As you have already learned, research is not the full picture of the reality; it is meant to condense the world into useful and insightful information. Importantly, it builds on the prior knowledge and accumulates to a body of literature. All these aspects mean that research needs theory.
For instance, let’s say we wish to understand what makes some entrepreneurs internationalise their business and even expand to global markets. Could be anything, right? Is there not enough demand in domestic environment, or a huge need abroad? The entrepreneurs’ personal background might influence, and the support programs by the state might give a push. As we cannot (and don’t even aspire to) explain the whole story, we turn to the most interesting insight offered by the prior theory: What is well-established knowledge that we could build on? What is something that could reasonably offer an explanation, but has not been fully explored? This offers us the theoretical model that proposes different concepts are connected to each other, and guides all our empirical research.
When moving on to the empirics, we must convert the theoretical propositions and concepts into hypotheses that describe relationships in the empirical world, and can thus be tested. For instance, we could argue that the theory suggests that start-ups heavily rely on the personal capabilities of their owners, and there is a connection between the owners’ characteristics and the internationalisation strategy of the company. To convert the theoretical concepts of “owners’ characteristics” and “internationalisation strategy” into a testable hypothesis, we could formulate that “the more international experience the owner has, the faster the start-up establishes international operations”.
But how do we actually test whether that is true? As discussed during the previous session, we must select variables that represent these empirical concepts, and collect and measure data. This means, that we could measure international experience by collecting survey data on, for instance, how many years the person has worked abroad, and/or how many times a year they travel abroad in business. Respectively, our variables for speed of internationalisation could be drawn from the financial statements of the companies by looking at when was the first time the company has e.g. revenue/personnel/affiliate in a country outside their original domicile. And, if we have too many similar variables, we could think of putting them into a common score by using factor analysis.
Our theory and hypotheses also formulate a conceptual model that describes the relationship between the variables: what is the dependent variable that we are trying to explain and which independent variables that could help to predict the outcome? In the example we have used, this is pretty straight-forward: we suggest company internationalisation depends on owner characteristics. The first would thus be the dependent variable and the second the independent variable; and they have a direct relationship.
However, whilst this might turn out to be a significant relationship, could there not be other things that effect the internationalisation of a company? What can you think of? Maybe, for instance, ICT sector would tend to internationalise faster than, say, construction industry. In order to see whether owner background alone has statistical significance, we must rule out these “rival explanations”. To do this we define control variables, collect data on them too and include them in our model.
Also, in many cases the accumulation of knowledge has led to the situation that we already know a lot about the phenomenon and wish to gain more nuanced insight. Maybe the relationship is not as straight-forward as we suggested above. What if there are other issues that influence the relationship, such as the composition of the management team as a whole. In this case, we could include that as a moderating variable in our model; it is something that makes the relationship between owner and internationalisation different in some ways, e.g. stronger/weaker, or negative/positive. Or could it be so that there is another element through which the independent variable (owner background) produces the outcome (internationalisation)? In this case, we need mediating variables in the model.
Does that sound complicated? Remember, it is simplification after all, and actually kind of fun!
Kibler, E., Salmivaara, V, Stenholm, P., & Terjesen, S. 2018. The Evaluative Legitimacy of Social Entrepreneurship in Capitalist Welfare Systems. Journal of World Business, 53(6), 944-957.
Sparrowe, R. T., & Mayer, K. J. (2011). Publishing in AMJ--Part 4: Grounding Hypotheses. Academy of Management Journal, 54(6), 1098-1102.
Theory and Hypotheses
Research Questions, Hypothesis and Variables
Variables: IV, DV, control, Mediation, Moderation
Exploratory Factor Analysis
Exercise 6.1 – Comprehend
After getting acquainted with the materials (readings, videos), explain briefly (max 1 page) why it is important to pay due attention to the empirical research design when either conducting research, or reading a research report or article.
Exercise 6.2 – Critique
Read the Theory and hypotheses, and Research design and method sections of the article by Kibler et al., (2018), and answer the questions on it.
What kind of theory/-ies do the authors use to build their model and ground their hypotheses? How do they operationalise their model into variables (DV, IV, control)? Would you have ideas for improving the process of building a bridge between the theory and the empirics (e.g. do you think they argue their hypotheses convincingly, do the variables fit well with the concepts, and could they have included other control variables)?
Please check. Did you gain an understanding of the following?
- The idea of moving from the conceptual level to the empirical world
- Formulation of testable hypotheses
- The role of different types of variables (DV, IV, control)
If you can answer everything with a confident Yes! then you have achieved the learning objective of this session.