E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 192 European Integration Studies No. 14 / 2020, pp. 192-206 doi.org/10.5755/j01.eis.1.14.26381 Abstract Idea Management System Application Type Impact on Idea Quantity http://dx.doi.org/10.5755/j01.eis.1.14.26381 Idea management system application considers idea quantity as the key to idea management success. The aim of this paper is to examine how different idea management system application types impact idea quantity. The authors conducted empirical research by conducting a survey based on adaptive structuration theory framework. In the research paper, an analysis of 447 responses was included. The study shows how to separate idea management system application types impact by idea quantity. The target group consisting of commercially available web-based idea management system applied enterprises bias present in the survey research may limit the generalisability of the results. The study contributes to the discussion about the idea management system application type impact on the idea management results by showing that different idea management system application types lead to different idea management results. KEYWORDS: Idea management systems, Idea quantity, Survey research, Web-based idea management, Innovation. Introduction Elina Mikelsone BA School of Business and Finance Innovation management and application of information technologies in organisations has become increasingly more relevant over the last few decades. Web-based idea management systems (IMS) fall in line with the current developments (e.g. growing importance of ICT, the spread of open in- novation and co-innovation, etc.) in all previously mentioned considerations, IMS is a manageable systematic tool to generate and evaluate ideas. The use of web-based IMS has become a part of the organizational culture in various enterprises and Web-based IMS are used by many well-known or- ganizations such as Boeing, P&G, Volkswagen, Xerox, Pentax, Heineken, Panasonic, Sony, Fujitsu, Electrolux, Volvo, etc. The authors expect that throughout the following years the role of web-based IMS will grow as even more organizations will start to apply them. Many good examples show positive effects on organizations performance that use web-based IMS. For example, BT Group is using its IMS Webstorm which helped the company acquiring 10 000 new ideas in the seven years between 2005 and 2012. Realization of these ideas has helped the company to increase its revenue by 100 million pounds and improve customer loyalty (Bright Idea, 2010). Another example is Bruce Submitted 05/2020 Accepted for publication 07/2020 Idea Management System Application Type Impact on Idea Quantity EIS 14/2020 Aivars Spilbergs BA School of Business and Finance Tatjana Volkova BA School of Business and Finance Elita Lielā BA School of Business and Finance Jānis Frišfelds Maastricht University School of Business and Economics http://doi.org/10.5755/j01.eis.0.13.23562 193 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 Power, the only privately owned nuclear power station in Canada. In two years since it started using IMS Idealink Open, it has acquired more than 2700 new ideas and more than 10 000 participants have participated in their IMS process (generation, development). The use of IMS can lead to both a decrease in costs and an increase in revenue (Brain Bank, 2014). Application cases show that this tool gives the possibility to connect internal and external idea creators and evaluators in the idea management (IM) process and these systems could connect different entrepreneurship areas, for example, intrapreneurship and innovation management, opportunity identification and creation. But there is a lack of research on the web-based IMS application types and their respective results. Authors of this paper aim to explore web-based IMS application type impact on its application results. To fill the gap, authors apply theoretical and empirical approach with the main aim to ex- amine how different IMS application types impact IMS results. Applegate (1986) is the first researcher mentioning IM and to begin IM and IMS research. Since then there have been several academic perspectives on how to research IM and IMS. A majority of researches focuses on systematic aspects of IM and IMS (e.g. Bailey and Horvitz, 2010; Barczak et al., 2009; Bjork and Magnusson, 2009; Coughlan and Jahanson, 2008; Flynn et al., 2003; Gal- brait, 1982; Gish, 2011; Green et al., 1983; Korde and Paulus, 2016; Vandenbosch et al., 2006) and structural (e.g. Bassiti and Ajhoun, 2013; Bergendah and Magnusson, 2014; Divakaran, 2016; Luo and Tobia, 2015; Narvaez and Gardoni, 2015; Poveda et al., 2012; Summa, 2004; Voigt and Brem, 2006; Westerski and Iglesias, 2011; Wooten and Ulrich, 2015). Structural literature sources focus on design and the process, but systematic literature sources focus on social capital, creativity, cognition, etc. (Rose and Jensen, 2012). Authors have revealed in the previous researches that there are multiple types of research availa- ble with a structural perspective that provide a theoretical base for IMS concept exploration. Litera- ture about IMS overviews mostly focus on existing IMS and their application and potential improve- ments (e.g. Summa, 2004; Bakker et al., 2006; Coughlan et al., 2008; Bothos et al., 2008; Bjork et al., 2009; Barczak et al., 2009; Beretta, 2015; Tung et al., 2009; Bailey et al., 2010; Hrastinski et al., 2010; Holzblatt et al.,2011), but some researches also aim to research development of new IMS (e.g. Flynn et al., 2003; Vandenbosch et al., 2006; Bothos, et al, 2009; Iversen, et al., 2009; Bansemir et al., 2009; Bettoni et al., 2010; Xie et al.,2010; Bothos et al., 2012; Lowe and Heller, 2014). This paper aims to be part of the first type of papers which explores existing systems focusing on commer- cially available systems. Most researches that explore existing IMS research focus on one or a few IMS but this research is based on a survey across multiple different IMS users. IMS has not been given sufficient scientific attention and it should be researched how different IMS types impact its application results (van den Ende et al., 2015). This research is aiming at providing a contribution to fill this gap. First, the paper will help researchers and IMS users to understand the ba- sic IMS application types and their potential results. Second, the exploration of different IMS types and their results could motivate entrepreneurs to re-evaluate their current approach to IM. Third, develop- ers and users of web-based IMS see the potential of these systems but positive outcomes often do not occur and that is one of the reasons why organizations do not use them in the long term (DeSanctis and Poole, 1994). Due to these reasons, it is important to explore web-based IMS application types and their results, to explain what results companies could expect based on different application types. In this paper, IM is defined as a systematic, manageable process of idea generation, evaluation, and repeated idea generation and evaluation. The IMS is defined as a tool, tool kit or complex system which provides systematic, manageable process in IM (Mikelsone and Liela, 2015). The authors use 2 IMS classifications: based on involved idea sources (internal, external, mixed) and based on the application focus (active, passive). The research aims to answer the research ques- tion: How different IMS application types impact idea quantity? To answer these questions 4 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 194 hypotheses will be tested that are based on the 3 types of IMS and 4 dimensions of the results: _ (H1) Active IMS provide higher idea quantity than passive. _ (H2) External IMS provide higher idea quantity than internal. _ (H3) Mixed IMS provide higher idea quantity than internal. _ (H4) Mixed IMS provide higher idea quantity than external. Figure 1 summarizes the motivation for this paper. Aim Gap to fill (academic topicality) Overall topicality (trends and practical topicality) Use theoretical and empirical approach to examine how different IMS application types impact IMS application idea quantity. Main literature gap – no research on how IMS application types impact idea quantity. IT application and innovation management in organizations is more relevant now than ever before. Trends to match – (1) in the age of knowledge tools that provide means for acquiring, evaluation and development of knowledge and ideas are extremely important; (2) the growing role of ICT increases the importance of web-based tools that support the innovation process; (3) web-based IMS is becoming more important in the context of open innovation and co-innovation, giving them access to both internal and external sources of ideas and knowledge. Figure 1 Research development motivation Source: created by author’s This paper fulfils an identified need to clarify IMS types and their impact on idea quantity. This pa- per creates academical contribution: it researches different classifications of IMS and their impact on idea quantity. Practical contribution - web-based IMS application types and their idea quantity could help to evaluate the potential application of IMS in different scenarios. The rest of the paper is structured as follows. The second section introduces the theoretical background. The third section continues by presenting the research methodology. The fourth section provides the answers to the research questions through conducted research. This structure can be seen in Figure 2. In tr o du ct io n Th eo re ti ca l B ac kg ro u nd Methodology Findings Conclusions Research instrument Data collection Data analysis Basic data characteristics IMS application type impact on idea quantity Summary Implications Future research directions Source: created by author’s Figure 2 Paper Structure Secon 1 Secon 2 Secon 3 Secon 4 Secon 5 Theoretical background Idea management system basic assumptions In this paper, the definition of the IM is based on the following assumptions that that IM is: (1) sys- tematic process; (2) manageable process; (3) main parts of IM are idea generation, evaluation, and repeated idea generation and evaluation (if it is needed). Based on these assumptions IMS is a tool, tool kit or complex system which provides systematic, manageable process in IM (Mikel- sone and Liela, 2015). Table 1 provides a detailed description of IMS, characterising all previously mentioned elements with its sub-elements. 195 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 Table 1 IMS Main Characteristics IMS - tool, tool kit or complex system which provides systematic, manageable process of: Idea generation (preparation, capture/gathering of ideas, retention, enhancement) Idea evaluation (screening, selection, retention) Continuation of IM (concept development, distribution of ideas, support during implementation with repeated IM and rewarding, retention) e.g. Korde and Paulus, 2016; Wooten and Ulrich, 2015; Summa, 2004 e.g. Westerski, 2013; Summa, 2004 e.g.Summa, 2004 Source: based on Mikelsone, Volkova and Liela (2019) There are 3 IM types: external, internal and mixed IM. External IM means external idea generation and evaluation (main IM sources – experts, partners, customers and other stakeholders outside the organization) (e.g. Bothos et al., 2008; Tung et al., 2009; Westerski et al, 2011; Bothos et al., 2012). Internal IM is internal idea generation and evaluation in an organization (main IM source employees) (e.g. Iversen et al., 2009; Fatur et al., 2009; Bansemir et al., 2009; Glassmann, 2009; Klein and Lechner, 2010; Bettoni et al., 2010; Selart and Johansen, 2011; Shani et al., 2011; Moss et al., 2011; Deichmann, 2012; Aagaard, 2012, 2013; Poveda et al., 2012; Bassiti and Ajhoun, 2013, Wood, 2003). Mixed IM - idea generation and evaluation involving internal and external sources (e.g. Fritz, 2002; Nilsson et al., 2002; Voigt et al., 2006; Brem et al., 2007; Enkel et al., 2009; Brem et al., 2009; Sandstrom and Bjork, 2010; Baez and Convertino, 2012). Idea management system application types For this research the authors have applied two categories for classifying IMS application: (1) based on the involved IM source; (2) based on the IM application focus. There are other possible categories for IMS classification: based on the provided process functions (lim- ited, full, extra) and based on the IMS price type (monthly payment, yearly payments). The last two types of classifications will not be investigated further as they focus on systems, not on their application type. Authors based on the ideas divide all IMS application cases as follows: _ internal IMS by involvement internal idea creators and evaluators; _ external IMS by involvement external idea creators and evaluators; _ mixed IMS by involvement internal and external idea creators and evaluators. Based on the application focus all systems could be divided as “active” and “passive’’, therefore, there are passive and active IMS. Passive IMS collect all ideas in an unfocused manner, but active IMS provide functions to collect ideas in a focused manner and most cases includes idea evalua- tion possibilities. IMS type descriptions are provided in Table 2. Table 2 IMS Application Types Classifications Classification criteria: based on the application focus Passive IMS Active IMS Functions Type of focus Functions Type of focus Focus on idea generation Unfocused process Focus on all IM dimensions Focused process Classification criteria: based on the involved IM source Description Main IM source Description Main IM source Description Main IM source IMS that allows involving only internal IM sources Employees IMS that allows involving only external IM sources Crowds, experts, clients, etc. IMS that allows involving internal and external IM sources Employees; clients, experts, crowds, etc. Source: based on Mikelsone, Volkova and Liela (2019) E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 196 Idea Management System Application Results Quality and quantity of ideas are most often used as measurements for IM and IMS application, and as a result, should be considered as the main elements of the web-based IMS application outputs. Denis and Garfield (2003) have revealed that decision support system processes may encourage more participation that also provides a challenge to research this element in the IMS context. Woo- ten and Ulrich (2015) had researched feedback importance in the idea management process, based on the conclusion that managers face a decision about if and how to provide in-process feedback to the idea generators about the quality of submissions. Their research revealed that directed feed- back benefits the average quality of entries submitted. The stimulus that impacts web-based IMS application and its results can also be researched. The authors have concluded that there is no common view on IMS output elements, except idea quality and quantity, and involvement. It would be advisable to create IMS effectiveness evaluation tool that would include the most important output elements. In this research, authors will apply the most frequent researched IMS output variables - idea quantity, idea quality and involvement. Idea quality could be defined as the average quality of generated ideas (idea creativity) (Selart & Johansen, 2011; Deichmann, 2012; Bjork & Magnusson, 2009). Idea quantity could be defined as a number of ideas generated (MacCrimmon & Wagner, 1994; Korde & Paulus, 2016; Girotra & Ulrich, 2010; Deichmann, 2012). There is an ad- ditional variable chosen to research results – involvement or number of involved people (Dennis & Garfield, 2003; Deichmann, 2012). In this research, authors focus on idea quantity. Methodology Research instrument for measuring web-based IMS application results A questionnaire was created for web-based IMS applied companies. The survey was conducted in the summer/autumn of 2018. Methods for obtaining primary data are described in Table 3. This survey results allowed to compile data on IMS in 8 blocks, according to Adaptive Structura- tion Theory: (1) type of IMS; (2) tasks; (3) organization system; (4) adaptation and type of use; (5) IMS results; (6) organisational effectiveness; (7) new structures; (8) problems with the use of IMS. Table 4 highlights the survey block - IMS results. A total of 186 elements are summarized in 8 question blocks. In this paper, the applied survey block is IMS results. The questionnaire was created and distributed in English, as the dominant language of the IMS and its use in English. All criteria were based on literature analysis and updated scales were based on the results of the case studies as used to describe the results of the application of IMS. Data collection The survey was conducted on the survey platform 'The QuestBack' (https://www.unipark.com/) created by UNIPARK. This platform was chosen because it is: (1) focused on academic surveys; (2) widely recommended by world-class researchers; (3) provides data security required by IMS representatives - BSI-certified data centre in accordance with ISO 27001; (4) complies with the requirements of the EU General Data Protection Regulation. To test the questionnaire, it was sent to 9 companies that conducted a survey and were able to comment on any question. The test was done in 3 rounds, the questionnaire was sent to 3 companies using the IMS when comments were received and based on the feedback the ques- tionnaire was improved. In the third round, comments on the structure or clarity of the questions were no longer provided. Based on the tests, the time of completing the questionnaire was de- termined (20-30 minutes). After the test, the survey results were deleted. It should be noted that to reach the target audience more accurately, the authors asked IMS devel- opers to distribute the survey to their clients. It was stipulated that the survey should only be sent to companies using the system in question to the person in charge of the IMS (mostly think-tanks, innovation managers or business managers). In the authors' private communication with 107 IMS 197 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 Table 3 Primary data selection un analysis description Aim Data selection Data analysis Period Steps IMS application types impact on the results Survey of the enterprises that applies web-based IMS (n>400) Statistical analysis (frequency distribution, standard deviation, t-statistics, degrees of freedom (df), critical values (tc) and p-values, etc.) 3r quarter of 2018 1 Survey development based on literature analysis and developed classifications. 2 Round survey test (data not included in the analysis). 3 Survey distribution to 107 web-based IMS develop- ers, that they could distribute to their clients. 4 400 valid surveys. 5 Standard deviations to evaluate the data consisten- cy for the analysis. 6 Data analysis through selected methods Source: created by author’s Question Scale Based on Idea quantity - What is the average number of ideas created per task? None Based on the literature studies and empirical case studies conducted by the authors Using internal IMS To 10 Using external IMS 11 - 100 Using mixed IMS 101 - 1000 Using active IMS 1001 - 5000 Using passive IMS 5001 - 10 000 More than 10 001 Table 4 Questionnaire Section – IMS Results Source: created by author’s developers and the information published by the IMS concerned, it was concluded that the IMS employs around 70, 000 - 100,000 companies (derived from the average number of IMS clients). Invitation for their system applicants to involve sent to 108 idea management system crea- tors – Crowdicity, Viim, IdeasMine, Idea Drop, Ideanote, Receptive, CrowdWorx Innovation Engine, Ideawake, Sideways 6, OrganisedFeedback, Exago SMART, Ideation360, ProdPad, Vetter Online Suggestion Box, IdeaLab, GroupMap, Ideaflip, IDEAFOX, iMindQ, Innovation Cloud, innovation5g. com, MindView 7, WE THINQ, eXo Platform, IdeaScale, HYPE Innovation, MindManager, Milanote, Innovation Platform, Kindling, Coggle, DataStation Innovation Cloud, SprintGround, Be-novative, BrainStorm, Idearium, Stormboard, MangoApps, Nova-Innovate Innovation Management Soft- ware, Wizeline, Comapping, FeatureMap, Glint Innovation, Ideacomb, Mindomo, SoapBox, Sprint, Wave, Work by InnoCentive, Braincatena, BrightSpars, Bubbl.us, CogniStreamer, De Idee Manage- ment Tool, e-Zassi, easycrit, eVSM, Firefly, GainX, germ.io, Headstarter, HiFISH, Hives.co, id-Force, ID8 Enterprise, Ideabox, IdeaBridge, Ideafactory, Ideakeep, IdeaLinker Accelerate, Ideasbank, IDhall SC, Includer, INDONIS, Inno360, InnoEngines, innosabi, Innovation Agora, Innovation Central, In- novationCast, InnovationStation, Innovbook, ITONICS Ideation, MindApp, Nosco, NovaMind, Onyx Cloud Ideas, Orchidea, PIT, Postwaves, SocialJsIdeas, Solverboard, Sophia, TalkFreely, Verve, Voco- li, Yambla, Accept360, Idea Glow, IdeaSpotlight, Idea Management Software by Planbox, Brightidea, IDEALYST, Online Suggestion Box, Flagpole, Spigit Idea Management Software, Academy of Ideas. Data analysis To validate data for the further analysis the pre-analysis was conducted by using the following methods: _ Point estimation and interval estimation – ‘’the process of providing a numerical value for a pop- ulation parameter based on information collected from a sample. If a single figure is calculated for the unknown parameter, the process is called point estimation. The process of providing a E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 198 numerical value for a population parameter based on information collected from a sample. If an interval is calculated which is likely to contain the parameter, then the procedure is called interval estimation” (Everitt & Skrondal, 2010); _ Frequency distribution – “the division of a sample of observations into a number of classes, together with the number of observations in each class. Acts as a useful summary of the main features of the data such as location, shape, and spread” (Everitt & Skrondal, 2010); _ Mean of the group - to get the average value of the group; _ Standard deviation – to measure the spread of a set of observations; _ Modal and medial class (interval) – to observe the most frequent and ‘’the value in a set of ranked observations that divides the data into two parts of equal size” (Everitt & Skrondal, 2010); _ Coefficient of variation - to measure “the spread for a set of data defined” (Everitt & Skrondal, 2010); _ Confidence interval – to range the values, calculated from the sample observations, that is believed, with a particular probability, to contain the true parameter value (of the population); _ Sampling error – to observe “the difference between the sample result and the population characteristic being estimated” (Everitt & Skrondal, 2010). _ To test the hypothesis the following data analysis methods were applied: _ Significance tests for a population mean number for the result variable. _ The t-test was used to measure statistically significant variations between IMS types. It was applied to test the hypothesis. _ Calculating p-values for the given test statistics and the degrees of freedom. Findings Basic data characteristics – idea quantity Respondents frequency distribution based on survey data is shown in Figure 3. The further detailed analysis consists of the arithmetic mean of the group, standard deviation, modal and medial class (group), coefficient of variation. Arithmetic mean of grouped data is calculated as follows: for the unknown parameter, the process is called point estimation. The process of providing a numerical value for a population parameter based on information collected from a sample. If an interval is calculated which is likely to contain the parameter, then the procedure is called interval estimation” (Everitt & Skrondal, 2010);  Frequency distribution – “the division of a sample of observations into a number of classes, together with the number of observations in each class. Acts as a useful summary of the main features of the data such as location, shape, and spread” (Everitt & Skrondal, 2010);  Mean of the group - to get the average value of the group;  Standard deviation – to measure the spread of a set of observations;  Modal and medial class (interval) – to observe the most frequent and ‘’the value in a set of ranked observations that divides the data into two parts of equal size” (Everitt & Skrondal, 2010);  Coefficient of variation - to measure “the spread for a set of data defined” (Everitt & Skrondal, 2010);  Confidence interval – to range the values, calculated from the sample observations, that is believed, with a particular probability, to contain the true parameter value (of the population);  Sampling error – to observe “the difference between the sample result and the population characteristic being estimated” (Everitt & Skrondal, 2010). To test the hypothesis the following data analysis methods were applied:  Significance tests for a population mean number for the result variable.  The t-test was used to measure statistically significant variations between IMS types. It was applied to test the hypothesis.  Calculating p-values for the given test statistics and the degrees of freedom. 4. Findings 4.1.Basic data characteristics – idea quantity Respondents frequency distribution based on survey data is shown in Figure 3. Figure 3. Respondents frequency distribution for idea quantity based on sample data Source: created by author’s The further detailed analysis consists of the arithmetic mean of the group, standard deviation, modal and medial class (group), coefficient of variation. Arithmetic mean of grouped data is calculated as follows: �̅�𝑥 = ∑ �� ∗ �� �� � (1) where 𝑚𝑚� - ith class (group) midpoint, 𝑓𝑓� - frequency of the ith class (interval), 𝑛𝑛 - sample size, 𝑛𝑛 = ∑ 𝑓𝑓��� . The standard deviation of grouped data is calculated as follows: 𝑠𝑠 = ���� ∗ �1 − � �� (2) where 𝑠𝑠� = ���� ∑ 𝑓𝑓� ∗ �𝑚𝑚� − �̅�𝑥� �� � 𝑚𝑚�- ith class (group) midpoint. 0 50 100 150 200 1-10 11-100 101-1000 1001-5000 5001-10000 >10000 Active IMS Pasive IMS Internal IMS External IMS Mixed IMS (1) where: mi - ith class (group) midpoint, ƒi - frequency of the ith class (interval), n - sample size, for the unknown parameter, the process is called point estimation. The process of providing a numerical value for a population parameter based on information collected from a sample. If an interval is calculated which is likely to contain the parameter, then the procedure is called interval estimation” (Everitt & Skrondal, 2010);  Frequency distribution – “the division of a sample of observations into a number of classes, together with the number of observations in each class. Acts as a useful summary of the main features of the data such as location, shape, and spread” (Everitt & Skrondal, 2010);  Mean of the group - to get the average value of the group;  Standard deviation – to measure the spread of a set of observations;  Modal and medial class (interval) – to observe the most frequent and ‘’the value in a set of ranked observations that divides the data into two parts of equal size” (Everitt & Skrondal, 2010);  Coefficient of variation - to measure “the spread for a set of data defined” (Everitt & Skrondal, 2010);  Confidence interval – to range the values, calculated from the sample observations, that is believed, with a particular probability, to contain the true parameter value (of the population);  Sampling error – to observe “the difference between the sample result and the population characteristic being estimated” (Everitt & Skrondal, 2010). To test the hypothesis the following data analysis methods were applied:  Significance tests for a population mean number for the result variable.  The t-test was used to measure statistically significant variations between IMS types. It was applied to test the hypothesis.  Calculating p-values for the given test statistics and the degrees of freedom. 4. Findings 4.1.Basic data characteristics – idea quantity Respondents frequency distribution based on survey data is shown in Figure 3. Figure 3. Respondents frequency distribution for idea quantity based on sample data Source: created by author’s The further detailed analysis consists of the arithmetic mean of the group, standard deviation, modal and medial class (group), coefficient of variation. Arithmetic mean of grouped data is calculated as follows: �̅�𝑥 = ∑ �� ∗ �� �� � (1) where 𝑚𝑚� - ith class (group) midpoint, 𝑓𝑓� - frequency of the ith class (interval), 𝑛𝑛 - sample size, 𝑛𝑛 = ∑ 𝑓𝑓��� . The standard deviation of grouped data is calculated as follows: 𝑠𝑠 = ���� ∗ �1 − � �� (2) where 𝑠𝑠� = ���� ∑ 𝑓𝑓� ∗ �𝑚𝑚� − �̅�𝑥� �� � 𝑚𝑚�- ith class (group) midpoint. 0 50 100 150 200 1-10 11-100 101-1000 1001-5000 5001-10000 >10000 Active IMS Pasive IMS Internal IMS External IMS Mixed IMS . 0 50 100 150 200 1-10 11-100 101-1000 1001-5000 5001-10000 >10000 Active IMS Pasive IMS Internal IMS External IMS Mixed IMS 0 20 40 60 80 100 120 140 160 Active IMS Pasive IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS External IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS Mixed IMS 0 20 40 60 80 100 120 140 External IMS Mixed IMS Source: created by author’s Figure 3 Respondents frequency distribution for idea quantity based on sample data 199 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 The standard deviation of grouped data is calculated as follows: for the unknown parameter, the process is called point estimation. The process of providing a numerical value for a population parameter based on information collected from a sample. If an interval is calculated which is likely to contain the parameter, then the procedure is called interval estimation” (Everitt & Skrondal, 2010);  Frequency distribution – “the division of a sample of observations into a number of classes, together with the number of observations in each class. Acts as a useful summary of the main features of the data such as location, shape, and spread” (Everitt & Skrondal, 2010);  Mean of the group - to get the average value of the group;  Standard deviation – to measure the spread of a set of observations;  Modal and medial class (interval) – to observe the most frequent and ‘’the value in a set of ranked observations that divides the data into two parts of equal size” (Everitt & Skrondal, 2010);  Coefficient of variation - to measure “the spread for a set of data defined” (Everitt & Skrondal, 2010);  Confidence interval – to range the values, calculated from the sample observations, that is believed, with a particular probability, to contain the true parameter value (of the population);  Sampling error – to observe “the difference between the sample result and the population characteristic being estimated” (Everitt & Skrondal, 2010). To test the hypothesis the following data analysis methods were applied:  Significance tests for a population mean number for the result variable.  The t-test was used to measure statistically significant variations between IMS types. It was applied to test the hypothesis.  Calculating p-values for the given test statistics and the degrees of freedom. 4. Findings 4.1.Basic data characteristics – idea quantity Respondents frequency distribution based on survey data is shown in Figure 3. Figure 3. Respondents frequency distribution for idea quantity based on sample data Source: created by author’s The further detailed analysis consists of the arithmetic mean of the group, standard deviation, modal and medial class (group), coefficient of variation. Arithmetic mean of grouped data is calculated as follows: �̅�𝑥 = ∑ �� ∗ �� �� � (1) where 𝑚𝑚� - ith class (group) midpoint, 𝑓𝑓� - frequency of the ith class (interval), 𝑛𝑛 - sample size, 𝑛𝑛 = ∑ 𝑓𝑓��� . The standard deviation of grouped data is calculated as follows: 𝑠𝑠 = ���� ∗ �1 − � �� (2) where 𝑠𝑠� = ���� ∑ 𝑓𝑓� ∗ �𝑚𝑚� − �̅�𝑥� �� � 𝑚𝑚�- ith class (group) midpoint. 0 50 100 150 200 1-10 11-100 101-1000 1001-5000 5001-10000 >10000 Active IMS Pasive IMS Internal IMS External IMS Mixed IMS (2) where: for the unknown parameter, the process is called point estimation. The process of providing a numerical value for a population parameter based on information collected from a sample. If an interval is calculated which is likely to contain the parameter, then the procedure is called interval estimation” (Everitt & Skrondal, 2010);  Frequency distribution – “the division of a sample of observations into a number of classes, together with the number of observations in each class. Acts as a useful summary of the main features of the data such as location, shape, and spread” (Everitt & Skrondal, 2010);  Mean of the group - to get the average value of the group;  Standard deviation – to measure the spread of a set of observations;  Modal and medial class (interval) – to observe the most frequent and ‘’the value in a set of ranked observations that divides the data into two parts of equal size” (Everitt & Skrondal, 2010);  Coefficient of variation - to measure “the spread for a set of data defined” (Everitt & Skrondal, 2010);  Confidence interval – to range the values, calculated from the sample observations, that is believed, with a particular probability, to contain the true parameter value (of the population);  Sampling error – to observe “the difference between the sample result and the population characteristic being estimated” (Everitt & Skrondal, 2010). To test the hypothesis the following data analysis methods were applied:  Significance tests for a population mean number for the result variable.  The t-test was used to measure statistically significant variations between IMS types. It was applied to test the hypothesis.  Calculating p-values for the given test statistics and the degrees of freedom. 4. Findings 4.1.Basic data characteristics – idea quantity Respondents frequency distribution based on survey data is shown in Figure 3. Figure 3. Respondents frequency distribution for idea quantity based on sample data Source: created by author’s The further detailed analysis consists of the arithmetic mean of the group, standard deviation, modal and medial class (group), coefficient of variation. Arithmetic mean of grouped data is calculated as follows: �̅�𝑥 = ∑ �� ∗ �� �� � (1) where 𝑚𝑚� - ith class (group) midpoint, 𝑓𝑓� - frequency of the ith class (interval), 𝑛𝑛 - sample size, 𝑛𝑛 = ∑ 𝑓𝑓��� . The standard deviation of grouped data is calculated as follows: 𝑠𝑠 = ���� ∗ �1 − � �� (2) where 𝑠𝑠� = ���� ∑ 𝑓𝑓� ∗ �𝑚𝑚� − �̅�𝑥� �� � 𝑚𝑚�- ith class (group) midpoint. 0 50 100 150 200 1-10 11-100 101-1000 1001-5000 5001-10000 >10000 Active IMS Pasive IMS Internal IMS External IMS Mixed IMS , mi - ith class (group) midpoint. The median (Me) of grouped data is calculated as follows: The median (𝑀𝑀𝑀𝑀) of grouped data is calculated as follows: 𝑀𝑀𝑀𝑀 = 𝑥𝑥��,� + �∑�� �������� ∗ ∆�� ��� (3) where 𝑥𝑥��,� - lower class boundary of the interval containing the median, 𝑐𝑐𝑐𝑐���� – cumulative frequency of the interval before the median interval, 𝑐𝑐�� - frequency of the median interval, ∆��- the median interval width. Medial interval is interval for which accumulated frequencies first time is equal or larger than half of the sample size. Coefficient of variation (CV) is calculated as follows: 𝐶𝐶𝐶𝐶 = ��̅ *100% (4) Point estimates were aggregated and are provided in Table 5. Table 5 Point estimates for idea quantity Source: created by author’s As shown in Table 5, the medians for all IMS types are less than the means of the generated ideas. These differences indicate some asymmetry in the distribution of respondents - more often a smaller number of ideas are generated, but less often - a large number of ideas. There is a particularly large difference between these indicators for passive IMS as well as for internal IMS - as frequency distributions are skewed mean values does not give a good idea of a typical value that can be expected in case of using these types of IMS. The calculated coefficients of variation also indicate similar - passive and internal IMS has more variation, relative to its arithmetic means than other IMS application types. Further described is the interval estimation for the population mean. The confidence interval for the population means µ is calculated as follows: �̅�𝑥 ± 𝑀𝑀𝑀𝑀 (5) where 𝑀𝑀𝑀𝑀 - margin error, 𝑀𝑀𝑀𝑀 = 𝑡𝑡���,�/� ∗ �√� (6) 𝑡𝑡���,�/�- value of t distribution for the selected confidence level and sample size, 𝛼𝛼 - level of significance and 100*(1- 𝛼𝛼)% - confidence interval. The upper confidence limit (UCL) is calculated as follows: UCL = �̅�𝑥 + 𝑀𝑀𝑀𝑀 (7) The lower confidence limit (LCL) is calculated as follows: LCL = �̅�𝑥 − 𝑀𝑀𝑀𝑀 (8) Confidence intervals (CI) provide the lower confidence limit (LCL) and the upper confidence limit (UCL) that are likely to contain the true parameter value (of the population). The value 95% refers to the probability that the interval will capture the parameter being estimated (Tan & Tan, 2010). 95% confidence interval estimates are aggregated in Table 6. Table 6 Mean values, margin errors and 95% confidence interval estimates for idea quantity IMS type n Mean Median s CV Active IMS 439 4138,6 3392,1 3500,6 85% Pasive IMS 306 1141,1 96,7 2313,7 203% Internal IMS 411 1284,3 396,1 2299,2 179% External IMS 371 4377,1 3875,5 3543,5 81% Mixed IMS 337 4420,1 3842,6 3771,4 85% IMS type Mean ME LCL UCL Active IMS 4138,6 167,1 3810,2 4466,9 Pasive IMS 1141,1 132,3 880,9 1401,4 Internal IMS 1284,3 113,4 1061,4 1507,3 External IMS 4377,1 184,0 4015,4 4738,9 Mixed IMS 4420,1 205,4 4016,0 4824,2 (3) where: xMe,l - lower class boundary of the interval containing the median, cfMe-1 – cumulative frequency of the interval before the median interval, fMe - frequency of the median interval, ΔMe - the median interval width. Medial interval is interval for which accumulated frequencies first time is equal or larger than half of the sample size. Coefficient of variation (CV) is calculated as follows: The median (𝑀𝑀𝑀𝑀) of grouped data is calculated as follows: 𝑀𝑀𝑀𝑀 = 𝑥𝑥��,� + �∑�� �������� ∗ ∆�� ��� (3) where 𝑥𝑥��,� - lower class boundary of the interval containing the median, 𝑐𝑐𝑐𝑐���� – cumulative frequency of the interval before the median interval, 𝑐𝑐�� - frequency of the median interval, ∆��- the median interval width. Medial interval is interval for which accumulated frequencies first time is equal or larger than half of the sample size. Coefficient of variation (CV) is calculated as follows: 𝐶𝐶𝐶𝐶 = ��̅ *100% (4) Point estimates were aggregated and are provided in Table 5. Table 5 Point estimates for idea quantity Source: created by author’s As shown in Table 5, the medians for all IMS types are less than the means of the generated ideas. These differences indicate some asymmetry in the distribution of respondents - more often a smaller number of ideas are generated, but less often - a large number of ideas. There is a particularly large difference between these indicators for passive IMS as well as for internal IMS - as frequency distributions are skewed mean values does not give a good idea of a typical value that can be expected in case of using these types of IMS. The calculated coefficients of variation also indicate similar - passive and internal IMS has more variation, relative to its arithmetic means than other IMS application types. Further described is the interval estimation for the population mean. The confidence interval for the population means µ is calculated as follows: �̅�𝑥 ± 𝑀𝑀𝑀𝑀 (5) where 𝑀𝑀𝑀𝑀 - margin error, 𝑀𝑀𝑀𝑀 = 𝑡𝑡���,�/� ∗ �√� (6) 𝑡𝑡���,�/�- value of t distribution for the selected confidence level and sample size, 𝛼𝛼 - level of significance and 100*(1- 𝛼𝛼)% - confidence interval. The upper confidence limit (UCL) is calculated as follows: UCL = �̅�𝑥 + 𝑀𝑀𝑀𝑀 (7) The lower confidence limit (LCL) is calculated as follows: LCL = �̅�𝑥 − 𝑀𝑀𝑀𝑀 (8) Confidence intervals (CI) provide the lower confidence limit (LCL) and the upper confidence limit (UCL) that are likely to contain the true parameter value (of the population). The value 95% refers to the probability that the interval will capture the parameter being estimated (Tan & Tan, 2010). 95% confidence interval estimates are aggregated in Table 6. Table 6 Mean values, margin errors and 95% confidence interval estimates for idea quantity IMS type n Mean Median s CV Active IMS 439 4138,6 3392,1 3500,6 85% Pasive IMS 306 1141,1 96,7 2313,7 203% Internal IMS 411 1284,3 396,1 2299,2 179% External IMS 371 4377,1 3875,5 3543,5 81% Mixed IMS 337 4420,1 3842,6 3771,4 85% IMS type Mean ME LCL UCL Active IMS 4138,6 167,1 3810,2 4466,9 Pasive IMS 1141,1 132,3 880,9 1401,4 Internal IMS 1284,3 113,4 1061,4 1507,3 External IMS 4377,1 184,0 4015,4 4738,9 Mixed IMS 4420,1 205,4 4016,0 4824,2 (4) Point estimates were aggregated and are provided in Table 5. Table 5 Point estimates for idea quantity IMS type n Mean Median S CV Active IMS 439 4138,6 3392,1 3500,6 85% Pasive IMS 306 1141,1 96,7 2313,7 203% Internal IMS 411 1284,3 396,1 2299,2 179% External IMS 371 4377,1 3875,5 3543,5 81% Mixed IMS 337 4420,1 3842,6 3771,4 85% Source: Authors’ constructions As shown in Table 5, the medians for all IMS types are less than the means of the generated ide- as. These differences indicate some asymmetry in the distribution of respondents - more often a smaller number of ideas are generated, but less often - a large number of ideas. There is a par- ticularly large difference between these indicators for passive IMS as well as for internal IMS - as frequency distributions are skewed mean values does not give a good idea of a typical value that can be expected in case of using these types of IMS. The calculated coefficients of variation also indicate similar - passive and internal IMS has more variation, relative to its arithmetic means than other IMS application types. Further described is the interval estimation for the population mean. The confidence interval for the population means µ is calculated as follows: The median (𝑀𝑀𝑀𝑀) of grouped data is calculated as follows: 𝑀𝑀𝑀𝑀 = 𝑥𝑥��,� + �∑�� �������� ∗ ∆�� ��� (3) where 𝑥𝑥��,� - lower class boundary of the interval containing the median, 𝑐𝑐𝑐𝑐���� – cumulative frequency of the interval before the median interval, 𝑐𝑐�� - frequency of the median interval, ∆��- the median interval width. Medial interval is interval for which accumulated frequencies first time is equal or larger than half of the sample size. Coefficient of variation (CV) is calculated as follows: 𝐶𝐶𝐶𝐶 = ��̅ *100% (4) Point estimates were aggregated and are provided in Table 5. Table 5 Point estimates for idea quantity Source: created by author’s As shown in Table 5, the medians for all IMS types are less than the means of the generated ideas. These differences indicate some asymmetry in the distribution of respondents - more often a smaller number of ideas are generated, but less often - a large number of ideas. There is a particularly large difference between these indicators for passive IMS as well as for internal IMS - as frequency distributions are skewed mean values does not give a good idea of a typical value that can be expected in case of using these types of IMS. The calculated coefficients of variation also indicate similar - passive and internal IMS has more variation, relative to its arithmetic means than other IMS application types. Further described is the interval estimation for the population mean. The confidence interval for the population means µ is calculated as follows: �̅�𝑥 ± 𝑀𝑀𝑀𝑀 (5) where 𝑀𝑀𝑀𝑀 - margin error, 𝑀𝑀𝑀𝑀 = 𝑡𝑡���,�/� ∗ �√� (6) 𝑡𝑡���,�/�- value of t distribution for the selected confidence level and sample size, 𝛼𝛼 - level of significance and 100*(1- 𝛼𝛼)% - confidence interval. The upper confidence limit (UCL) is calculated as follows: UCL = �̅�𝑥 + 𝑀𝑀𝑀𝑀 (7) The lower confidence limit (LCL) is calculated as follows: LCL = �̅�𝑥 − 𝑀𝑀𝑀𝑀 (8) Confidence intervals (CI) provide the lower confidence limit (LCL) and the upper confidence limit (UCL) that are likely to contain the true parameter value (of the population). The value 95% refers to the probability that the interval will capture the parameter being estimated (Tan & Tan, 2010). 95% confidence interval estimates are aggregated in Table 6. Table 6 Mean values, margin errors and 95% confidence interval estimates for idea quantity IMS type n Mean Median s CV Active IMS 439 4138,6 3392,1 3500,6 85% Pasive IMS 306 1141,1 96,7 2313,7 203% Internal IMS 411 1284,3 396,1 2299,2 179% External IMS 371 4377,1 3875,5 3543,5 81% Mixed IMS 337 4420,1 3842,6 3771,4 85% IMS type Mean ME LCL UCL Active IMS 4138,6 167,1 3810,2 4466,9 Pasive IMS 1141,1 132,3 880,9 1401,4 Internal IMS 1284,3 113,4 1061,4 1507,3 External IMS 4377,1 184,0 4015,4 4738,9 Mixed IMS 4420,1 205,4 4016,0 4824,2 (5) where: ME - margin error, E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 200 The median (𝑀𝑀𝑀𝑀) of grouped data is calculated as follows: 𝑀𝑀𝑀𝑀 = 𝑥𝑥��,� + �∑�� �������� ∗ ∆�� ��� (3) where 𝑥𝑥��,� - lower class boundary of the interval containing the median, 𝑐𝑐𝑐𝑐���� – cumulative frequency of the interval before the median interval, 𝑐𝑐�� - frequency of the median interval, ∆��- the median interval width. Medial interval is interval for which accumulated frequencies first time is equal or larger than half of the sample size. Coefficient of variation (CV) is calculated as follows: 𝐶𝐶𝐶𝐶 = ��̅ *100% (4) Point estimates were aggregated and are provided in Table 5. Table 5 Point estimates for idea quantity Source: created by author’s As shown in Table 5, the medians for all IMS types are less than the means of the generated ideas. These differences indicate some asymmetry in the distribution of respondents - more often a smaller number of ideas are generated, but less often - a large number of ideas. There is a particularly large difference between these indicators for passive IMS as well as for internal IMS - as frequency distributions are skewed mean values does not give a good idea of a typical value that can be expected in case of using these types of IMS. The calculated coefficients of variation also indicate similar - passive and internal IMS has more variation, relative to its arithmetic means than other IMS application types. Further described is the interval estimation for the population mean. The confidence interval for the population means µ is calculated as follows: �̅�𝑥 ± 𝑀𝑀𝑀𝑀 (5) where 𝑀𝑀𝑀𝑀 - margin error, 𝑀𝑀𝑀𝑀 = 𝑡𝑡���,�/� ∗ �√� (6) 𝑡𝑡���,�/�- value of t distribution for the selected confidence level and sample size, 𝛼𝛼 - level of significance and 100*(1- 𝛼𝛼)% - confidence interval. The upper confidence limit (UCL) is calculated as follows: UCL = �̅�𝑥 + 𝑀𝑀𝑀𝑀 (7) The lower confidence limit (LCL) is calculated as follows: LCL = �̅�𝑥 − 𝑀𝑀𝑀𝑀 (8) Confidence intervals (CI) provide the lower confidence limit (LCL) and the upper confidence limit (UCL) that are likely to contain the true parameter value (of the population). The value 95% refers to the probability that the interval will capture the parameter being estimated (Tan & Tan, 2010). 95% confidence interval estimates are aggregated in Table 6. Table 6 Mean values, margin errors and 95% confidence interval estimates for idea quantity IMS type n Mean Median s CV Active IMS 439 4138,6 3392,1 3500,6 85% Pasive IMS 306 1141,1 96,7 2313,7 203% Internal IMS 411 1284,3 396,1 2299,2 179% External IMS 371 4377,1 3875,5 3543,5 81% Mixed IMS 337 4420,1 3842,6 3771,4 85% IMS type Mean ME LCL UCL Active IMS 4138,6 167,1 3810,2 4466,9 Pasive IMS 1141,1 132,3 880,9 1401,4 Internal IMS 1284,3 113,4 1061,4 1507,3 External IMS 4377,1 184,0 4015,4 4738,9 Mixed IMS 4420,1 205,4 4016,0 4824,2 (6) where: tn – 1,α/2 - value of t distribution for the selected confidence level and sample size, α - level of significance and 100*(1– α)% - confidence interval. The upper confidence limit (UCL) is calculated as follows: The median (𝑀𝑀𝑀𝑀) of grouped data is calculated as follows: 𝑀𝑀𝑀𝑀 = 𝑥𝑥��,� + �∑�� �������� ∗ ∆�� ��� (3) where 𝑥𝑥��,� - lower class boundary of the interval containing the median, 𝑐𝑐𝑐𝑐���� – cumulative frequency of the interval before the median interval, 𝑐𝑐�� - frequency of the median interval, ∆��- the median interval width. Medial interval is interval for which accumulated frequencies first time is equal or larger than half of the sample size. Coefficient of variation (CV) is calculated as follows: 𝐶𝐶𝐶𝐶 = ��̅ *100% (4) Point estimates were aggregated and are provided in Table 5. Table 5 Point estimates for idea quantity Source: created by author’s As shown in Table 5, the medians for all IMS types are less than the means of the generated ideas. These differences indicate some asymmetry in the distribution of respondents - more often a smaller number of ideas are generated, but less often - a large number of ideas. There is a particularly large difference between these indicators for passive IMS as well as for internal IMS - as frequency distributions are skewed mean values does not give a good idea of a typical value that can be expected in case of using these types of IMS. The calculated coefficients of variation also indicate similar - passive and internal IMS has more variation, relative to its arithmetic means than other IMS application types. Further described is the interval estimation for the population mean. The confidence interval for the population means µ is calculated as follows: �̅�𝑥 ± 𝑀𝑀𝑀𝑀 (5) where 𝑀𝑀𝑀𝑀 - margin error, 𝑀𝑀𝑀𝑀 = 𝑡𝑡���,�/� ∗ �√� (6) 𝑡𝑡���,�/�- value of t distribution for the selected confidence level and sample size, 𝛼𝛼 - level of significance and 100*(1- 𝛼𝛼)% - confidence interval. The upper confidence limit (UCL) is calculated as follows: UCL = �̅�𝑥 + 𝑀𝑀𝑀𝑀 (7) The lower confidence limit (LCL) is calculated as follows: LCL = �̅�𝑥 − 𝑀𝑀𝑀𝑀 (8) Confidence intervals (CI) provide the lower confidence limit (LCL) and the upper confidence limit (UCL) that are likely to contain the true parameter value (of the population). The value 95% refers to the probability that the interval will capture the parameter being estimated (Tan & Tan, 2010). 95% confidence interval estimates are aggregated in Table 6. Table 6 Mean values, margin errors and 95% confidence interval estimates for idea quantity IMS type n Mean Median s CV Active IMS 439 4138,6 3392,1 3500,6 85% Pasive IMS 306 1141,1 96,7 2313,7 203% Internal IMS 411 1284,3 396,1 2299,2 179% External IMS 371 4377,1 3875,5 3543,5 81% Mixed IMS 337 4420,1 3842,6 3771,4 85% IMS type Mean ME LCL UCL Active IMS 4138,6 167,1 3810,2 4466,9 Pasive IMS 1141,1 132,3 880,9 1401,4 Internal IMS 1284,3 113,4 1061,4 1507,3 External IMS 4377,1 184,0 4015,4 4738,9 Mixed IMS 4420,1 205,4 4016,0 4824,2 (7) The lower confidence limit (LCL) is calculated as follows: The median (𝑀𝑀𝑀𝑀) of grouped data is calculated as follows: 𝑀𝑀𝑀𝑀 = 𝑥𝑥��,� + �∑�� �������� ∗ ∆�� ��� (3) where 𝑥𝑥��,� - lower class boundary of the interval containing the median, 𝑐𝑐𝑐𝑐���� – cumulative frequency of the interval before the median interval, 𝑐𝑐�� - frequency of the median interval, ∆��- the median interval width. Medial interval is interval for which accumulated frequencies first time is equal or larger than half of the sample size. Coefficient of variation (CV) is calculated as follows: 𝐶𝐶𝐶𝐶 = ��̅ *100% (4) Point estimates were aggregated and are provided in Table 5. Table 5 Point estimates for idea quantity Source: created by author’s As shown in Table 5, the medians for all IMS types are less than the means of the generated ideas. These differences indicate some asymmetry in the distribution of respondents - more often a smaller number of ideas are generated, but less often - a large number of ideas. There is a particularly large difference between these indicators for passive IMS as well as for internal IMS - as frequency distributions are skewed mean values does not give a good idea of a typical value that can be expected in case of using these types of IMS. The calculated coefficients of variation also indicate similar - passive and internal IMS has more variation, relative to its arithmetic means than other IMS application types. Further described is the interval estimation for the population mean. The confidence interval for the population means µ is calculated as follows: �̅�𝑥 ± 𝑀𝑀𝑀𝑀 (5) where 𝑀𝑀𝑀𝑀 - margin error, 𝑀𝑀𝑀𝑀 = 𝑡𝑡���,�/� ∗ �√� (6) 𝑡𝑡���,�/�- value of t distribution for the selected confidence level and sample size, 𝛼𝛼 - level of significance and 100*(1- 𝛼𝛼)% - confidence interval. The upper confidence limit (UCL) is calculated as follows: UCL = �̅�𝑥 + 𝑀𝑀𝑀𝑀 (7) The lower confidence limit (LCL) is calculated as follows: LCL = �̅�𝑥 − 𝑀𝑀𝑀𝑀 (8) Confidence intervals (CI) provide the lower confidence limit (LCL) and the upper confidence limit (UCL) that are likely to contain the true parameter value (of the population). The value 95% refers to the probability that the interval will capture the parameter being estimated (Tan & Tan, 2010). 95% confidence interval estimates are aggregated in Table 6. Table 6 Mean values, margin errors and 95% confidence interval estimates for idea quantity IMS type n Mean Median s CV Active IMS 439 4138,6 3392,1 3500,6 85% Pasive IMS 306 1141,1 96,7 2313,7 203% Internal IMS 411 1284,3 396,1 2299,2 179% External IMS 371 4377,1 3875,5 3543,5 81% Mixed IMS 337 4420,1 3842,6 3771,4 85% IMS type Mean ME LCL UCL Active IMS 4138,6 167,1 3810,2 4466,9 Pasive IMS 1141,1 132,3 880,9 1401,4 Internal IMS 1284,3 113,4 1061,4 1507,3 External IMS 4377,1 184,0 4015,4 4738,9 Mixed IMS 4420,1 205,4 4016,0 4824,2 (8) Confidence intervals (CI) provide the lower confidence limit (LCL) and the upper confidence limit (UCL) that are likely to contain the true parameter value (of the population). The value 95% refers to the probability that the interval will capture the parameter being estimated (Tan & Tan, 2010). 95% confidence interval estimates are aggregated in Table 6. Table 6 Mean values, margin errors and 95% confidence interval estimates for idea quantity IMS type Mean Me LCL UCL Active IMS 4138,6 167,1 3810,2 4466,9 Pasive IMS 1141,1 132,3 880,9 1401,4 Internal IMS 1284,3 113,4 1061,4 1507,3 External IMS 4377,1 184,0 4015,4 4738,9 Mixed IMS 4420,1 205,4 4016,0 4824,2 Source: created by author’s A 95% CI means that if the study will be conducted multiple times with corresponding 95% CI for the mean constructed, author’s expect 95% of these CI’s to contain the true population mean (Tan & Tan, 2010) and it could be between 3810 to 4467 ideas generated in active IMS, for passive IMS between 881 to 1401, for internal IMS between 1061 to 1507, external IMS between 4015 to 4739 and mixed IMS between 4016 to 4824. Hypothesis testing – idea quantity Basic data set analysis showed that it is possible to test the hypothesis on the gathered data. That is the reason why further in this paper the authors conduct significance tests for population mean number of ideas created (idea quantity). A respondent’s frequency distribution shows the main trends that will be tested: (1) active IMS provides higher idea quantity than passive IMS; (2) external IMS provides higher idea quantity than internal IMS; (3) mixed IMS provides higher idea quantity than internal and external IMS. See in Figure 8. Hypothesis tested: (H1) Active IMS provide higher idea quantity than passive: Source: created by author’s A 95% CI means that if the study will be conducted multiple times with corresponding 95% CI for the mean constructed, author’s expect 95% of these CI’s to contain the true population mean (Tan & Tan, 2010) and it could be between 3810 to 4467 ideas generated in active IMS, for passive IMS between 881 to 1401, for internal IMS between 1061 to 1507, external IMS between 4015 to 4739 and mixed IMS between 4016 to 4824. 4.2.Hypothesis testing – idea quantity Basic data set analysis showed that it is possible to test the hypothesis on the gathered data. That is the reason why further in this paper the authors conduct significance tests for population mean number of ideas created (idea quantity). A respondent’s frequency distribution shows the main trends that will be tested: (1) active IMS provides higher idea quantity than passive IMS; (2) external IMS provides higher idea quantity than internal IMS; (3) mixed IMS provides higher idea quantity than internal and external IMS. See in Figure 8. Figure 8. Respondents frequency distribution for idea quantity based on sample data Source: created by author’s Hypothesis tested: (H1) Active IMS provide higher idea quantity than passive: H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H2) External IMS provide higher idea quantity than internal: H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H3) Mixed IMS provide higher idea quantity than internal H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H4) Mixed IMS provide higher idea quantity than external H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 As sample sizes un standard deviations differs, the test statistics are calculated as follows, (Moore et al. 2016): 0 20 40 60 80 100 120 140 160 Active IMS Pasive IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS External IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS Mixed IMS 0 20 40 60 80 100 120 140 External IMS Mixed IMS (H2) External IMS provide higher idea quantity than internal: Source: created by author’s A 95% CI means that if the study will be conducted multiple times with corresponding 95% CI for the mean constructed, author’s expect 95% of these CI’s to contain the true population mean (Tan & Tan, 2010) and it could be between 3810 to 4467 ideas generated in active IMS, for passive IMS between 881 to 1401, for internal IMS between 1061 to 1507, external IMS between 4015 to 4739 and mixed IMS between 4016 to 4824. 4.2.Hypothesis testing – idea quantity Basic data set analysis showed that it is possible to test the hypothesis on the gathered data. That is the reason why further in this paper the authors conduct significance tests for population mean number of ideas created (idea quantity). A respondent’s frequency distribution shows the main trends that will be tested: (1) active IMS provides higher idea quantity than passive IMS; (2) external IMS provides higher idea quantity than internal IMS; (3) mixed IMS provides higher idea quantity than internal and external IMS. See in Figure 8. Figure 8. Respondents frequency distribution for idea quantity based on sample data Source: created by author’s Hypothesis tested: (H1) Active IMS provide higher idea quantity than passive: H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H2) External IMS provide higher idea quantity than internal: H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H3) Mixed IMS provide higher idea quantity than internal H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H4) Mixed IMS provide higher idea quantity than external H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 As sample sizes un standard deviations differs, the test statistics are calculated as follows, (Moore et al. 2016): 0 20 40 60 80 100 120 140 160 Active IMS Pasive IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS External IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS Mixed IMS 0 20 40 60 80 100 120 140 External IMS Mixed IMS (H3) Mixed IMS provide higher idea quantity than internal Source: created by author’s A 95% CI means that if the study will be conducted multiple times with corresponding 95% CI for the mean constructed, author’s expect 95% of these CI’s to contain the true population mean (Tan & Tan, 2010) and it could be between 3810 to 4467 ideas generated in active IMS, for passive IMS between 881 to 1401, for internal IMS between 1061 to 1507, external IMS between 4015 to 4739 and mixed IMS between 4016 to 4824. 4.2.Hypothesis testing – idea quantity Basic data set analysis showed that it is possible to test the hypothesis on the gathered data. That is the reason why further in this paper the authors conduct significance tests for population mean number of ideas created (idea quantity). A respondent’s frequency distribution shows the main trends that will be tested: (1) active IMS provides higher idea quantity than passive IMS; (2) external IMS provides higher idea quantity than internal IMS; (3) mixed IMS provides higher idea quantity than internal and external IMS. See in Figure 8. Figure 8. Respondents frequency distribution for idea quantity based on sample data Source: created by author’s Hypothesis tested: (H1) Active IMS provide higher idea quantity than passive: H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H2) External IMS provide higher idea quantity than internal: H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H3) Mixed IMS provide higher idea quantity than internal H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H4) Mixed IMS provide higher idea quantity than external H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 As sample sizes un standard deviations differs, the test statistics are calculated as follows, (Moore et al. 2016): 0 20 40 60 80 100 120 140 160 Active IMS Pasive IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS External IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS Mixed IMS 0 20 40 60 80 100 120 140 External IMS Mixed IMS (H4) Mixed IMS provide higher idea quantity than external Source: created by author’s A 95% CI means that if the study will be conducted multiple times with corresponding 95% CI for the mean constructed, author’s expect 95% of these CI’s to contain the true population mean (Tan & Tan, 2010) and it could be between 3810 to 4467 ideas generated in active IMS, for passive IMS between 881 to 1401, for internal IMS between 1061 to 1507, external IMS between 4015 to 4739 and mixed IMS between 4016 to 4824. 4.2.Hypothesis testing – idea quantity Basic data set analysis showed that it is possible to test the hypothesis on the gathered data. That is the reason why further in this paper the authors conduct significance tests for population mean number of ideas created (idea quantity). A respondent’s frequency distribution shows the main trends that will be tested: (1) active IMS provides higher idea quantity than passive IMS; (2) external IMS provides higher idea quantity than internal IMS; (3) mixed IMS provides higher idea quantity than internal and external IMS. See in Figure 8. Figure 8. Respondents frequency distribution for idea quantity based on sample data Source: created by author’s Hypothesis tested: (H1) Active IMS provide higher idea quantity than passive: H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H2) External IMS provide higher idea quantity than internal: H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H3) Mixed IMS provide higher idea quantity than internal H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 (H4) Mixed IMS provide higher idea quantity than external H0: �̅�𝑥�� − �̅�𝑥�� ≤ 0 and HA: �̅�𝑥�� − �̅�𝑥��> 0 As sample sizes un standard deviations differs, the test statistics are calculated as follows, (Moore et al. 2016): 0 20 40 60 80 100 120 140 160 Active IMS Pasive IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS External IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS Mixed IMS 0 20 40 60 80 100 120 140 External IMS Mixed IMS 201 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 0 50 100 150 200 1-10 11-100 101-1000 1001-5000 5001-10000 >10000 Active IMS Pasive IMS Internal IMS External IMS Mixed IMS 0 20 40 60 80 100 120 140 160 Active IMS Pasive IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS External IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS Mixed IMS 0 20 40 60 80 100 120 140 External IMS Mixed IMS Figure 8 Respondents frequency distribution for idea quantity based on sample data 0 50 100 150 200 1-10 11-100 101-1000 1001-5000 5001-10000 >10000 Active IMS Pasive IMS Internal IMS External IMS Mixed IMS 0 20 40 60 80 100 120 140 160 Active IMS Pasive IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS External IMS 0 20 40 60 80 100 120 140 160 180 Internal IMS Mixed IMS 0 20 40 60 80 100 120 140 External IMS Mixed IMS Source: created by author’s As sample sizes un standard deviations differs, the test statistics are calculated as follows, (Moore et al. 2016): 𝑡𝑡 = �̅� � �̅� ��� � �� � �� � �� (9) where �̅�𝑥� and �̅�𝑥� means of comparable sample variables, 𝑠𝑠�� and 𝑠𝑠�� variance (standard deviation squared) of comparable samples variables, 𝑛𝑛� and 𝑛𝑛� sample sizes of comparable samples and compared with t-statistics critical values 𝑡𝑡�,�, where degrees of freedom (𝑑𝑑𝑑𝑑) for the t-test statistics are calculated as follows: 𝑑𝑑𝑑𝑑 = ���� � �� � � � �� � �� �� � � �� � �� � � ������ � � �� � �� � � ������ (10) Also, p-values were calculated for given test statistics and the degrees of freedom. The p-value is the probability of obtaining a value of the test statistic as extreme as or more extreme than the actual value obtained when the null hypothesis is true. Thus, the p-value is the smallest significance level at which a null hypothesis can be rejected, given the observed sample statistic. Calculated t-statistics, degrees of freedom (df), critical values (tc) and p-values are aggregated in following Table 7. Table 7 Hypothesis test statistics for idea quantity Source: created by author’s From test results, authors would reject the null hypothesis 1, 2, 3 and conclude that sample data provide strong evidence to support conclusions that: (H1) Active IMS provide higher idea quantity than passive; (H2) External IMS provide higher idea quantity than internal; (H3) Mixed IMS provide higher idea quantity than internal. These conclusions are supported by very low p-values (<0,0001). Regarding H4 authors cannot reject the null hypothesis because t < tc and p-value >0,05 – so sample data does not give sufficient evidence that mixed IMS provide higher idea quantity than external. 5. Conclusions  There is strong statistical evidence to support conclusions that: o Active IMS provide higher idea quantity than passive. o External IMS provide higher idea quantity than internal. o Mixed IMS provide higher idea quantity than internal. o Paper does not have sufficient evidence that mixed IMS provide higher idea quantity than external.  This research fulfils an identified need to clarify IMS types and their impact on the results – idea quantity. This research delivers the following academical contribution: ‘ o (1) it is the widest web-based IMS empirical research based on the survey; o (2) approbated classifications of IMS; o (3) it researches different classifications of IMS and their impact on idea quantity.  The practical contribution of the research results helps to understand what kind of results enterprises could expect from different IMS application types.  Research results highlight the benefits/implications of adopting different types of IMS for organizations. These contributions also provide managers with a richer set of theoretical tools, Pairs tested t df tc p -value AIMS vs PIMS 14,07 741,05 1,96 <0,0001 IIMS vs EIMS 14,31 623,41 1,96 <0,0001 IIMS vs MIMS 13,36 531,54 1,96 <0,0001 EIMS vs MIMS 0,16 688,74 1,96 >0,05 (9) where 𝑡𝑡 = �̅� � �̅� ��� � �� � �� � �� (9) where �̅�𝑥� and �̅�𝑥� means of comparable sample variables, 𝑠𝑠�� and 𝑠𝑠�� variance (standard deviation squared) of comparable samples variables, 𝑛𝑛� and 𝑛𝑛� sample sizes of comparable samples and compared with t-statistics critical values 𝑡𝑡�,�, where degrees of freedom (𝑑𝑑𝑑𝑑) for the t-test statistics are calculated as follows: 𝑑𝑑𝑑𝑑 = ���� � �� � � � �� � �� �� � � �� � �� � � ������ � � �� � �� � � ������ (10) Also, p-values were calculated for given test statistics and the degrees of freedom. The p-value is the probability of obtaining a value of the test statistic as extreme as or more extreme than the actual value obtained when the null hypothesis is true. Thus, the p-value is the smallest significance level at which a null hypothesis can be rejected, given the observed sample statistic. Calculated t-statistics, degrees of freedom (df), critical values (tc) and p-values are aggregated in following Table 7. Table 7 Hypothesis test statistics for idea quantity Source: created by author’s From test results, authors would reject the null hypothesis 1, 2, 3 and conclude that sample data provide strong evidence to support conclusions that: (H1) Active IMS provide higher idea quantity than passive; (H2) External IMS provide higher idea quantity than internal; (H3) Mixed IMS provide higher idea quantity than internal. These conclusions are supported by very low p-values (<0,0001). Regarding H4 authors cannot reject the null hypothesis because t < tc and p-value >0,05 – so sample data does not give sufficient evidence that mixed IMS provide higher idea quantity than external. 5. Conclusions  There is strong statistical evidence to support conclusions that: o Active IMS provide higher idea quantity than passive. o External IMS provide higher idea quantity than internal. o Mixed IMS provide higher idea quantity than internal. o Paper does not have sufficient evidence that mixed IMS provide higher idea quantity than external.  This research fulfils an identified need to clarify IMS types and their impact on the results – idea quantity. This research delivers the following academical contribution: ‘ o (1) it is the widest web-based IMS empirical research based on the survey; o (2) approbated classifications of IMS; o (3) it researches different classifications of IMS and their impact on idea quantity.  The practical contribution of the research results helps to understand what kind of results enterprises could expect from different IMS application types.  Research results highlight the benefits/implications of adopting different types of IMS for organizations. These contributions also provide managers with a richer set of theoretical tools, Pairs tested t df tc p -value AIMS vs PIMS 14,07 741,05 1,96 <0,0001 IIMS vs EIMS 14,31 623,41 1,96 <0,0001 IIMS vs MIMS 13,36 531,54 1,96 <0,0001 EIMS vs MIMS 0,16 688,74 1,96 >0,05 and 𝑡𝑡 = �̅� � �̅� ��� � �� � �� � �� (9) where �̅�𝑥� and �̅�𝑥� means of comparable sample variables, 𝑠𝑠�� and 𝑠𝑠�� variance (standard deviation squared) of comparable samples variables, 𝑛𝑛� and 𝑛𝑛� sample sizes of comparable samples and compared with t-statistics critical values 𝑡𝑡�,�, where degrees of freedom (𝑑𝑑𝑑𝑑) for the t-test statistics are calculated as follows: 𝑑𝑑𝑑𝑑 = ���� � �� � � � �� � �� �� � � �� � �� � � ������ � � �� � �� � � ������ (10) Also, p-values were calculated for given test statistics and the degrees of freedom. The p-value is the probability of obtaining a value of the test statistic as extreme as or more extreme than the actual value obtained when the null hypothesis is true. Thus, the p-value is the smallest significance level at which a null hypothesis can be rejected, given the observed sample statistic. Calculated t-statistics, degrees of freedom (df), critical values (tc) and p-values are aggregated in following Table 7. Table 7 Hypothesis test statistics for idea quantity Source: created by author’s From test results, authors would reject the null hypothesis 1, 2, 3 and conclude that sample data provide strong evidence to support conclusions that: (H1) Active IMS provide higher idea quantity than passive; (H2) External IMS provide higher idea quantity than internal; (H3) Mixed IMS provide higher idea quantity than internal. These conclusions are supported by very low p-values (<0,0001). Regarding H4 authors cannot reject the null hypothesis because t < tc and p-value >0,05 – so sample data does not give sufficient evidence that mixed IMS provide higher idea quantity than external. 5. Conclusions  There is strong statistical evidence to support conclusions that: o Active IMS provide higher idea quantity than passive. o External IMS provide higher idea quantity than internal. o Mixed IMS provide higher idea quantity than internal. o Paper does not have sufficient evidence that mixed IMS provide higher idea quantity than external.  This research fulfils an identified need to clarify IMS types and their impact on the results – idea quantity. This research delivers the following academical contribution: ‘ o (1) it is the widest web-based IMS empirical research based on the survey; o (2) approbated classifications of IMS; o (3) it researches different classifications of IMS and their impact on idea quantity.  The practical contribution of the research results helps to understand what kind of results enterprises could expect from different IMS application types.  Research results highlight the benefits/implications of adopting different types of IMS for organizations. These contributions also provide managers with a richer set of theoretical tools, Pairs tested t df tc p -value AIMS vs PIMS 14,07 741,05 1,96 <0,0001 IIMS vs EIMS 14,31 623,41 1,96 <0,0001 IIMS vs MIMS 13,36 531,54 1,96 <0,0001 EIMS vs MIMS 0,16 688,74 1,96 >0,05 means of comparable sample variables, si 2 and sj2 variance (standard deviation squared) of comparable samples variables, ni and nj sample sizes of comparable samples and compared with t-statistics critical values tν,α, where degrees of freedom (df) for the t-test statistics are calculated as follows: 𝑡𝑡 = �̅� � �̅� ��� � �� � �� � �� (9) where �̅�𝑥� and �̅�𝑥� means of comparable sample variables, 𝑠𝑠�� and 𝑠𝑠�� variance (standard deviation squared) of comparable samples variables, 𝑛𝑛� and 𝑛𝑛� sample sizes of comparable samples and compared with t-statistics critical values 𝑡𝑡�,�, where degrees of freedom (𝑑𝑑𝑑𝑑) for the t-test statistics are calculated as follows: 𝑑𝑑𝑑𝑑 = ���� � �� � � � �� � �� �� � � �� � �� � � ������ � � �� � �� � � ������ (10) Also, p-values were calculated for given test statistics and the degrees of freedom. The p-value is the probability of obtaining a value of the test statistic as extreme as or more extreme than the actual value obtained when the null hypothesis is true. Thus, the p-value is the smallest significance level at which a null hypothesis can be rejected, given the observed sample statistic. Calculated t-statistics, degrees of freedom (df), critical values (tc) and p-values are aggregated in following Table 7. Table 7 Hypothesis test statistics for idea quantity Source: created by author’s From test results, authors would reject the null hypothesis 1, 2, 3 and conclude that sample data provide strong evidence to support conclusions that: (H1) Active IMS provide higher idea quantity than passive; (H2) External IMS provide higher idea quantity than internal; (H3) Mixed IMS provide higher idea quantity than internal. These conclusions are supported by very low p-values (<0,0001). Regarding H4 authors cannot reject the null hypothesis because t < tc and p-value >0,05 – so sample data does not give sufficient evidence that mixed IMS provide higher idea quantity than external. 5. Conclusions  There is strong statistical evidence to support conclusions that: o Active IMS provide higher idea quantity than passive. o External IMS provide higher idea quantity than internal. o Mixed IMS provide higher idea quantity than internal. o Paper does not have sufficient evidence that mixed IMS provide higher idea quantity than external.  This research fulfils an identified need to clarify IMS types and their impact on the results – idea quantity. This research delivers the following academical contribution: ‘ o (1) it is the widest web-based IMS empirical research based on the survey; o (2) approbated classifications of IMS; o (3) it researches different classifications of IMS and their impact on idea quantity.  The practical contribution of the research results helps to understand what kind of results enterprises could expect from different IMS application types.  Research results highlight the benefits/implications of adopting different types of IMS for organizations. These contributions also provide managers with a richer set of theoretical tools, Pairs tested t df tc p -value AIMS vs PIMS 14,07 741,05 1,96 <0,0001 IIMS vs EIMS 14,31 623,41 1,96 <0,0001 IIMS vs MIMS 13,36 531,54 1,96 <0,0001 EIMS vs MIMS 0,16 688,74 1,96 >0,05 (10) lso, p-values were calculated for given test statistics and the degrees of freedom. The p-value is the probability of obtaining a value of the test statistic as extreme as or more extreme than the actual value obtained when the null hypothesis is true. Thus, the p-value is the smallest significance level at which a null hypothesis can be rejected, given the observed sample statistic. Calculated t-statistics, degrees of freedom (df), critical values (tc) and p-values are aggregated in following Table 7. E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 202 Table 7 Hypothesis test statistics for idea quantity Pairs tested t df tc p- value AIMS vs PIMS 14,07 741,05 1,96 <0,0001 IIMS vs EIMS 14,31 623,41 1,96 <0,0001 IIMS vs MIMS 13,36 531,54 1,96 <0,0001 EIMS vs MIMS 0,16 688,74 1,96 >0,05 Source: created by author’s From test results, authors would reject the null hypothesis 1, 2, 3 and conclude that sample data provide strong evidence to support conclusions that: _ (H1) Active IMS provide higher idea quantity than passive; _ (H2) External IMS provide higher idea quantity than internal; _ (H3) Mixed IMS provide higher idea quantity than internal. These conclusions are supported by very low p-values (<0,0001). Regarding H4 authors cannot reject the null hypothesis because t < t c and p-value >0,05 – so sam- ple data does not give sufficient evidence that mixed IMS provide higher idea quantity than external. Conclusions There is strong statistical evidence to support conclusions that: _ Active IMS provide higher idea quantity than passive. _ External IMS provide higher idea quantity than internal. _ Mixed IMS provide higher idea quantity than internal. _ Paper does not have sufficient evidence that mixed IMS provide higher idea quantity than external. This research fulfils an identified need to clarify IMS types and their impact on the results – idea quantity. This research delivers the following academical contribution: ‘ 1 it is the widest web-based IMS empirical research based on the survey; 2 approbated classifications of IMS; 3 it researches different classifications of IMS and their impact on idea quantity. _ The practical contribution of the research results helps to understand what kind of results en- terprises could expect from different IMS application types. _ Research results highlight the benefits/implications of adopting different types of IMS for organ- izations. These contributions also provide managers with a richer set of theoretical tools, ena- bling them to make better decisions regarding the selection of IMS that are the best for achieving the results in the given context. Web-based IMS types and their impact on the IMS results could help to evaluate the potential application of these systems in different application scenarios. _ This research concentrated only on commercially available web-based IMS, but further re- search could include insights from non-commercially/private IMS, as well as real-life IMS. Additional research should be done to compare real-life and web-based IMS. Also, detailed research is needed to explore why standard deviations and coefficient of variations are so high. _ Further research should provide evidence on what benefits the different classes of IMS provide to organizations and also on other IMS results, such as, idea quality and involvement. This also coincides with van den Ende et. al. (2015) call to research different IMS types and their results. The research delivered by authors is the first attempt to answer this question, but there are a lot of additional questions to be answered. The authors are convinced that this paper will attract the interest of more researchers regarding IMS types and their impact on its results. 203 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 ReferencesAagaard, A. (2012). Idea Management in support of Pharmaceutical Front End of Innovation. Interna- tional Journal of Technology, Policy and Manage- ment, 12(4), 373-386. https://doi.org/10.1504/IJT- PM.2012.050138 Aagaard, A. (2013). A theoretical model of support- ing open source front end innovation through idea management. International Journal of Business In- novation and Research, 7(4), 446-465. https://doi. org/10.1504/IJBIR.2013.054869 Applegate, L.M. (1986). Idea management in organi- zation planning (brainstorming, strategy). University of Arizona: Tucson, Arizona. Baez, M. & Convertino, G. (2012). Innovation Cockpit: A Dashboard for Facilitators in Idea Management. In ACM 2012 conference proceedings of the internation- al conference in Seattle, Washington, 2012, ACM: NY, 47-48. https://doi.org/10.1145/2141512.2141536 Bailey, B.P. & Horvitz, E. (2010). What's Your Idea? A Case Study of a Grassroots Innovation Pipe- line within a Large Software Company. In CHI2010 proceedings on the 28th annual CHI conference on human factors in computing systems in Atlan- ta, USA, 2010, ACM: NY, 2065-2074. https://doi. org/10.1145/1753326.1753641 Bakker, H., Boersma, K. & Oreel, S. (2006). Creativ- ity (Ideas) Management in Industrial R&D Orga- nizations: A Crea- Political Process Model and an Empirical Illustration of Corus RD&T. Creativity and Innovation Management, 15(3), 296- 309. https://doi. org/10.1111/j.1467-8691.2006.00397.x Bansemir, B. & Neyer, A.K. (2009). From idea manage- ment systems to interactive innovation management systems: Designing for interaction and knowledge exchange. Retrieved June14, 2020, from https://aisel. aisnet.org/wi2009/3 Barczak, G., Griffin, A. & Kahn, B.K. (2009). PERSPEC- TIVE: Trends and Drivers of Success in NPD Practices: Results of the 2003 PDMA Best Practices Study. Jour- nal of Product Innovation Management, 26(1), 3-23. https://doi.org/10.1111/j.1540-5885.2009.00331.x Bassiti, L. & Ajhoun, R. (2013). Toward an Innovation Management Framework: A Life-cycle Model with an Idea Management focus. International Journal of Innovation, Management and Technology, 4 (6), 551- 559. https://doi.org/10.7763/IJIMT.2013.V4.460 Beretta, M. (2015). The Role of Idea Management Sys- tems for Innovation in Large Organizations: 3 essays. Aarhus University: Denmark, Aarhus. Bergendah, M. & Magnusson, M. (2014). Combining collaboration and competition: a key to improved idea management. European Journal of International Management, 8(5), 528-547. https://doi.org/10.1504/ EJIM.2014.064603 Bettoni, M., Bernhard, W., Eggs, C. & Schiller, G. (2010). Idea Management by Role Based Networked Learning. In 11th European conference on knowledge manage- ment proceedings of the international conference in Universidade Lusíada de Vila Nova de Famalicão, Por- tugal, Academic Publishing Limited: London, 107-116. Bjork, J. & Magnusson, M. (2009). Where Do Good In- novation Ideas Come From? Exploring the Influence of Network Connectivity on Innovation Idea Quality. Jour- nal of Product Innovation Management, 26 (1), 662-670. https://doi.org/10.1111/j.1540-5885.2009.00691.x Bothos, E., Apostolou, D. & Mentzas, G. (2008). A Col- laborative Information Aggregation System for Idea Management. In Conference on Internet and Web Ap- plications and Sevices proceedings of the internation- al conference in Athens, Greece, 2008, IEEE: Athens, 289-296. https://doi.org/10.1109/ICIW.2008.107 Bothos, E., Apostolou, D. & Mentzas, G. (2009). Collective intelligence for idea management with Internet-based information aggregation mar- kets. Internet Research, 19(1), 26-41. https://doi. org/10.1108/10662240910927803 Bothos, E., Apostoulou, D. & Mentzas, G. (2012). Collective intelligence with web-based information aggregation markets: The role of market facilitation in idea management. Experts Systems with Appli- cations, 39(1),1333-1345. https://doi.org/10.1016/j. eswa.2011.08.014 BrainBank. (2014). BrucePower. Retrieved June14, 2020, from http://www.brainbankinc.com/case-stud- ies/bruce-power Brem, A. & Voigt, K.I. (2007). Innovation management in emerging technology ventures - the concept of an integrated idea management. Journal of Technology, Policy and Management, 7(3), 304 - 321. https://doi. org/10.1504/IJTPM.2007.015113 Brem, A., & Voigt, K.I. (2009). Integration of market pull and technology push in the corporate front end and innovation management - Insights from the Ger- man software industry. Technovation, 29(1), 351-367. https://doi.org/10.1016/j.technovation.2008.06.003 BrightIdea. (2010). BTGroup. Retrieved June14, 2020, from http://www.brightidea.com/pdf_casestudy/ casestudy_bt.pdf E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 204 Coughlan, T. & Johnson, P. (2008). Idea Management in Creative Lives. In Conference on Human Factors in Computing Systems proceedings of the internation- al conference, Florence, Italy, 2008, ACM: NY, 3081- 3086. https://doi.org/10.1145/1358628.1358811 Crowdcity. (2018). Costumers. Retrieved June14, 2020, from https://crowdicity.com/customer-stories Deichmann, D. (2012). Idea Management: Perspec- tives from Leadership, Learning, and Network Theo- ry. ERIM: Rotterdam. DeSanctis, G. & Poole, M.S. (1994). Capturing the Complexity in Advanced Technology Use: Adaptive Structuration Theory. Organization Science, 5(2),121- 147. https://doi.org/10.1287/orsc.5.2.121 Divakaran, K.P.P. (2016). When users offer up ideas: how to evaluate them. Journal of Business Strategy, 37(5), 32-38. https://doi.org/10.1108/JBS-05-2015- 0046 Enkel, E., Grassmann, O. & Chesbrough, H. (2009). Open R&D and open innovation: exploring the phe- nomenon. R&D Management, 39(4), 311-316. https:// doi.org/10.1111/j.1467-9310.2009.00570.x Everitt B.S. & Skrondal, A. (2010). The Cambridge Dic- tionary of Statistics. Cambridge University Press: Edin- burgh. https://doi.org/10.1017/CBO9780511779633 Fatur, P. & Likar, P. (2009). The development of a performance measurement methodology for idea management. International Journal of Innovation and Learning, 6(4), 422-437. https://doi.org/10.1504/ IJIL.2009.024137 Flynn, M., Dooley, L., Sullivan, D. & Cormican, K. (2003). Idea management for organizational innovation. In- ternational Journal of Innovation Management, 7(4), 1-26. https://doi.org/10.1142/S1363919603000878 Fritz, M. (2002). Idea management enabler. EContent, 25(8), 50. Galbraith, J.R. (1982). Designing the innovating or- ganization. Organizational Dynamics, 10(3), 5-25. https://doi.org/10.1016/0090-2616(82)90033-X Gamlin, J.N., Yourd, R. & Patric, V. (2007). Unlock creativity with ''active'' idea management. Research Technology Management, 50(1), 13-16. https://doi.or g/10.1080/08956308.2007.11657413 Gish, L. (2011). Experience with idea promoting initia- tives: why they don't always work. In 18th Conference on Engineering Design: Impacting Society through Engineering Design proceedings of the international conference, Lyngby/Copenhagen, Denmark, 2011, Copenhagen: Design Society, 83-92. Glassmann, B.S. (2009). Improving idea generation and idea management in order to better manage the fuzzy front end of innovation, Prude University: West Lafayette. Green, S.G., Bean, S.A. & Snavely, B.K. (1983). Idea management in R&D as a human information pro- cessing analog. Human Systems Management, 4(2), 98-112. https://doi.org/10.3233/HSM-1983-4205 Holzblatt, L. & Tierney, L.M. (2011). Measuring the ef- fectiveness of social media on an innovation process. In CHI EA '11 proceedings, Vancouver, BC, ACM: NY, 697-712. https://doi.org/10.1145/1979742.1979669 Hrastinski, S., Kviselius, N.Z., Ozan, H. & Edenius, M. (2010). A review of technologies for open innovation: Characterastics and future trends. In HICSS '10 Pro- ceedings of the 2010 43rd Hawaii International Con- ference on System Sciences, Honolulu, HI, 2010, IEEE, Washington, DC: USA, 1-10. https://doi.org/10.1109/ HICSS.2010.29 Iversen, H., Kristensen, K., Liland, C.S., Berman, T., Enger, N. & Losnedahl, T. (2009). Idea Management: A Life- cycle Perspective on Innovation. Retrieved June14, 2020, from http://www.ict-21.ch/com-ict/ IMG/pdf/39_IdeaManagement_Kjetil_Kristensen_FI- NAL.pdf https://doi.org/10.1109/ITMC.2009.7461410 Klein, D. & Lechner, U. (2010). Ideenmanagement im Rahmen von organisatorischem Wandel. In MKWI 2010-IKT-gestutzte Unternehmens kommunikation proceedings, Gottingen, 2010, Universitatsverlag Got- tingen: Gottingen, 1831-1842. Korde, R. & Paulus, P.B. (2016). Alternating individual and group idea generation: Finding the elusive syner- gy. Journal of Experimental Social Psychology, 70(1), 177-190. https://doi.org/10.1016/j.jesp.2016.11.002 Lowe, M. & Heller, J.E. (2014). PLM Reference Model for Integrated Idea and Innovation Management. Prod- uct Lifecycle Management for a Global Market, IFIP Advances in Information and Communication Tech- nology, 442(1), 257-266. https://doi.org/10.1007/978- 3-662-45937-9_26 Luo, L. & Toubia, O. (2015). Improving Online Idea Generation Platforms and Customizing the Task Structure on the Basis of Consumers' Domain-Spe- cific Knowledge. Journal of Marketing, 79(5), 100-114. https://doi.org/10.1509/jm.13.0212 Mikelsone, E. & Liela, E. (2015). Discussion on the Terms of Idea Management and Idea Manage- 205 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 ment Systems. Journal of Regional Formation and Development Studies, 3(17), 97-110. https://doi. org/10.15181/rfds.v16i3.1124 Mikelsone, E., Volkova, T. & Liela, E. (2019). Practi- cal Evidence Of Web-Based Idea Management Sys- tems: Classification And Application. Unpublished Manuscript, Riga: Latvia. https://doi.org/10.22616/ rrd.25.2019.080 Moore, D., McCabe, G., Alwan, L., & Craig, B. (2016). The practice of statistics for business and economics. 4th ed. W.H.Freeman & Co, New York. Moss, B., Beimborn, D., Wagner, H.T. & Weitzel, T. (2011). The role of innovation governance and knowl- edge management for innovation success. In HICSS '10 Proceedings of the 2010 43rd Hawaii Interna- tional Conference on System Sciences, Kauai, HI, USA,2011, IEEE, Washington, DC: USA, 1-10. https:// doi.org/10.1109/HICSS.2011.429 Naravaez, J.L.M. & Gardoni, M. (2015). Harnessing idea management in the process of technology trans- fer at Canadian Space Agency. International Journal on Interactive Design and Manufacturing, 9 (33), 1-11. Nilsson L., Elg, M. & Bergman, B. (2002). Managing ideas for the development of new products. Interna- tional Journal of Technology Management, 24 (5/6), 498-513. https://doi.org/10.1504/IJTM.2002.003067 O'Brien, S.F. & Yi, Q.L. (2016). How do I interpret a confidence interval. Transfusion, 56(7), 1680-1683. https://doi.org/10.1111/trf.13635 Poveda, G., Westerski, A. &Iglesias, C. A. (2012). Appli- cation of Semantic Search in Idea Management Sys- tems. In conference for Internet Technology Secured Transactions proceedings of the international confer- ence in London, 2012, UK, London: IEEE, 230-236. Rose, A. & Jensen, V. (2012). A literature review of idea management. Retrieved June14, 2020, from https:// www.designsociety.org/publication/35610/a_litera- ture_review_of_idea_management Sandriev, A.R. & Pratchenko, O.V. (2014). Idea man- agement in the system of innovative management. Mediterranean Journal of Social Sciences, 5(12),155- 158. https://doi.org/10.5901/mjss.2014.v5n12p155 Sandstrom, C. & Bjork, J. (2010). Idea management systems for a changing innovation landscape. Inter- national Journal of Product Development, 11(3/4), 310-324. https://doi.org/10.1504/IJPD.2010.033964 Selart, M. & Johansen, S.T. (2011). Understanding the Role of Value- Focused Thinking in Idea Man- agement. Creativity and Innovation Management, 20(3), 196-206. https://doi.org/10.1111/j.1467- 8691.2011.00602.x Shani, N. & Divyapriya, P. (2011). A Role of Innovative Idea Management in HRM. International Journal of Management, 2 (1), 69-78. Summa, A. (2004). Software tools to support innovation process- focus on idea management", Unpublished Manuscript, Working paper [29], Innovation Manage- ment Institute: Helsinki University of Technology. Tan, S.H. & Tan S.B. (2010). The Correct Interpre- tation of Confidence Intervals. In proceedings of Singapore Healthcare, 19(3), 276-278. https://doi. org/10.1177/201010581001900316 Tung, W.F., Yuan, S.T. & Tsai, J.R. (2009). A custom collaboration service system for idea management of mobile phone design. Human Factors and Ergo- nomics in Manufacturing, 19(5), 494-509. https://doi. org/10.1002/hfm.20147 van den Ende, J., Frederiksen, L. & Prencipe, A. (2015). The front end of innovation: Organizing search for ideas. Journal of Product Innovation Management, 32 (4), 482-487. https://doi.org/10.1111/jpim.12213 Vandenbosch, B., Saatcioglu, A. & Fay, S. (2006). Idea management: A systematic view. Journal of Management Studies, 43(2), 259-288. https://doi. org/10.1111/j.1467-6486.2006.00590.x Voigt, K.I. & Brem, A. (2006). Integrated Idea Man- agement in Emerging Technology Ventures. In IEEE International Conference on Management of Innova- tion and Technology proceedings of the international conference in Singapore, China, IEEE: Singapore, 211- 215.https://doi.org/10.1109/ICMIT.2006.262153 Westerski, A. (2013). Semantic Technologies in Idea Management Systems: A Model for Interoperability, Linking and Filtering, Universidad Politécnica de Ma- drid: Spain. Westerski, A. & Iglesias, C.A. (2011). Exploiting struc- tured linked data in enterprise knowledge manage- ment systems: An idea management case study. In Enterprise Distributed Object Computing Conference Workshops proceedings of the 15th IEEE Internation- al in Finland, Helsinki, 2011, IEEE: Helsinki, 395-403. https://doi.org/10.1109/EDOCW.2011.14 Westerski, A., Iglesias, C.A. & Nagle, T. (2011). The road from community ideas to organisational in- novation: a life cycle survey of idea management systems. International Journal of Web Bases Com- E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 2 0 / 1 4 206 munities, 7(4), 493-506. https://doi.org/10.1504/IJW- BC.2011.042993 Wood, A. (2003). Managing Employees' Ideas From Where do Ideas Come. Journal for Quality & Partici- pation, 26(2), 22-26. Wooten, J.O. & Ulrich, K.T. (2015). Idea Generation and the Role of Feedback: Evidence from Field Experi- ments with Innovation Tournaments. Production and Operations Management, 26(1), 80-99. https://doi. org/10.1111/poms.12613 Xie, L. and Zhang, P. (2010). Idea Management System for Team Creation. Journal of Soft- ware. 5(11), 1187-1194. https://doi.org/10.4304/ jsw.5.11.1187-1194 About the authors MIĶELSONE ELĪNA Phd. BA School of Business and Finance Fields of interests Idea management, innovation management, IT management. Address K.Valdemara str.161, Riga, Latvia Phone:+371 26418258 E-mail: elina.mikelsone@ba.lv AIVARS SPILBERGS Dr.oec. BA School of Business and Finance Fields of interests Ddata analysis, risk management. Address K.Valdemara str.161, Riga, Latvia E-mail: aivars.spilbergs@ba.lv TATJANA VOLKOVA Dr.oec. BA School of Business and Finance Innovation management, business model innovations. Address K.Valdemara str.161, Riga, Latvia E-mail: volkova@ba.lv ELITA LIELĀ Dr.oec. BA School of Business and Finance Fields of interests Innovation management, business model innovations. Address K.Valdemara str.161, Riga, Latvia E-mail: elita.liela@ba.lv JĀNIS FRIŠFELDS Masters student Maastricht University School of Business and Economics. Fields of interests IT management, idea management, supply chains, logistics. Address Nassaustraat 36, 5911 BV Venlo E-mail: jfrisfelds@gmail.com This article is an Open Access article distributed under the terms and conditions of the Creative Commons Attribution 4.0 (CC BY 4.0) License (http://creativecommons.org/licenses/by/4.0/).