Received: 21 January 2021 | Accepted: 27 January 2022 DOI: 10.1111/ijsw.12528 ORIGINAL ARTICLE Using a hybrid data collection tool: Analysis of youth labour market trajectories integrating quantitative, qualitative and social network data Joan M. Verd Department of Sociology, Centre d'Estudis Sociològics sobre la Vida Abstract Quotidiana i el Treball (QUIT), Institut This article explains the factors that led to the use of a hybrid data collection d'Estudis del Treball (IET), Universitat tool—­i.e., a single questionnaire designed to collect both quantitative and quali- Autònoma de Barcelona, Barcelona, Spain tative data—­ in two different research projects that focused on the relation between the labour market trajectories of young people and the use and accu- Correspondence Joan M. Verd, Department of Sociology, mulation of network social capital in these trajectories. The article describes the Centre d'Estudis Sociològics sobre la characteristics of this tool and the methodological challenges posed by its use. Vida Quotidiana i el Treball (QUIT), The advantages and disadvantages of this specific type of data collection are also Institut d'Estudis del Treball (IET), Universitat Autònoma de Barcelona, addressed. Among the main advantages, the improvement of data quality and the Barcelona, Spain. overarching picture obtained of the career path and use of personal networks are Email: joanmiquel.verd@uab.cat highlighted. The most important downside found is the amount of time needed Funding information in the interview, which shows the need to pay attention to the balance between The research projects mentioned quantitative and qualitative elements in the structure of the questionnaire. in the article have been funded by the Spanish Ministry of Economy KEYWORDS and Competitiveness with contracts hybrid methods, job seeking, labour market trajectories, mixed methods data collection, CSO2012-­36055 (REDEMAS project) and CSO2016-­77905 (CAPSINES personal networks, sequential data, Spain, youth project) I N T RO DU CT ION the use and accumulation (or loss) of network social capital in these trajectories. The methodological issue addressed in this The concurrent production of quantitative, qualitative and article is twofold: (1) the rationale for building and using one network data by means of a single “hybrid” data collection single tool to collect different types of data, and (2) the advan- tool (a mixed data collection tool in which qualitative and tages and disadvantages of using this type of tool in regard to quantitative characteristics are combined) and the challenges the fieldwork and characteristics of the data obtained. The ar- posed by this methodological decision constitute the focus ticle argues that the use of a hybrid data collection tool is use- of this article. More specifically, the article presents and dis- ful to produce quality data to address complex, multifaceted cusses the use of a hybrid data collection tool implemented and interrelated processes, such as those mentioned above. in two different research projects examining the relationship However, the design and use of this kind of tool involves cer- between the labour market trajectories of young people and tain complexities and difficulties, which are also described and Abbreviations: REDEMAS, Social networks as a mechanism for job search and labour market insertion in young people; CAPSINES, The accumulation and use of social capital among young people with unstable labour market trajectories. This is an open access article under the terms of the Creative Commons Attribution-­NonCommercial-­NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-­commercial and no modifications or adaptations are made. © 2022 The Authors. International Journal of Social Welfare published by Akademikerförbundet SSR (ASSR) and John Wiley & Sons Ltd. Int J Soc Welfare. 2022;00:1–11.  wileyonlinelibrary.com/journal/ijsw | 1 2 |    VERD discussed. Although this article presents the characteristics of areas of the Chitwan Valley in Nepal. They adapted the life the data produced and the findings, it does not elaborate on history calendar (Freedman et al., 1988) to their research how the quantitative and qualitative data simultaneously ob- question and the population they wanted to study. What tained were used during the data analysis stage. The article was initially a survey questionnaire with a timeline to an- focuses on the data collection stage and does not discuss data chor memories—­for example, by placing events that were integration issues, which have been widely discussed in many collectively remembered, such as national elections, floods well-­known mixed methods manuals (e.g., Creswell & Plano or the resignation of the king—­became a new, single data Clark, 2018; Morgan, 2014; Morse & Niehaus, 2009) and in collection tool with characteristics of both structured meth- empirical research (e.g., Harrits & Møller, 2021; Kerrigan, ods, such as surveys, and less structured methods, such as 2014). Thus, the article concentrates on a topic not sufficiently unstructured interviewing (Axinn & Pearce, 2006, pp. 106–­ discussed in the mixed methods literature and helps readers to 112). This new hybrid questionnaire was often answered avoid in their future research some of the difficulties derived collectively, in the frame of an open conversation. from the use of a hybrid data collection tool. For their part, Gobo and Mauceri’s (2014) proposal As shown by Creswell and Plano-­Clark (2011; 2018), little develops the basic idea of the conversational interview has been written about mixed methods data collection proce- (Schober & Conrad, 1997) to propose a new type of sur- dures, with the exception of discussions on sampling strategies. vey, which they call inter-­vey, in which both qualitative The mixed methods methodology literature has concentrated and quantitative elements are merged in a single instru- on the combination of quantitative and qualitative compo- ment. As in a conversational interview, one of the objec- nents in the stages of research design and data analysis and tives of an inter-­vey is to clarify the concepts mentioned usually assumes that quantitative and qualitative data are in the questions, but this clarification is framed as a ver- collected separately by using two different tools, one quanti- bal interaction that is similar to a chat between an inter- tative and the other qualitative (see, for instance, Creswell & viewer and interviewee in qualitative interviews (Gobo & Plano Clark, 2018; Morgan, 2014; Morse & Niehaus, 2009). Mauceri, 2014). Gobo classifies this type of interview as However, the collection of quantitative and qualitative data what he more generically calls merged methods, highlight- using one single tool (and a single data source) is also possible ing that they combine “both qualitative and quantitative (Axinn & Pearce, 2006; Creswell & Plano Clark, 2018). Axinn approaches in a single instrument, squeezing the advan- and Pearce (2006) use the term hybrid to label a data collec- tages of both in a single technique” (Gobo, 2015, p. 331). tion procedure that produces both quantitative and qualitative In the following sections, different aspects of the building data, while Cresswell and Plano-­Clark (2018) prefer using the and use of a hybrid data collection tool applied to the study term “one form data collection”, in contrast to “different form of the relationship between youth labour market trajectories data collection” when different quantitative and qualitative and social capital are presented and discussed. First, the char- methods are used (Creswell & Plano Clark, 2018, p. 189). For acteristics of the two research projects where the hybrid tool the sake of simplicity, the term hybrid is used in this article. was used and the characteristics of the tool are discussed, be- This adjective is not unknown in the mixed methods litera- ginning with a short theoretical section presenting the com- ture, as it has also been used to describe data analysis methods plexity and multiple directions of causality in the relationship that combine both quantitative and qualitative characteristics between labour market trajectories and social capital. Second, in a single procedure (Bazeley, 2010, 2018; Fielding, 2012). the advantages and downsides of the use of this type of hybrid The collection of different types of data using the same method for research are evaluated and discussed. Finally, in tool is usually carried out in research conducted through the concluding section, the findings and knowledge resulting surveys (Creswell & Plano Clark, 2018; Morgan, 2014): from the use of this hybrid tool are summarised, and a gen- qualitative data obtained from open-­ ended questions eral reflection on the use of this kind of method is presented. within a survey are used for further exploration of the quantitative data or quantitised to add further details to the closed-­ended questions of a survey. However, some STUDYING YOUTH LABOUR authors have gone beyond the collection of qualitative MARKET TRAJECTORIES AND information embedded in survey questionnaires and have SOCIAL CAPITAL USING A HY B RID explicitly conceived hybrid tools to obtain qualitative data DATA COLLECTION TOOL that are not just complementary to quantitative survey data. This is the case for researchers such as Axinn and The relationship between youth labour Pearce (2006) and Gobo and Mauceri (2014). market trajectories and social capital Axinn and Pearce's (2006) methodological design is part of their research focusing on changes in resources and insti- The literature has highlighted the process of feedback tutions that occurred over 42 years (1954–­1996) in various and reinforcement of the initial unequal endowments of USING A HYBRID DATA COLLECTION TOOL     | 3 social capital in labour market integration (Lin, 2000; The projects carried out Moerbeek & Flap, 2008): a high endowment of social capital facilitates labour market integration and ca- According to Creswell and Plano-­Clark (2018, p. 189), the reer stability, which eventually provides access to bet- use of a single data collection tool “may be preferred if ter information on job vacancies and opportunities for the researcher's interest is to capture multiple facets of a professional achievement; a reverse process occurs for phenomenon from each participant”. This was, indeed, people with a low endowment of social capital. These one of our interests. The objective of obtaining the above unequal endowments of social capital are linked to mentioned three types of data—­quantitative career data, individual characteristics such as gender, age and so- quantitative network data and qualitative biographical-­ cial background, as access to resources embedded in narrative data—­in an integrated and articulated way was personal networks is strongly associated with the web an essential aspect of the two research projects described of (unequal) social relations individuals are a part of next. The reasons for building and using a hybrid data col- (Granovetter, 1985). The present-­ day context, where lection tool, the fieldwork carried out, the data obtained in young people face non-­standard and particularly diffi- these projects and the findings are addressed in the subse- cult labour market transitions, adds extra layers of com- quent sections. plexity to this relationship. The uncertainty of youth Table 1 presents the objectives and characteristics of career paths and the multifaceted relationship between the fieldwork for the two projects1 in which a very similar variables, marked by the time dimension, shows the hybrid data collection tool was used. The main objective importance of using sequential data for the analysis of of the first project was to find out to what extent personal the link between labour market positions and network contacts that offered useful support along the labour mar- social capital. Both from an aggregate (macro) point of ket trajectory overlapped or not with those who offered view and from an individual (micro) point of view, a other types of support (economic, expressive, emotional, cross-­sectional snapshot of the labour market situation etc.). The age range of the population considered was 20 to and the personal network of contacts at a certain point 34 years. In the second project, the main objective was to in time may feature a different picture if taken a few understand the extent to which spells of precarious em- years before or a few years later. ployment or unemployment along the labour market tra- The existing literature has already addressed the jectory influenced the characteristics of network social time dimension in the analysis of the role of personal capital and the way this social capital was used in subse- networks in the labour market integration and advance- quent transitions in the labour market. In this project, the ment of young people (e.g., Ryan et al., 2019; Verhaeghe population under study was young people between 29 and et al., 2015); but this literature, to the best of our knowl- 34 years of age whose working trajectory had periods of edge, has not addressed sequentially the two-­way re- contractual instability or unemployment. lationship between the composition and structure of In both projects, the complexity and the multiple direc- young people's personal networks (contacts among their tions of causality required a comprehensive and holistic personal contacts) and their trajectory in the labour perspective and a careful and detailed data collection pro- market. This is because existing sequential databases, as cedure, very much in the style of the qualitative biograph- well as surveys developed in specific research projects, ical approach (Wengraf, 2001). Thus, the use of a hybrid are usually limited to the characteristics of personal con- data collection tool provided the wealth of information tacts and do not consider the structure of connections offered by a qualitative approach while at the same time among these contacts, which requires much more time offering the possibility of producing different standard to obtain. A third type of information could be added to quantitative data matrices, useful, for example, to apply the previous ones in order to obtain the most complete sequence analysis methods. The use of a single tool com- and comprehensive view possible of the relationship be- bining quantitative and qualitative characteristics was tween labour market trajectories and the accumulation risky, as it involved methodological difficulties, but after of network social capital: biographical-­narrative data. weighing the pros and cons, this methodological route This kind of data is useful for understanding individ- was explored. The characteristics of the hybrid tool and ual decision making in trajectories (Hollstein, 2019) and identifying those events and transitions that constitute 1 key moments of biographical rupture (Verd & López, These projects were called REDEMAS (Social networks as a mechanism for job search and labour market insertion in young people) 2011). For example, it is at such turning points that the and CAPSINES (The accumulation and use of social capital among (potentially crucial) role played by specific personal or young people with unstable labour market trajectories). The fieldwork institutional support in the redirection of a career path for both projects was carried out in the metropolitan area of Barcelona, can be best assessed. Spain. 4 |    VERD TABLE 1 Projects where a hybrid data collection tool was used. Objectives and characteristics of the fieldwork Project 1 (REDEMAS) Project 2 (CAPSINES) Objectives Finding out the extent to which personal contacts Finding out the extent to which spells of who offered useful support in finding employment precarious employment or unemployment along the labour market trajectory overlapped along the labour market trajectory with those who offered other types of support influenced the characteristics of social (economic, expressive, emotional, etc.) capital and the way this social capital was used in subsequent transitions in the labour market Characteristics of the Young people between 20 and 34 years of age residing Young people between 29 and 34 years of age population considered in Barcelona metropolitan area who had at least residing in Barcelona metropolitan area who one employment experience had been unemployed for at least six months or with a temporary contract for at least 12 months Sampling strategy Quota sampling Quota sampling Sample size 250 150 Data collection tool Hybrid questionnaire Hybrid questionnaire Year of fieldwork 2014 2018 the way it was built are described below, while the pros tool with both quantitative and qualitative characteris- and cons of the tool are discussed further down. tics, in line with the proposal of Axinn and Pearce (2006). These changes also involved the reordering of the blocks and questions, as presented in Table 2 for the two projects. The characteristics of the hybrid tool It should be noted that the final configuration of this new tool was the result of multiple pilot tests carried out during In both projects, the point of departure for the design of the first project and its subsequent use in the fieldwork in the data collection tool was a quantitative questionnaire that project. The knowledge obtained will be discussed in aimed at recording all the events related to employment the discussion section. or training done after or in parallel with the first job that The first qualitative element that was included in the an individual had after reaching the age of 16 (Vacchiano questionnaire was a life grid (Nico, 2016; Parry et al., et al., 2018). Additionally, in both projects, the quantita- 1999). In its most basic format, this instrument consists tive (closed-­ended) questions in the survey were organ- of a timeline in the form of a grid, normally organised in ised into three main blocks. First, information about the months or years, on which the interviewee can retrospec- interviewee's sociodemographic characteristics and his/ tively locate personal or collective events so that it is easy her most recent position in the labour market was ob- to recall and record the start date, end date and duration tained; second, the interviewee's career was reconstructed of the events narrated. Sometimes, to facilitate recall, im- retrospectively together with the characteristics of the portant events anchored in the collective or individual contacts that helped throughout this career; and third, the memory are placed on the baseline timeline so that the characteristics of the network of supportive contacts was person interviewed can link their personal memories with explored. The content of this third block of the question- those events. This instrument was used in both projects. naire was slightly different in the two projects; in the first Although the literature points to the use of life grids as project, a closed list of 20 active contacts at the time of the both a quantitative template and a qualitative tool (Parry interview was requested, while in the second project, this et al., 1999; Wilson et al., 2007), in our case, a life grid was list of personal contacts was constructed only of people used only as a qualitative tool to facilitate and order the who were helpful in the job seeking process, without set- recall of events. Once the sketching of these events on the ting a minimum or maximum number. As accessing and grid was completed (see Figure 1 for an example of the life storing the names of these contacts raises ethical issues, grid used), they were reviewed in detail and then recorded the interviewees were asked to provide a first name or a using the software used to collect the data. nickname so that only the interviewee could recognise The second qualitative element embedded in the sur- these people. vey was a block of biographical open-­ended questions, Several qualitative blocks were incorporated into this which could be extended with new questions raised during initial quantitative framework, resulting in a new single the conversation, in the style of a short semi-­structured USING A HYBRID DATA COLLECTION TOOL     | 5 TABLE 2 Questionnaires’ structure and blocks used in Project 1 (REDEMAS) and Project 2 (CAPSINES) Project 1 (REDEMAS) Project 2 (CAPSINES) 1st block of questions 1st block of questions Interviewer asks close-­ended questions about sociodemographic Interviewer asks close-­ended questions about sociodemographic characteristics and most recent labour market status of characteristics and most recent labour market status of interviewees interviewees 2nd block of questions 2nd block of questions Interviewee represents his/her labour market trajectory on a life Interviewee represents his/her labour market trajectory on a life grid and interviewer asks open-­ended questions about the grid and interviewer asks open-­ended questions about the events on this representation events on this representation 3rd block of questions 3rd block of questions Interviewer asks close-­ended questions about all the employment Delving into the information on the life grid, interviewer asks or training events along the trajectory and characteristics of the open-­ended biographical questions to explore the objectives social capital mobilised (when applicable) of interviewee along the trajectory (if any), his/her degree of agency, and the subjective turning points along the trajectory 4th block of questions 4th block of questions Interviewer asks close-­ended questions about a closed list of 20 Interviewer asks close-­ended questions about all the employment active personal contacts who offered different kinds of support or training events along the trajectory and characteristics of the (economic, expressive, emotional, etc.) along his/her trajectory social capital mobilised (when applicable). 5th block of questions 5th block of questions Interviewer asks interviewee to visualise and comment on the Interviewer asks close-­ended questions about an open list of network representation of his/her personal contacts. personal contacts who were helpful in any job seeking process along the labour market trajectory 6th block of questions Interviewer asks open-­ended biographical questions to explore the objectives of interviewee along his/her trajectory (if any), his/ her degree of agency, and the subjective turning points along the trajectory FIGURE 1 Qualitative life grid completed by an interviewee interview (it took approximately 15 min). In the first as a sort of scaffolding for the responses to biographical project, this qualitative interview was conducted at the questions. end of the questionnaire, with the objective of captur- The third qualitative element that was incorporated ing interviewees’ views on their careers and the labour into the questionnaire was used only in the first research market-­related decisions they had made. In the second project. It consisted of the visualisation of the network project, this short qualitative interview was developed si- representation of contacts reported by the interviewee; multaneously with the use of the life grid, since we re- this representation was obtained by means of the software alised that this positioning improved both the use of the programme used to collect the data (EgoNet), and was pre- life grid and the development of the interaction with inter- sented to the interviewee before the qualitative biograph- viewees. Qualitative interviews helped to better recall the ical questions. Although the visualisation of the network events represented in the life grid, and the life grid worked could potentially produce uncomfortable representations 6 |    VERD of the interviewees’ relationships (see Tubaro et al., 2021), at instructing them to use the questionnaire and survey our experience was highly satisfactory and similar to that software used in the fieldwork, as well as in briefings reported by McCarty et al. (2007), in which network vi- held during the projects to share fieldwork experiences sualisation worked as a cue to develop further reflection and to harmonise the coding of certain difficult respons- on the interviewee's network of supports and to check es—­in the sense of knowing what category to assign to the accuracy of the information obtained. In addition, the them—­offered by interviewees. graphic representation of the contacts made it possible to Indeed, by not using “hired hands” the interviewers’ deepen the importance of the different spheres of sociabil- interest in the information provided by interviewees was ity where interviewees had acquired these contacts. more evident and the interaction between interviewers and interviewees was open and flexible, easing commu- nication and facilitating rapport between them. In other Sample and fieldwork words, the style of interaction was that of a qualitative interview. This conception of the whole interview as a The sample and fieldwork for both projects also combined conversation-­like interaction and the attention to details quantitative and qualitative characteristics. First, Project also implied that there were no problems to add or modify 1 (REDEMAS) included 250 individuals and Project 2 events on the timeline initially represented in the life grid (CAPSINES), 150. These relatively small sample sizes during the conversation; sometimes, an event in the life were justified by the characteristics of the data collec- grid was only recalled after some time by interviewees. In tion tool and the volume of information collected in the both research projects, the software used did not require a survey. Given the dilemma of choosing either a greater chronological ordering of events as the primary focus was number of respondents or higher quality—­reliability and on recording the start and end dates of each event. Later, internal validity—­of the information collected, the second in the quantitative treatment of the data, these events option was chosen. Moreover, this sample size is common were ordered chronologically. The last qualitative trait of in social network analysis, as the final number of individ- the interaction between interviewer and interviewee was uals analysed is the result of multiplying the interview- the audio recording of the whole conversation. ees’ sample size (called egos) by the number of contacts (called alters) mentioned by every interviewee (Crossley et al., 2015). In addition, a non-­probabilistic quota sam- Data and analysis pling based on sex, age, neighbourhood, country of birth and education level was used to guarantee that the most The fieldwork developed in the two studies yielded inter- relevant social structuring variables were present in the esting quantitative and qualitative data. On the one hand, sample in the same proportion as found in the popula- three different quantitative matrices with three kinds of tion of the metropolitan area of Barcelona. Keeping these units of analysis (cases) were produced: (1) interviewees proportions allowed a statistical descriptive analysis of the (egos) and their sociodemographic characteristics and tra- sample to be carried out, avoiding a possible bias regard- jectories, (2) the separate events of each trajectory, and (3) ing these variables. the personal contacts mentioned in the interview (alters) The sample size of the projects made it possible for and their sociodemographic characteristics. On the other members of the research team to conduct all the inter- hand, qualitative information was obtained that contained views, so outsourcing of questionnaire administration the narrative of participants’ trajectories and reflections in was not necessary; as a result, the impersonal and de- relation to those trajectories. contextualised form of interaction typical of fieldwork These two types of data could be analysed inde- in survey studies was avoided. As stated by Seale (2004), pendently using the corresponding quantitative or quali- historically, with the standardisation of information for tative data analysis methods, or integrated, giving priority the purposes of statistical analysis and the development to either the qualitative or quantitative data (see Morgan, of large-­scale surveys, the recruitment of “hired hand” 2014; Morse & Niehaus, 2009). Regarding qualitative interviewers has been required, resulting in interview- data, all or a subset of the narratives obtained could be ers being simple “stimulus deliverers” that “would not directly analysed following a purely qualitative strategy, know much about the purpose of the investigation” (p. which would result in a qualitative approach to the study 104). In contrast, in our case, all interviewers were mem- and findings. This strategy was followed by Vacchiano bers of the research team, participated in the meetings (2021). Data analysis could also follow the rationale of a held about the projects and understood the objectives sequential quant → QUAL design (see Morgan, 2014), in and theoretical framework of the research. Moreover, all which a quantitative analysis of the close-­ended questions the interviewers participated in training meetings aimed is first carried out, and then, the qualitative information USING A HYBRID DATA COLLECTION TOOL     | 7 belonging to the cases that were particularly outstanding for young people with a low family socioeconomic back- in the quantitative results in regard to specific research ground than for those with a middle and high socioeco- questions was selected for a detailed qualitative analysis nomic background. Family ties among the former provide of their narratives. The difference with the most com- contact with persons of higher status and/or belonging to mon quant → QUAL design is that only one tool (and one the labour market sphere (for example, employers or co-­ source) was used to collect the quantitative and qualitative workers of family members); therefore, family members data. This was the strategy followed by Vacchiano et al. (strong ties) serve as intermediaries with helpful contacts (2021). However, other types of integrated analyses in outside the family sphere (weak ties). In contrast, young which either qualitative or quantitative information takes people from middle and high socioeconomic backgrounds priority in the analysis are possible. For instance, quan- obtained more direct support within the family network. titative information about the characteristics of personal In addition, in terms of agency, the findings show that contacts and network structure, which could be anal- young people from higher socioeconomic backgrounds ysed with standard Social Network Analysis quantitative and with stable occupational paths were able to mobilise methods, is helpful to provide (relational) context to the fewer contacts more efficiently, obtaining better outcomes. qualitative information on the strategies used to mobil- This can be explained by the learning process that young ise contacts and find a job (Muntanyola et al., 2019). This people with this profile undergo in job-­seeking, as well kind of use of quantitative and qualitative data adopts the as by the fact that having “quality contacts” in a personal logic of a QUAL + quant design (Morgan, 2014). network reduces the number of jobs individuals apply for before successfully finding employment. These additional findings should always be placed in the context of the dif- Findings ferences in the use of family contacts highlighted above. As mentioned above, the findings derived from the pro- jects sometimes relied mostly on the qualitative data, as in DISCUSSION: ADVANTAGES AND Muntanyola et al. (2019) or Vacchiano (2021), sometimes DISADVANTAGES OF USING A relied mostly on the quantitative data, as in Vacchiano HYBRID DATA COLLECTION TO O L et al. (2018) or Bolíbar et al. (2019), and sometimes com- bined both kinds of data, as in Vacchiano et al. (2021). In this section, the advantages of using the above-­described Each publication focused on specific aspects of the re- hybrid data collection tool are first presented, followed by search, but some of the main findings are briefly and ge- the disadvantages. nerically mentioned in what follows. First, regarding the data collection process, the interac- An initial interesting finding of this research was that, tion between interviewer and interviewee was far from the unlike the canonical hypothesis about the strength of weak impersonal and distant relationship typically found in the ties (Granovetter, 1973, 1974), family ties (i.e., strong ties) administration of a survey questionnaire. Given the char- were the main source of useful support for labour market acteristics of the tool used, this interaction was close to integration among young people. However, it should be the (usually) flexible and enjoyable interaction present in added that the importance of family depends on different qualitative interviews. This conversation-­like exchange in- elements: (1) it was more important at the beginning of cluded questions about the social, family and work context the working career than later on; (2) it was more important in which the labour market trajectory was developed, as after the 2008 recession than before; (3) it was more im- well as requests for clarification and details (probing) with portant among young people with a low family socioeco- the aim of facilitating the recall of the events described nomic background than among those with a middle and and the collection of all information with maximum accu- high socioeconomic background. These findings, which racy. The exchange also allowed for good rapport between were obtained from the analysis of the quantitative se- the interviewer and interviewee, and therefore, had a pos- quential data on employment and social contacts, coincide itive effect on the quality of the data obtained. in part with findings reported in previous literature (see a Regarding the reliability and internal validity of the review in Kramarz & Skans, 2014). However, the multiple data—­or credibility and consistency, using the terms pro- facets from the same phenomenon obtained thanks to the posed by Lincoln and Guba (1985)—­different advantages quantitative and qualitative data gathered simultaneously were identified. First, the possibility of recording quali- from the same individuals provided us with a broader and tative information and the depth and richness associated more nuanced picture of the importance of family ties in with it provided a good context for deepening and probing labour market integration. In this sense, it was found that the responses to the most standard quantitative closed-­ the support offered by family ties was of a different nature ended questions in the questionnaire. This advantage is 8 |    VERD directly connected to the capacity of hybrid tools to help which provided a qualitative framing to the administration researchers clarify and better understand how interview- of the questionnaire. In addition, care was taken to avoid ees interpret question categories, as highlighted by Gobo the potential ethical problems of these lengthy interviews. and Mauceri (2014). Additionally, the use of a qualitative First, the persons interviewed were warned about the life grid gave researchers confidence about the accuracy length of the interview and were advised to have a look at of the sequence of events collected and the findings ob- their curriculum vitae before the research encounter or to tained through their analysis, knowing the potential recall even bring it to the interview. Second, interviewees were bias of retrospective sequential data (Parry et al., 1999). selected based on their willingness to participate in the sur- Second, the qualitative information obtained provided im- vey, their time availability was respected, and interview ses- portant insight into the agency, objectives and representa- sions were adapted to their convenience.2 tions of the interviewees, i.e., it provided the interviewees’ The learning process just mentioned regarding the use perspectives in connection with their actions. In this way, of our tool was taken into account in the second project, in interpretation of the data was improved, as the motives which a better balance between quantitative and qualita- or reasoning behind the observed actions was known. tive elements was sought so that the quantitative part was Third, the collecting of quantitative, qualitative and net- shorter, even if it continued to have more weight than the work data together, and as part of an individual narrative, qualitative part. Thus, in the second research project, the provided a comprehensive view of the life of interview- presence of the quantitative and qualitative elements in ees; their labour market trajectory and personal network the questionnaire was more similar, and the overall inter- were parts of a single overarching picture, which included action between interviewer and interviewee was shorter other information regarding social background, educa- (between 90 and 120 min). Moreover, in the second re- tional level and other relevant variables associated with search project, qualitative biographical questions were their labour market status and network social capital. placed earlier in the questionnaire, just after the block Finally, the richness and dimensions of the information where the life grid was used (see Table 2). This produced and the internal connections revealed in the data allowed a synergy between these two blocks and proves that the for an articulated approach to the complexity and multiple sequence of the blocks is very relevant in the design of a dimensions of the research object. A good example of the hybrid questionnaire. Biographical questions were some- outcomes obtained from this articulated approach is the times taken up throughout the conversation between the broader and nuanced picture of the use of family ties for interviewer and the interviewee, too. This increased the labour market integration and its relationship to the in- flexibility of the questionnaire and enhanced the rapport dividual's socioeconomic background, as reported above. between interviewer and interviewee. The most important limitation was the amount of time the tool, as initially designed, required. The objective of col- lecting different types of data with a single tool in one re- CONC LUSIONS search encounter resulted in a lengthy interview process. We realised in the pilot tests during the first research proj- This article describes the use of a hybrid data collection ect that the initial questionnaire was too ambitious regard- tool applied in two research projects that examined the ing information on the network of contacts. Therefore, in relationship between the labour market trajectories of the final version of the questionnaire, the part on the net- young people and the use and accumulation (or loss) of work of contacts was trimmed down. This decision was network social capital along these trajectories. First, the made after identifying a sort of trade-­off between the de- rationale for using this kind of tool was presented. The tails of sequential information and those of personal net- research questions of the projects were multifaceted and work information, and after considering that the detailed had many levels and dimensions, and causality among information on labour market trajectories was more im- the variables could develop in multiple directions. The portant for the objectives of the project. However, even use of a single data collection tool allowed these multiple after reducing the number of questions about contacts, and given that the number of events in some labour market tra- jectories was very high—­reflecting unstable and precarious 2 Although a statistical generalization of the results was not sought careers—­some interviews were still quite long, lasting up to —­given the sample sizes and the non-­probabilistic character of the 3 h. This length could potentially result in a decrease in re- sampling strategy—­the potential bias this willingness to participate could introduce in the sample was controlled by means of a close and liability towards the end of the interview caused by inter- strict monitoring of the interviewees’ fit to the established sample viewee fatigue. This problem was solved by ensuring that quotas, which were based on the variables considered as having the the interaction between the interviewer and interviewee most relevant influence on labour market trajectories and social capital was loose and flexible, in the style of a casual conversation, characteristics. USING A HYBRID DATA COLLECTION TOOL     | 9 levels, dimensions and variables to be gathered from every implications. Two solutions were applied in the second interviewee, while at the same time maintaining a com- project to overcome this drawback. The most obvious and prehensive and overarching view of the interrelationship straightforward solution was to reduce the number of between labour market trajectories and changes in per- questions so that the interview duration was two hours at sonal networks. most. Even with this length limitation, a second strategy Of course, using a single tool producing different was applied: interviewers tried to turn the survey into a data types was not the only possible option, as data conversation-­like interaction, introducing more open-­ could be produced separately with different tools and ended (qualitative) questions throughout the interview later combined. However, synergy was gained (see process and playing a role far from a simple “stimulus Fetters & Freshwater, 2015) by the combination of both deliverer”. This produced a more pleasant and enjoyable qualitative and quantitative characteristics in a single dialogue, reducing fatigue and the research burden on data collection tool. Axinn and Pearce (2006) and Gobo interviewees. and Mauceri (2014) have already highlighted some of This leads to a more general reflection on the use the advantages of this kind of data collection tool, al- of hybrid data collection tools and the need to balance though the tools they designed were different from ours their qualitative and quantitative elements. The collec- and were not applied to the broad field of the relation- tion of different types of data—­quantitative, qualitative ship between young people and the labour market. This and network in our case—­with the same tool is a less article emphasises the advantages regarding the charac- flexible strategy than is the use of two different forms teristics and quality of the data obtained, although the of data collection, as in more common mixed methods downsides have also been addressed. designs. When using different forms, the design and ap- One of the main advantages that should be men- plication of each data collection tool is developed sep- tioned is that the data obtained permitted us to address arately, which allows for multiple options, regarding, simultaneously the different aspects of the research for instance, just to name a few, the ordering and tim- questions posed. Although some questions in the ques- ing of fieldwork, the size of the sample to which data tionnaire had the usual closed-­ended format of a stan- collection tools will be applied and the kind of interac- dard survey, offering the possibility of producing tion between interviewer and interviewee. When using different quantitative data matrices, the introduction of a hybrid data collection tool, there is only one single various qualitative blocks in the questionnaire produced moment of data collection, only one sample and only qualitative data that could be easily integrated with the one kind of interaction possible. Therefore, a “commit- quantitative data.3 ment decision” must be made to avoid jeopardising the It can be argued that the comprehensive character of essence and quality of the different kinds of data that the data collection process is, in a certain way, a return to are simultaneously obtained. In return, the above men- the characteristics that, until the 1940s, were widespread tioned advantages are obtained. Therefore, the use of a in survey research. Until then, survey research was not hybrid tool should be given serious consideration if the synonymous with quantitative research since, in many research objectives and research object justify its use. cases, both quantitative and qualitative data were col- lected in interviews (Fontana & Frey, 2000). This more ACKNOWLEDGEMENTS artisanal, less impersonal way of obtaining data was lost I would like to thank Dr. Oriol Barranco and Dr. Mireia due to the standardisation of information for statisti- Bolíbar for their comments on earlier drafts of this article. cal analysis and the development of large-­scale surveys (Seale, 2004). In the case of the two research projects CONFLICT OF INTEREST reviewed, this impersonal way of obtaining data was The author declares that he has no competing interests. avoided because the sample sizes allowed the research team to be involved in the data collection process. A re- DATA AVAILABILITY STATEMENT turn to this type of research approach would make great The data bases mentioned in the article are available upon sense for addressing certain complex research topics, request to the author. such as those mentioned here. The downside of the hybrid tool was that the increased ORCID insight gained was at the expense of a lengthy interview. Joan M. Verd https://orcid.org/0000-0002-8636-3659 This could lead to interviewee fatigue and has ethical REFERENCES 3 The way this information was integrated during the data analysis stage Axinn, W. G., & Pearce, L. D. (2006). Mixed method data collection is not discussed in the article. strategies. Cambridge University Press. 10 |    VERD Bazeley, P. (2010). Computer-­assisted integration of Mixed Methods Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Sage. data sources and analyses. In A. Tashakkori & C. Teddlie (Eds), McCarty, C., Molina, J. L., Aguilar, C., & Rota, L. (2007). A com- SAGE handbook of mixed methods in social and behavioral re- parison of social network mapping and personal network search (2nd ed., pp. 431–­467). Sage. visualization. Field Methods, 19(2), 145–­ 162. https://doi. Bazeley, P. (2018). Integrating analyses in mixed methods research. org/10.1177/15258​22X06​298592 Sage. Moerbeek, H., & Flap, H. (2008). Social resources and their effect on Bolíbar, M., Verd, J. M., & Barranco, O. (2019). The downward spiral occupational attainment through the life course. In N. Lin & of youth unemployment: An approach considering social net- B. H. Erickson (Eds.), Social capital. An international research works and family background. Work, Employment and Society, program (pp. 133–­156). Oxford University Press. 33(3), 401–­421. https://doi.org/10.1177/09500​17018​822918 Morgan, D. L. (2014). Integrating qualitative and quantitative meth- Creswell, J.W., & Plano Clark, V.L. (2011) Designing and conducting ods: A pragmatic approach. Sage. mixed methods research. 2nd ed. Sage. Morse, J. M., & Niehaus, L. (2009). Mixed method design. Principles Creswell, J. W., & Plano Clark, V. L. (2018). Designing and conducting and procedures. Left Coast Press. mixed methods research. 3rd ed. Sage. Muntanyola, D., Barranco, O., & Vacchiano, M. (2019). Inequality Crossley, N., Belloti, E., Edwards, G., Everett, M. G., Koskinen, J., & beyond networking: Personal networks and mobilization of Tranmer, M. (2015). Social network analysis for ego-­nets. Sage. contacts by young job seekers in Barcelona. Revista Española Fetters, M. D., & Freshwater, D. (2015). The 1 + 1 = 3 integration de Sociología, 28(2), 207–­226. https://doi.org/10.22325/​fes/ challenge. Journal of Mixed Methods Research, 9(2), 115–­117. res.2019.01 https://doi.org/10.1177/15586​89815​581222 Nico, M. L. (2016). Bringing life “back into life course research”: Fielding, N. G. (2012). Triangulation and mixed methods designs: Using the life grid as a research instrument for qualitative data Data integration with new research technologies. Journal collection and analysis. Quality & Quantity, 50, 2107–­2120. of Mixed Methods Research, 6(2), 124–­ 136. https://doi. https://doi.org/10.1007/s1113​5-­015-­0253-­6 org/10.1177/15586​89812​437101 Parry, O., Thomson, C., & Fowkes, G. (1999). Life course data Fontana, A., & Frey, J. (2000). The interview from structured ques- collection: Qualitative interviewing using the life grid. tions to negotiated text. In N. K. Denzin & Y. S. Lincoln (Eds.), Sociological Research Online, 4(2), 102–­ 112. https://doi. Handbook of qualitative research (pp. 361–­376). Sage. org/10.5153/sro.233 Freedman, D., Thornton, A., Camburn, D., Alwin, D., & Young-­ Ryan, L., D’Angelo, A., Kaye, N., & Lorinc, M. (2019). Young peo- DeMarco, L. (1988). The life history calendar: A technique for ple, school engagement and perceptions of support: A mixed collecting retrospective data. Sociological Methodology, 18, 37–­ methods analysis. Journal of Youth Studies, 22(9), 1272–­1288. 68. https://doi.org/10.2307/271044 https://doi.org/10.1080/13676​261.2019.1571178 Gobo, G. (2015). The next challenge: From mixed to merged meth- Schober, M. F., & Conrad, F. G. (1997). Does conversational inter- ods. Qualitative Research in Organizations and Management: An viewing reduce survey measurement error? Public Opinion International Journal, 10(4), 329–­331. https://doi.org/10.1108/ Quarterly, 61, 576–­602. https://doi.org/10.1086/297818 QROM-­07-­2015-­1309 Seale, C. (2004). History of qualitative methods. In C. Seale (Ed.), Gobo, G., & Mauceri, S. (2014). Constructing survey data: An interac- Researching society and culture (2nd ed., pp. 99–­113). Sage. tional approach. Sage. Tubaro, P., Ryan, L., Casilli, A. A., & D’Angelo, A. (2021). Social Granovetter, M. S. (1973). The strength of weak ties. American Journal network analysis: New ethical approaches through collec- of Sociology, 78(6), 1360–­1380. https://doi.org/10.1086/225469 tive reflexivity. Introduction to the special issue of social net- Granovetter, M. S. (1974). Getting a job: A study of contacts and ca- works. Social Networks, 67, 1–­ 8. https://doi.org/10.1016/j. reers. Harvard University Press. socnet.2020.12.001 Granovetter, M. S. (1985). Economic action and social structure: Vacchiano, M. (2021, May). Nine mechanisms of job-­ searching The problem of embeddedness. American Journal of Sociology, and job-­ finding through contacts among young adults. 91(3), 481–­510. https://doi.org/10.1086/228311 Sociological Research Online. First published online. https:// Harrits, G. H., & Møller, M. Ø. (2021). Qualitative vignette exper- doi.org/10.1177/13607​80421​1009525 iments: A mixed methods design. Journal of Mixed Methods Vacchiano, M., Martí, J., Yepes-­Cayuela, L., & Verd, J. M. (2018). Research, 15(4), 526–­545. https://doi.org/10.1177/15586​89820​ Personal networks in job insertion among young adults in 977607 times of crisis: An analysis in Barcelona. Revista Española de Hollstein, B. (2019). What autobiographical narratives tell us about Investigaciones Sociológicas, 161, 121–­140. the life course. Contributions of qualitative sequential ana- Vacchiano, M., Yepes-­Cayuela, L., & Martí, J. (2021). The fam- lytical methods. Advances in Life Course Research, 41, Article ily as (one-­or two-­step) social capital: Mechanisms of sup- 100248. port during labor market transitions. Community, Work Kerrigan, M. R. (2014). A framework for understanding community & Family, 24(4), 471–­ 487. https://doi.org/10.1080/13668​ colleges’ organizational capacity for data use: A convergent par- 803.2019.1687425 allel mixed methods study. Journal of Mixed Methods Research, Verd, J. M., & López, M. (2011). The rewards of a qualitative ap- 8(4), 341–­362. https://doi.org/10.1177/15586​89814​523518 proach to life-­course research. the example of the effects of Kramarz, F., & Skans, O. N. (2014). When strong ties are strong –­ social protection policies on career paths. Forum Qualitative Networks and youth labor market entry. Review of Economic Sozialforschung/Forum: Qualitative Social Research, 12(3), Art. Studies, 81, 1164–­1200. 15. https://doi.org/10.17169/​fqs-­12.3.1753 Lin, N. (2000). Inequality in social capital. Contemporary Sociology, Verhaeghe, P., Van der Bracht, K., & Van de Putte, B. (2015). 29(6), 785–­795. https://doi.org/10.2307/2654086 Inequalities in social capital and their longitudinal effects on USING A HYBRID DATA COLLECTION TOOL     | 11 the labour market entry. Social Networks, 40, 174–­184. https:// doi.org/10.1016/j.socnet.2014.10.001 How to cite this article: Verd, J. M. (2022). Using Wengraf, T. (2001). Qualitative research interviewing. Biographic nar- a hybrid data collection tool: Analysis of youth rative and semi-­structured methods. Sage. Wilson, S., Cunningham-­Burley, S., Bancroft, A., Backett-­Milburn, labour market trajectories integrating quantitative, K., & Masters, H. (2007). Young people, biographical narra- qualitative and social network data. International tives and the life grid: Young people’s accounts of parental Journal of Social Welfare, 00, 1–­11. https://doi. substance use. Qualitative Research, 7(1), 135–­151. https://doi. org/10.1111/ijsw.12528 org/10.1177/14687​94107​071427