The theory framework is the set of concepts that underlie the research question. One of the jobs of the literature review is to select the theory framework you intend to use and describe it. Note that science uses the term “theory” differently from common use, e.g. if I say “My theory is people don’t listen” that is my opinion and others are entitled to have a different opinion. In contrast in science, not all theories are equal as it is about selecting the best from the rest based on data, e.g. to say “Evolution is just a theory” is to confuse what is based on an enormous amount of evidence with an opinion. The difference is that while opinions often lead nowhere, a good theories lead to better questions, e.g. one asks “Why did God give giraffes long necks?” or “Why did giraffes evolve long necks?” depending on the theory base. One question leads to speculation and the other to a search for evidence. Theory is the beginning and end of research, as theories lead to more research questions. The literature review makes the theory behind the research question explicit. So review the literature to pick the theory that underlies the research question.
Theory directs data. Science is not the search for data but for knowledge based on data, as the world has data without end but knowledge is rare. Knowledge comes from the questions asked of data not the data itself. For example, in 1887 Michelson and Morley went to great lengths to find out if light went at the same speed in every direction not because they cared but to test the theory that everything moved in an invisible ether. Searching for data without theory is like a blind man casting about to invariably find nothing, e.g. student questionnaires that ask every question they can think of. When asked what they are looking for, they say everything but the result is a lot of useless questions that lead nowhere. The key to finding is to know what you are looking for, and theory tells you where to look.
Theory is abstract. Scientific theories connect abstract concepts to evidence from the vastness of data. The power of the abstract is that one theory can describe countless specifics, e. g. the law of gravity applies to all matter, from a falling apple to the moon. The power of science is based on the abstractness of theory but that is also its great weakness, as it can disconnect from reality. Every paper needs the power of theory, as research without theory is like a car without an engine – it doesn’t go far. A theory is abstract concepts connected.
Topic construct. Every scientific theory is built from ideas but a construct is a concept connected to data evidence, that is “constructed” as it were from data. A construct measured as a number is a variable, e. g. the construct Usage as measured by web hits is a variable. In contrast, the construct consciousness is based on personal experience so science can address it but currently has no measure. The topic construct is the construct the research investigates. If there is more than one, combine them into a higher construct, e. g. combine software Satisfaction and Enjoyment into a more abstract Response construct. A construct and its measure are the same thing at different levels, so if your study measures fear by changes in blood pressure, don’t say that fear “causes” changes in blood pressure. The literature review identifies construct(s) relevant to the topic, i.e. concepts and their evidence base.
The literature review must state the topic construct, e. g. in a thesis entitled “Web Site Effects” the topic construct was unclear, as the question “Effects on what?” was ignored. Possible topic constructs included the customer relationship, user site acceptance, web site usage, marketing, communication, etc., any one of which was a study in itself. The research meandered and gave weak results because the literature review didn’t define a topic construct. Check that the literature review defines a topic construct.
Causality. A literature review not only identifies the key constructs but also how they might cause each other. In describing theoretical constructs, it is critical to distinguish what is causing from what is being caused. Quantitative science divides variables into:
- Dependent variables. Variables that are caused, i.e. change depending on other variables. So if the topic construct is a variable, it is always the dependent variable, e. g. for research into how ease of use affects web site use, the latter topic construct measured as number of views is the dependent variable.
- Independent variables. Variables that cause others to change. Research often look at independent variables that change a dependent variable e.g. web usage is also affected by the value a site offers.
- Moderating variables. Variables that alter causal effects, often individual differences, e. g. expertise may moderate how ease of use affects usage, as experts care less about ease of use and more about value.
Note: A construct can be both a cause and an effect, e. g. if ease of use increases web usage that in turn increases web sales, web usage is both dependent and independent. The literature review must distinguish what is causing from what is being caused.
Theory figure. After a literature review picks the theory used in the research, it is often a good idea to provide a figure. This may be an existing theory, a combination of existing theories, a modification of an existing theory, or a new theory of your own making. Since the theory is what directs your research, it is worth showing it in a figure. A good literature review gives a figure of the theory the research uses.