Our main aim in this book is to look closely at Theories of Change. But as Theories of Change are a special type of Theory, we need to look at Theories in general first. In this first section of the book, we start to show how potentially any social science Theory can be expressed purely as an assemblage of simple Theories, basic units in which one or more Variables influence a single consequence Variable.
In the first sections of this book we will try to keep things as simple as possible, leaving some of the more “wicked” problems (those that involve wicked Rules - those which are chaotic, emerging, etc.) to Section IV.
To get a feel for the Chapters in this section, you can just to click through each Chapter (by pressing the right-pointing chevron at the right of this screen) and read the brief summary at the top of each one. You can also see the titles of the Chapters in this section listed in the table of contents to the left of the screen.
What makes a good Theory: A check-list
- A Theory says how certain Variables (things that can have different Levels) influence one another.
- A Theory of Change (see the next section) is what someone has when they have a Theory about how doing one thing can influence something else they value. So in order to build a good Theory of Change, first we need to know how to build a good Theory.
With ordinary Theories, such as those we see in Theories of Change, readers are usually left to guess what a box or an arrow actually means - some typical ambiguities are presented in xx. Do the boxes represent variables or something else, like people? (We don’t recommend the latter.) Are the arrows causal connections, or critical paths xx? Do they claim that the variables at the tail of the arrow make a substantial contribution to the variable at the head of it? Or just any kind of contribution at all? Do they represent necessary or sufficient conditions or something else? If you construct your Theory in Theorymaker, you can make it clear to colleagues who also understand Theorymaker exactly what you mean. So Theorymaker gives us a standardised toolkit for expressing and understanding Theories and related ideas like “Impact” and “Output”.
The check-list below will provide us with a very brief overview of the main features of Theorymaker and the problems they address. Using Theorymaker will help us to tick these boxes.
So, a good Theory of Change has to be a good Theory. This means being sufficiently clear in particular on the following questions. (By “sufficiently clear”, I certainly don’t mean that all Theories always need to cover all these demands. A good Theory is as simple as possible, but not simpler. Where there is room for important and relevant ambiguity, it should be cleared up as described here.)
Checklist for a good Theory
- does the Theory identify the relevant basic elements, the Variables; things that can be different, e.g., no/yes Variables like “the law is passed” or count Variables like “number of teachers completing the training course”), or percentages or something else … .
- Are the names of the Variables distinguished from information about the kind of Difference we want to make with our Intervention?
- Often, Variables are best defined as belonging to specific people or groups of people or institutions. Is this clarified?
- how are the Variables connected into simple Theories; which Variables influence which others?
- how do these “simple Theories” fit together into composite Theories?
- what are the Rules describing how the Variables influence one another - e.g. positively, negatively, weakly, strongly?
- do some Variables share some attributes, e.g. the same time-point, the same stakeholder, etc, and if so, should they be visually grouped together?
- are all of the Variables supposed to be each assigned to a particular point in time, or are any of them supposed to stretch or repeat across time-points (what Theorymaker native speakers call “for time” Variables), and/or are they supposed to be common to groups of people?
- which of the key Variables are hard to directly assess/measure; for these, can other Variables be identified as proxy “Indicators”?
- are there any Variables which are defined in terms of existing Variables rather than influenced by them?
In the rest of Section 1 of this book, I go into the points on this check-list in a little more detail and also introduce Theorymaker, which helps you tick these boxes. Theorymaker is both written and visual because any piece of written Theorymaker corresponds to a particular Theory diagram, and vice versa. The diagrams below are Theorymaker diagrams, and above each diagram is the corresponding written Theorymaker (you can see it if you hover your mouse above the diagram).
If you use Theorymaker to construct your Theory, you will find it easy to tick the boxes on the check-list.