ccccccccccccceccceceeeeceeeeeeeeeeeess 27 1100 New to industry proCESSeS iisen iaae isenana nn kE rana aS 28 1510 Abandoned innovation activities ccccccceeceeececeeceeeeeeeeteeceeeeeeeseteeneeees 28 1520 Incomplete innovation activities sisirin ninani aE a 28 1501 Innovator MAKET vas sc cetceceteteskee deduce ceeeteecharbeseRecece te euds cack ok edeeaatetsietaeladeneded 28 Section D Context for Innovation ssssssssssssssssnsnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn nnmnnn 29 1210 Innovation factors Increasing range of goods or services seseee 29 1211 Innovation factors Entering new markets cccceceeeeeeeeeceeeeeesseeneeeeees 29 1220 Innovation factors Increasing market Share ccccceeeeseceeeeeeeeteneeeeeees 29 1230 Innovation factors Increasing quality of goods or services 0 cee 30 1240 Innovation factors Improving flexibility for producing goods or services 30 1250 Innovation factors Improving capacity for producing goods or services 30 1290 Innovation factors Increasing value added
52. reason to pick one or other of the duplicates They are therefore left in the dataset but should be removed or reduced before merging on RU references Overall these surveys require users to use them cautiously A detailed study of the survey forms and documentation is recommended before starting analysis UK Innovation Survey 2009 User Guide Contents IMtFOGDUGCHION sicisscssedeeviei ciel isistid ensai cin nineties ised eect 5 OVO NVI OW A E E A E A E A A 5 Coverage and Samphhgssreisnao e a ERA a aa oana 6 The QUESUIOMMAING sarson a aa aE 7 Response and weighting cviciccsiscco ccsccsttiesceceeescseceeetsiedeeessseescivesdadagessessesvesedadenddivensnst aa 8 Variable COING 2 iiteiccsisiectndiisceidnis edna wasn iid nite nia edininienedt 8 N te on MISSING VAlUGS i siirsin eena aiaa a i a OnE 8 Innovation COMCEPtS siiis iaaa aara a aaraa aa a aaia 9 Question Variable look up table cccceceseeeeceeeeeeeeeeeeeceeeeeeeeeeeeaeeeeeeeeeneseeeseseeeeeneees 10 Details of Variables sicion iaaiaee daaarna 11 D mographie a a 11 REFERENCE IDBR reference NUMDEL ccccceeeeeeeceeeeeceeeeeeeceeeeeeeeeeeeteneeaeeees 11 ENTERPRISE IDBR enterprise NUMDET c cccceccececeeeeeeeeeeeeceeeeeeeeeeeensneeeeeees 11 IC92 5 digit SIC 2003 classification ss risie rieni iinei tekin kine ii 11 INT_FOC country of immediate ownerShip 2 cccccceeeeeeeeeeeeeeeeeeeeeeeeeeeeeeetees 11 ULT_FOC count
53. reater the size of the firm the larger the weight it carries in the employment weighted results Employment weighted data provide a more robust means of estimating whole economy performance Given large firms dominance of economic activity in the UK the 6 500 large firms in the CIS4 population account for over 60 of total employment and around 70 of turnover in the population and the arguments for using employment weights to provide better measures of whole economy performance a census of large firms is recommended For the SME strata 10 49 and 50 250 sampling should be carried out using the Neyman allocation Recommendation Take a census for large firms 250 employees and sample based on optimal allocation for SMEs Stratification and issues of sample design Regional stratification A DTI paper Assessing the accuracy of published results of the UK CIS3 revealed that there are large disparities in the precision of the published CIS3 results for the 12 regions see chart 1 Chart 1 Innovation Activity by Region with 95 Confidence Interval UK Total W Mid S East N East EMid Wales I S West London Eastern 0 10 20 30 40 50 60 Per Cent Given that the results of CIS4 are likely to be used for regional benchmarking one conclusion of the paper was that a greater level of regional balance in the sample would be desirable to allow for more precise regional comparisons A sli
54. ry of ultimate Ownership ccccccceeseceeeeeeeeeeeneeeeeeeeseeeenseeeeees 11 REGION ONS TeGlOMS is scsissecacnaneshsannanctccastadaaneinsdeneecansiancctneedeenascccaaqenebatascereenanenaa 11 GOR Government office reQlOM iggescvscacactericceedenssntatanledataigas a nea aiaia e iekea 12 POSTCODE reference unit postcode first 4 characters Only ccceeeeeeeeeeees 12 EMPLOYMENT number of employees in enterprise c ccccceeseeceeeeteeteneeeeeees 12 SIZEBAND group number of employees in enterprise ccccccceeeceeeeeeseeeteeeeees 12 TURNOVER DBR nmiOWen nisiodo enne Ea aa EAT G 12 Section A Innovation activity ansaasenneeensnnnnnnnnnnnnnnnnnnnnnnnnnnnnnunnnnnnnnnnnnnnn nnna 12 210 Geographic market UK regional ccccccceeceeeeeeeeeeeeeeencneeceeeeeesensneaeeees 12 220 Geographic market UK national c cccceeeceeceeeeeeeeeeseeneneeeeeeeeessnsneeeeees 13 230 Geographic market European countries cccccecceeeeeeeeeeceeeeeeesneneeeeeees 13 240 Geographic market All other COUNTIICS 2 ceeceeceeceeeeeeeeeeeeeeeeeeteeteneeeeeees 13 410 Business changes Business was established c cccceeseceeeeeeeeeeeeeeeeees 13 420 Business changes Turnover increased due to Mergel ccceeeeeeereeeees 13 430 Business changes Turnover decreased due tO MEPGel cccccccceeeeeeeeeeees 14 440
55. s these are frequency weights that indicate the number of enterprises a respondent represents within their strata On average each respondent represents 13 enterprises in the population 2 Employment weights these are frequency weights that indicate the number of enterprises a respondent represents according the number of employees in their business and the total number employees within their strata taken from the IDBR Variable coding Details of the coding used for each variable are set out below under the Details of variables Typically variable labels adhere to the convention 1 yes 0 no and 3 High 2 Medium 1 Low Unless otherwise specified in this guide the coding for not answered and not applicable are 8 and 9 respectively Note on missing values The routing used throughout the questionnaire means that businesses are not required to answer every question some questions are not applicable In these instances where a question is not applicable responses for that particular business are coded as 9 and defined as missing i e their values will not be included in any analysis There are however occasions where a question is appropriate to the business but a certain category or response contained in the question may not be In this instance their response is considered as valid Like with many surveys there are occasions when a respondent does not or
56. s influence of organisational structure and age 60 50 40 30 20 Relative Odds Standalone Part of EG O 4years 5 9years 10 14 15 20 20 years years years B Product Process Wider Figure 7 shows that being part of an enterprise group significantly increases the probability that a firm states that it engages in innovation particularly wider innovation Those enterprises who are part of an enterprise group are approximately 30 more likely to engage in product and process innovation and 55 more likely to engage in wider innovation Note that this effect is separate and additional to other characteristics of these enterprises Finally it is estimated that the youngest enterprises are most likely to state that they engage in innovative activity Relative to those enterprises that have been established for 20 years or longer those enterprises who have been established for 0 4 years are estimated to be 50 more likely to engage in product innovation 27 more likely to engage in process innovation and 40 more likely to engage in wider innovation Bibliography Cereda et al 2005 Design and company performance Evidence from the Community Innovation Survey Report for DTI Crespi Criscuolo and Haskel 2007 Information Technology Organisational Change and Productivity CEPR Working Paper Crespi Criscuolo and Haskel 2007 Productivity Exporting and the Learning by Exporting Hypothesis
57. sing the CIS is shown below Innovative Business and the Science and Technology Base Swann 2002 This analysis uses CIS3 to assess the role of the university and other research institutions as a source of information and cooperation for innovative businesses and the effect of such cooperation on business performance It finds that companies are more likely to cooperate with universities when they are process innovators but less likely when they are product innovators The results suggest that universities play a relatively more important role through cooperation than as a source of information Cooperation with the university is especially effective in achieving better process performance i e greater production flexibility reduced unit labour costs and increased capacity Design and company performance Evidence from the Community Innovation Survey Cereda et al 2005 This research undertakes an analysis of the relation between design inputs and other innovation and economic performance indicators The authors find that around 9 of firms reported some spending on design and that design spending represented around 10 of all reported spending on innovative activities They also find that design has a positive and statistically significant association with product innovation but not process innovation They estimate a marginal return to design expenditure of about 17 which they state is likely to be an overestimate of the causal effect R
58. sssitaciecer i eat ieee ernie i 48 WEIGHT Business frequency Weights s o s cccce ce enai cat easeeeen deere inken naana 48 eWEIGHT Employment frequency weights ccccccceeceeceeeeeeeneneeeeeeeesesnsneeeeees 48 Introduction The UK Innovation Survey UKIS provides the main source of information on business innovation in the UK The survey data is a major resource for research into the nature and functioning of the innovation system and for policy formation It is used widely across government regions and by the research community The survey is funded and developed by the Department of Business Innovation and Skills BIS and administered by the Office for National Statistics ONS with assistance from the Northern Ireland Department of Enterprise Trade and Investment DETI The UK Innovation Survey also represents the UK s contribution to the Europe wide Community Innovation Survey CIS This user guide is based on the sixth iteration of the survey UKIS 2009 sometimes referred to as CIS6 or CIS 2008 Overview The purpose of the survey is to collect information about businesses innovation in the UK Like many innovation surveys across Europe the UK innovation Survey follows general guidelines set out in an OECD publication known as the Oslo manual OECD 2005 This manual provides guidelines on the conduct of innovation surveys including statistical procedures and a review of the range of concepts that fall
59. ter two Wider innovation also seems to be more prevalent relative to product and process innovation in the distribution and service sectors Figure 1 Innovation activity by industry Mining and quarrying A Food clothing wood paper publishing amp printing Fuels chemicals plastic metal amp minerals Electrical and optical equipment Transport equipment Manufacturing not elsewhere specified Electricity gas amp water supply Construction Wholesale trade including cars amp bikes Retail trade excluding cars amp bikes Hotels amp restaurants Transport storage amp communication Financial intermediation Real estate renting amp business activities 0 10 20 30 40 50 60 wider innovation H process innovation product innovation Innovation is divided into product process and wider innovation and the markers for these are available in the CIS4 data Product innovation occurs when a firm introduces a new or significantly improved good or service Process innovation occurs when a firm introduces new or significantly improved processes for producing or supplying goods or services and these processes are new to the enterprise Wider innovation occurs when an enterprise makes major changes in business structure and practices including corporate strategy advanced management techniques organisational struc
60. th technician staff ather than dedicated R amp D staff hence they are not doing R amp D continuously e Engineers in workforce e Use pc_qse as given in the survey that lies between 0 and 1 Checked with no_qse emp_96 and it is fine e Factors influencing INNOVATIVE ACTIVITY e Sometines there are some holes in the data i e some anwers are ticked others in the same observations are missings we replace them with zero e When they are all missings there is not much I can do I leave them out e Problem if both filter questions are no the firms should have not answered these questions but they have We create a dummy called filter for these observations and see what is the role they play in the analysis e SOURCES of INFORMATION for innovation projects problems as before e Sometines there are some holes in the data i e some anwers are ticked others in the same observations are missings we replace them with zero e When they are all missings there is not much I can do e Problem if both filter questions are no the firms should have not answered these questions but they have We create a dummy called filter for these observations and see what is the role they play in the analysis e SOURCES of TECHNOLGICAL KNOWLEDGE e Replace 0 when it is missing if at least one of the other is not missing e COOPERATION e In few cases cooperation 0 amp partner of cooperation is not zero we construct a marker for these cases weir
61. the point of view of UK and industry level estimates Larger relative burden on firms in smaller regions e g North East 40 of firms sampled compared to London 11 Scenario 2 Optimal National Allocation This scenario applies the same methodology as was used in CIS3 namely to apply the Neyman allocation to the whole UK dataset and select the sample to provide optimal national level estimates Table 7 Sample design under scenario 2 Optimal National Allocation by region Optimal National Allocation by industry 10 49 50 249 250 Total 10 49 50 249 250 Total North East 807 167 225 1199 10 14 65 55 45 165 North West 1616 498 705 2819 15 22 1140 489 720 2349 Yorks and Humber 1354 420 530 2304 23 29 1434 621 790 2845 East Midlands 1383 350 445 2178 30 33 507 210 275 992 West Midlands 1422 520 580 2522 34 35 235 126 240 601 East 1556 436 515 2507 36 37 510 124 95 729 London 1778 715 1190 3683 40 41 40 25 30 95 South East 1925 636 965 3526 45 1714 332 350 2396 South West 1639 394 445 2478 50 51 2334 631 595 3560 Wales 779 332 240 1351 52 1445 211 540 2196 Scotland 1014 397 485 1896 55 1965 364 350 2679 Northern Ireland 1184 167 175 1526 60 63 1004 241 380 1625 UK 16457 5032 6495 27984 64 1 78 50 25 153 64 2 80 49 55 184 65 67 548 163 345 1056 70 365 111 140 616 71 176 64 60 300 72 390 150 145 685 73 1 61 51 60 172 73 2 40 0 0 40 74
62. tive to a reference category Innovation activities are once again divided into product process and wider innovation The coloured bars are used to indicate where a variable is estimated to be significantly different from the reference category at the 5 level Figure 6 confirms the earlier finding that larger firms are more likely to innovate particularly in process and wider innovation Amongst those organisations with less than 15 employees the analysis distinguishes between those organisations where at least employee has a degree and those who do not As noted above we restrict this distinction to the smallest enterprises where it is more likely that the person filling out the survey will know if any of the employees has a degree Among those enterprises with less than 15 employees those enterprises with staff who have a science degree are approximately 200 more likely or 3 times as likely to engage in innovative activity The presence of staff with other degree subjects also increases the likelihood of innovative activity within small enterprises Figure 6 Probability of undertaking innovation activities influence of firm size and education Wicadddl lt 15 no lt IS other 15 24 25 49 50 99 100 249 250 499 500 degree science degree degree 450 400 350 300 250 200 150 100 50 Relative Odds E Product W Process H Wider Figure 7 Probability of undertaking innovation activitie
63. together under the umbrella term innovation The survey is based on a core questionnaire developed by Eurostat and Member States and covers a broad range of policy interests including General business information Innovation activity Goods services and process innovation Context for innovation and General economic information The UK survey was originally conducted every four years but since 2005 has been conducted biennially Previous Community Innovation surveys took place in 2007 2005 2001 In the UK the survey is voluntary and collected by means of a postal questionnaire 4 ittp www oecd org document 33 0 3343 en 2649 34273 35595607 1 1 1 37417 00 html http epp eurostat ec europa eu statistics explained index php Community innovation survey Coverage and Sampling The UK innovation survey consists of a nationally representative sample of businesses with 10 or more employees in sections C K of the Standard Industrial Classification CIS 2003 The sample is a stratified design drawn from the Inter Departmental Business Register IDBR with Neyman allocation used to determine the sample size in each stratum Overall roughly ten per cent of the target population in sampled Stratification was based on three variables 1 Region All regions and countries in the UK 9 Government Office Regions in England plus Scotland Wales and Northern Ireland are covered North East North West Yorkshire and The Hu
64. ture and marketing Figure 2 shows that standalone enterprises are less likely to innovate than enterprises which are part of an enterprise group This may partly reflect the fact that standalone enterprises are on average smaller mean employment of 129 compared with 546 for enterprises that are part of a group but it may also capture innovation spillovers between enterprises within an enterprise group For all types of innovation larger firms are more likely to innovate than smaller ones Finally innovation activity is found to be more prevalent among firms where at least one employee has a science or engineering degree compared to other degrees and compared to enterprises where no employee has a degree Figure 2 Innovation activity by enterprise characteristics 70 60 50 40 30 20 Stand Part of lt 15 15 24 25 49 50 99 100 249 250 499 500 No degree Other Science alone EG and engin W product innovation process innovation wider innovation Figure 3 shows that firms engaging in export activities are far more likely to innovate in all areas with a higher proportion of innovators among those firms who export beyond Europe Innovation activity is most prevalent among firms whose main customers are other businesses followed by those doing business with the public sector and finally consumers Comparing innovative activity between
65. uk e CIS 4 covers the period 2002 2004 It is the largest of the innovations surveys conducted so far sent to some 28 000 UK enterprises Of those 16 445 enterprises provided valid responses representing a response rate of 58 The sample of enterprises is drawn from the ONS Inter Departmental Business Register IDBR and is based upon those firms with more than 10 employees The sample is designed to be statistically representative of the 12 regions of the UK most industrial sectors and all sizes of firms The responses to the survey are weighed back to the population using the inverse sampling proportion in each stratum On average each respondent in CIS4 represents 11 enterprises in the population The CIS is a voluntary postal survey To boost response enterprises are sent the survey posted a reminder posted a second reminder with the survey again and finally telephoned There are a number of concerns about the data primarily as the result of inconsistent responses provided by survey respondents A number of marker variables were created in the cleaning process to identify problematic cases 2 Linking CIS to ONS Business Data The IDBR is the key sampling frame for business surveys within ONS Enterprises appearing within ONS surveys are assigned a unique IDBR reference number which can facilitate linking of information on the same enterprise over time and between surveys The reporting unit identifier in the CIS is given by ru_ref
66. ure on different kinds of innovative activity effects of innovation sources of information and co operation barriers to innovation protection methods for innovation and public support for innovation Within the CIS innovation is defined as major changes aimed at enhancing a firm s competitive position performance know how or capabilities for future enhancements These can be new or significantly improved goods services or processes for making or providing them Expenditure on innovative activities includes machinery and equipment R amp D training goods and service design or marketing The CIS is carried out at the level of the enterprise As such an enterprise may carry out one or more activities at one or more locations There have been four CIS surveys conducted to date each covering a three year period e CIS 1 covers the period 1991 1993 Due to the poor response rate 10 this survey is regarded as being of poor quality and is not available within the VML e CIS 2 covers the period 1994 1996 In total 5 416 enterprises were surveyed of which 2 339 responded to the survey achieving a response rate of 43 e CIS 3 covers the period 1998 2000 and was conducted in two waves The first wave sampled 13 340 enterprises Of the 19 625 enterprises to which the survey was sent 8 172 responded achieving a response rate of 42 For information on the CIS or other business data sets held within the VML please email vml ons gov
67. ustrial and national levels Regional analysts will have substantial sets of data to enable local analysis and the burden in the smaller regions and certain small firms will be reduced Recommendation UK CIS4 sample to be based on design described in scenario 4 DTI IG TESE Jan 2005 12 Technical details of the UK Innovation Survey 2005 CIS4 Methodology The UK Innovation Survey is funded by the Department of Trade and Industry DTI The survey was conducted on behalf of the DTI by the Office for National Statistics ONS with assistance from the Northern Ireland Department of Enterprise Trade and Investment DETINI The UK Innovation Survey is part of a wider Community Innovation Survey CIS covering European countries The survey is based on a core questionnaire developed by the European Commission EuroStat and Member States There have now been four innovation surveys more detail can be found from the following link www cordis lu innovation smes scoreboard home htm The UK Innovation Survey 2005 CIS4 sampled over 28 thousand UK enterprises The survey was voluntary and conducted by means of a postal questionnaire Coverage and sampling The survey covered enterprises with 10 or more employees in sections C K of the Standard Industrial Classification SIC 2003 The 2005 survey included the following sectors Sale maintenance amp repair of motor vehicles SIC 50 Retail Trade SIC 52 and Hotels amp restaurants
68. vation related expenditure e Formally co operated with other enterprises or institutions on innovation Sampling frame The Second Community Innovation Survey CIS 2 covered the period 1996 98 There is little information available about this survey now CIS 3 was in the field twice The first wave sampled 13 340 enterprises the second top up covered 6 285 to make the sample representative at the regional level The CIS 3 covers the period 1998 2000 Of the total 19 625 enterprises to which the survey was sent 8 172 responded Table 1 row 1 achieving a response rate of 42 The 2005 survey CIS 4 is the largest innovation survey so far conducted sent to 28 000 UK enterprises and achieving a 58 response rate The latest CIS 2007 survey contains approximately 14 000 observations Note this is a mini in between survey The next full survey of the CIS is the CIS 5 CIS2007 received a higher response rate than any of the previous surveys Importantly approximately half of the respondents to this latest survey had also previously responded to the CIS4 DIUS have kindly provided the VML with a CIS4 CIS2007 panel data set This file contains the observations for approximately 7000 respondents who completed both surveys Organisation of files The data is received from the DTI substantially cleaned and made available as the files cis2_clean_O1 dta cis3_clean_Ol dta and cis4 dta These are held within the clean drive U CIS For
69. was agreed at the CIS4 project board meeting of 2 July 2004 that the following additional sectors would be covered in the UK survey SIC 45 Construction SIC50 Sale maintenance and repair of motor vehicles SIC52 Retail trade SIC 55 Hotels and restaurants SIC70 Real estate SIC 71 Rental of machinery and equipment SIC73 R amp D Remainder of SIC 74 i e excl 74 2 and 74 3 Other Business Activities denotes sectors that were covered in CIS3 in the UK The detailed stratification is considered towards the end of this paper Region country coverage As in CIS3 all regions and countries in the UK 9 Government Office Regions in England plus Scotland Wales and Northern Ireland will be covered and these 12 areas used in the regional dimension of the stratification Sizebands All enterprises with 10 or more employees will be included in the target population and the following strata used 10 49 50 249 250 499 500 999 and 1000 Statistical units The sample will be drawn at the level of the reporting unit in line with other major relevant business surveys e g BERD ABI Some users have noted that by sampling at the reporting unit level we often lose sight of diffusion of innovation within the company at the local unit level and that this might lead to cross border innovation not being picked up Therefore a file detailing all local units within the selected reporting units will be extr
70. xtra xextram rdpers rdcont rdoccas For the observations that present inconsistencies set outrd 1 where e R amp D expenditure 0 even if no intramural R amp D R amp D expenditure e no R amp D but R amp D personnel e no internal or external R amp D or personnel but R amp D is being done continuously or occasionally e Create marker for R amp D done continuously or occasionally Part 5 some further inconsistency checks done e Prodnew prodimp produnc we check that the sum of the three is 100 by creating a new variable prodtot the sum of the three Results were consistent with almost all summing to 100 except a few cases where it was 99 e How are the new products processes developed prodwho and procwho in 15 cases prodinov equals 1 and prodwho is missing in 9 cases procinov equals and procwho equals missing We have not flagged these observations but it is worth keeping this into account for the analysis e Innovation activities not completed or abandoned Aband nyettime nyetlate noteven no case in which one part of the question had a response and another was unanswered In 4 710 cases this information is missing e Factors hampering innovation missing in 925 cases e support there are various inconsistencies but perhaps as noted by DTI is a problem with the survey itself and it is not possible to correct for those After the cleaning the dataset created is ready to be matched at the reporting unit level with any other datas

UK Innovation Survey - User Guide, 1st edition

Contents

Download Pdf Manuals

Related Search

Related Contents