Methodological Challenges regarding RBT


Hoskisson et al. (1999) assert that the research methods applied within empirical tests of RBT in the past, overall, do not seem to be suitable for the task at hand. The authors argue that it is due to the emphasis on the idiosyncratic nature of a firm’s resources and capabilities that empirical testing of RBT faces great challenges.556 As previously outlined in chapter 2.3, the power of RBT in explaining sustainable performance is based upon strategic resources, i.e., on valuable, rare, inimitable, and non-substitutable resources, which are, in part, by their nature unobservable (e.g., tacit knowledge, organizational culture).557 As a result, empirical testing of these unobservable resources and their effects on firm performance seems to be difficult. “Regarding these challenges, the need for a multiplicity of methods to identify, measure, and understand firm resources is increasing. Empirically we have some understanding of the what in many cases but now need to extend our methodology so we can know how as well.”558


Methodological Challenge Data Analysis Method Knowledge Resource Strategic Group Strategic Resource 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 556.
    Cf. Hoskisson et al. (1999), p. 420; Hitt et al. (1998), p. 13.Google Scholar
  2. 557.
    Cf. Godfrey/ Hill (1995), p. 523; Rouse/Daellenbach (1999), p. 488.Google Scholar
  3. 558.
    Rouse/ Daellenbach (2002), p. 965.Google Scholar
  4. 560.
    Cf. Brush/ Artz (1999), p. 229.Google Scholar
  5. 561.
    Cf. Combs/ Ketchen (1999), p. 876, footnote 4.Google Scholar
  6. 562.
    King/ Zeithaml (2001), p. 79.Google Scholar
  7. 563.
    Godfrey/ Hill (1995), p. 529f.Google Scholar
  8. 564.
    Cf. Godfrey/ Hill (1995), p. 523; Rouse/Daellenbach (1999), p. 488.Google Scholar
  9. 565.
    Cf. Fernández et al. (2000), p. 84.Google Scholar
  10. 566.
    Cf. Barney (1991), p. 108f; King/Zeithaml (2001), p. 77.Google Scholar
  11. 567.
    “Tacitness refers to the implicit and non-codifiable accumulation of skills that result from learning by doing.” Reed/ DeFillippi (1990), p. 89.Google Scholar
  12. 568.
    Cf. Godfrey/ Hill (1995), p. 531; Rouse/Daellenbach (1999).Google Scholar
  13. 569.
    Cf. Balogun et al. (2003); Ambrosini/Bowman (2001).Google Scholar
  14. 570.
    Cf. Amabile et al. (2001); Rynes et al. (2001); Rouse/Daellenbach (1999, 2002).Google Scholar
  15. 571.
    Cf. Eisenhardt (1989), p. 546.Google Scholar
  16. 572.
    Cf. Sharma/ Vredenburg (1998).Google Scholar
  17. 573.
    Cf. Zander/ Kogut (1995).Google Scholar
  18. 574.
    Cf. Henderson/ Cockburn (1994); Peng/York (2001).Google Scholar
  19. 575.
    Cf. Rouse/ Daellenbach (1999), p. 489.Google Scholar
  20. 576.
    Cf. Rouse/ Daellenbach (1999), p. 488 and (2002), p. 963.Google Scholar
  21. 577.
    The classification of traditional vs. specialized techniques follows Shook et al.’s (2003) classification. Here, only those methods were coded that were directly relevant to testing the study’s hypotheses or answering its research questions. If more than one data analysis method was used in a given study, coding was performed for all principal methods used.Google Scholar
  22. 578.
    Cf. Carmeli (2004), p. 116.Google Scholar
  23. 579.
    Cf. De Oliveira-Wilk/ Fensterseifer (2003), p. 1001ff. Also, the authors assert that the use of the software COPE, a mapping software, was very helpful. For further information, see De Oliveira-Wilk/Fensterseifer (2003), p. 1003.Google Scholar
  24. 580.
    Based on De Oliveira-Wilk/ Fensterseifer (2003), p. 1007.Google Scholar
  25. 581.
    Cf. King/ Zeithaml (2003), p. 769.Google Scholar
  26. 582.
    Cf. Rouse/ Daellenbach (1999), p. 489f. See also Rynes et al. (1999), p. 871f for merits and disadvantages of academic research inside organizations. In their empirical assessment of academic research inside organizations, the authors concluded that spending time at the research site enhanced scientific impact and also the prospects for research implementation. Cf. Rynes et al. (1999), p. 895.Google Scholar
  27. 583.
    Rouse/ Daellenbach (1999), p. 490. See also Ambrosini/Bowman (2001), p. 823f; Balogun et al. (2003), p. 201. Following Barney and Mackey (2005), many studies “examine the value potential of a firm’s resources at a level of analysis below that of the firm. Not surprisingly, the most correct level of analysis at which to examine the relationship between a firm’s resources and its strategies is at the level of the resource, not the level of the firm. However, the firm is usually the unit of accrual. We are likely to learn a great deal more about the relationship between resources and strategies if scholars are able to “get inside” the firm, where resources reside, rather than simply correlate aggregate measures of resources with aggregate measures of the value of a firm’s strategies (Rouse & Daellenbach, 1999).” Barney/Mackey (2005), p. 5.Google Scholar
  28. 584.
    Amabile et al. (2001), p. 418.Google Scholar
  29. 585.
    Cf. Amabile et al. (2001), p. 418. See also Rynes/McNatt (2001).Google Scholar
  30. 586.
    Cf. Amabile et al. (2001), p. 419.Google Scholar
  31. 587.
    E.g., McWilliams et al. (2002); Mauri/Michaels (1998); Borch et al. (1999); Boxall/Steeneveld (1999).Google Scholar
  32. 590.
    Cf. Liebold/ Trinczek (2002), p. 37f and p. 58.Google Scholar
  33. 591.
    Cf. Schütze (1983); Holtgrewe (2002), p. 72f; Hermanns (1991), p. 184f.Google Scholar
  34. 592.
    Cf. Ambrosini/ Bowman (2001), p. 820; Martin (1982), p. 257. Usually, participants are asked to tell two stories regarding what has in the past caused organizational success vs. failure. Conducting the interview within the participants’ organization and its familiar surroundings can serve as cues. Cf. Ambrosini/Bowman (2001), p. 820f; Balogun et al. (2003), p. 205.Google Scholar
  35. 593.
    Cf. Cook (2004), p. 27; Flanagan (1954), p. 327ff.Google Scholar
  36. 594.
    Cf. Balogun et al. (2003), p. 208; Breakwell/Wood (1995), p. 293f.Google Scholar
  37. 595.
    Cf. Easterby-Smith et al. (2002), p. 114f.Google Scholar
  38. 596.
    Cf. Breakwell/ Wood (1995), p. 294.Google Scholar
  39. 597.
    Cf. Kelly (1955); Easterby-Smith et al. (2002), p. 97; Reger (1990), p. 302; Fransella/Bannister (1997).Google Scholar
  40. 598.
    Cf. Bougon et al. (1989), p. 328; Ambrosini/Bowman (2001), p. 819. Through self-interviewing techniques “...the events, objects, and concepts [the participants] use to express their questions... reveal their tacit and explicit knowledge” Bougon et al. (1989), p. 329.Google Scholar
  41. 599.
    See chapter 5.1.2 as well as Ambrosini/ Bowman (2001), p. 817; Huff (1990), p. 15f; Gnyawali/Tyler (2005), p. 225ff. “Research instruments such as surveys and structured interviews are likely to be inappropriate insofar as individuals cannot be asked to state what they cannot readily articulate. The main challenge that may have to be faced is finding ways of expressing what is, or more correctly what has not been up to now, expressible.” Abrosini/Bowman (2001), p. 815.Google Scholar
  42. 600.
    “The causal mapping method [...] is an indirect way of surfacing tacit skills. It will be fragmented, not comprehensive, partial and biased but it should provide some insights to both participants and researchers into tacit skills and organizational success.” Ambrosini/ Bowman (2001), p. 825.Google Scholar
  43. 601.
    Cf. Millward (1995), p. 276; Liebig/Nentwig-Gesemann (2002), p. 145. See Steyaert/Bouwen (1994), p. 142ff for a comparison, benefits, and limitations of different types of group discussion methods.Google Scholar
  44. 602.
    Cf. Erbslöh, p. 36; Mayntz et al. (1971), p. 134; Greve/Goldeng (2004), p. 137.Google Scholar
  45. 603.
    Cf. Greve/ Goldeng (2004), p. 136. For instance, Bergmann-Lichtenstein and Brush (2001) use a panel design in their study on how resource bundles develop and change over time in new ventures.Google Scholar
  46. 604.
    There are two variations of participant observation. The first, called co-research, draws on the complementary perspectives, interests, skills, and knowledge of an academic (outsider view), a host manager from within the firm (insider), and a co-researcher from a different organization (insider in type of organization, outsider in the sense that his/her own company is composed differently). Cf. Hartley/ Benington (2000). The second, termed action research, describes a research approach of organizational intervention that attempts to result in practical transformation and advanced knowledge. Cf. Huxham/Vangen (2003), p. 384.Google Scholar
  47. 605.
    Cf. Waddington (1994), p. 108. Also, participant observation methods can easily be combined using interviewing techniques for getting more information. Cf. Bachman (2002), p. 335.Google Scholar
  48. 606.
    Data gathered through observations generally include detailed descriptions of people, events, and conversations as well as the observer’s actions, feelings and hypotheses. Cf. Waddington (1994), p. 109f.Google Scholar
  49. 607.
    Cf. Millward (1995), p. 288. See Mayring (1991), p. 210ff for respective procedures. Regarding RBT, Deephouse (2000), for instance, used content analysis of newspaper articles to measure the media reputation of commercial banks. His results indicate that media reputation, an intangible resource, is a resource influencing performance.Google Scholar
  50. 608.
    Cf. Easterby-Smith et al. (2002), p. 122; Wiedemann (1991), p. 442ff.Google Scholar
  51. 609.
    Cf. Tichy et al. (1980), p. 372. According to von Kardorff (1991), a qualitative approach to network analysis is advisable, i.e., case studies of rather small groups using methods, such as participant observation, interviews, diaries, or group discussion. Cf. von Kardorff (1991), p. 404. See Zwijze-Koning/De Jong (2005) for further details on underlying data collection techniques.Google Scholar
  52. 610.
    Cf. Christensen et al. (2002), p. 10.Google Scholar
  53. 611.
    An interesting approach of categorizing and attributing resources can be found within Jolly’s (2000) contribution. The author develops a continuum of attribute-pairs such as ‘tangible — intangible’, ‘marketable — unmarketable’, ‘discrete — systemic’, etc. and aligns resources on a scale between these continuums. Cf. Jolly (2000), p. 786. Here, attributes not yet refer to the four resource conditions.Google Scholar
  54. 612.
    Cf. Deephouse (2000), p. 1092. See chapter 3.2.1 for prime examples in this connection.Google Scholar
  55. 613.
    Cf. Peteraf/ Barney (2003), p. 320. For instance, Peteraf and Barney assert that competitive advantage is per definition a relative term and therefore requires an exogenous basis for comparison; rents as well.Google Scholar
  56. 614.
    Rouse/ Daellenbach (1999), p. 489.Google Scholar
  57. 615.
    Cf. Carmeli (2004), p. 113. For further examples, see Fahy (2002), O’Regan/Ghobadian (2004), and Santhanam/Hartono (2003).Google Scholar
  58. 616.
    Rouse/ Daellenbach (1999), p. 487.Google Scholar
  59. 617.
    Harrigan (1983), p. 400.Google Scholar
  60. 618.
    See Eisenhardt (1989) on further guidelines on how to conduct case studies.Google Scholar
  61. 619.
    Cf. Harrigan (1983), p. 403.Google Scholar
  62. 620.
    Cf. Levitas/ Chi (2002), p. 961.Google Scholar
  63. 621.
    See Miller and Friesen (1982) for advantages and limitations of longitudinal research in general and on five different research types in particular, classified according to three dimensions: breadth of focus, sample size, and the extent to which quantification occurs. Cf. Miller/ Friesen (1982), p. 1013ff.Google Scholar
  64. 622.
    “This, of course, implies that the best resource-based empirical work will involve collecting primary data from within firms in a carefully drawn sample.” Barney/ Mackey (2005), p. 5.Google Scholar
  65. 623.
    Cf. Chen et al. (1993), p. 1614ff; Harrigan (1983), p. 398; March/Sutton (1997), p. 701; Ambrosini/Bowman (2001), p. 824f; Balogun et al. (2003), p. 217. Studies from the present review using outsider information are for example Douglas/Ryman (2003), Christiaanse/Venkatraman (2002), Combs/Ketchen (1999), and King/Zeithaml (2001).Google Scholar
  66. 624.
    Cf. Chen et al. (1993), p. 1615.Google Scholar
  67. 625.
    Cf. Harrigan (1983), p. 401.Google Scholar
  68. 626.
    Cf. Chen et al. (1993), p. 1615.Google Scholar
  69. 627.
    Cf. Chen et al. (1993), p. 1618. For further details see Chen et al. (1993), Shrout/Fleiss (1979), Snow/Hambrick (1980), and Venkatraman/Grant (1986).Google Scholar
  70. 628.
    Cf. Chen et al. (1993), p. 1623ff.Google Scholar
  71. 629.
    Cf. Harrigan (1983), p. 400.Google Scholar
  72. 630.
    Cf. Shook et al. (2003), p. 1231. Shook et al. (2003) used data from 77 strategic management researchers who attended the Academy of Management’s Business Policy and Strategy Division Doctoral Consortium between 1996 and 2001, which is for Ph.D. students having defended their dissertation proposals and also being nominated by their institution as its best eligible student. Cf. Shook et al. (2003), p. 1233.Google Scholar
  73. 632.
    Cf. Barnett et al. (1994), p. 15ff. See also Jordá (1999) for details on random-time aggregation in partial adjustment models.Google Scholar
  74. 633.
    Cf. Barney (2001), p. 51f.Google Scholar
  75. 634.
    Cf. Majumdar (1998), p. 822.Google Scholar
  76. 636.
    Cf. Weller (2004), p. 520. Accordingly, event history studies are also suitable methods for assessing causal structures using longitudinal data.Google Scholar
  77. 637.
    Cf. McWilliams/ Siegel (1997), p. 628; McWilliams/McWilliams (2000), p. 1.Google Scholar
  78. 638.
    Cf. Segev et al. (1999).Google Scholar
  79. 639.
    Examples of SEMs are LISREL and PLS (partial least squares). Compared to LISREL, PLS has some additional advantages: (a) it makes no assumptions about multivariate normality on the data; and (b) it is also suitable for relatively small samples. Therefore, PLS is preferred to LISREL, especially at the beginning of theory building, and when the primary concern is the prediction of the dependent variable. Cf. Birkinshaw et al. (1998), p. 230; Delios/Beamish (1999), p. 717; Cortina et al. (2001), p. 325.Google Scholar
  80. 640.
    Hoopes et al. (2003), p. 889.Google Scholar

Copyright information

© Gabler | GWV Fachverlage GmbH, Wiesbaden 2008

Personalised recommendations