Designing Your Experiment
These articles describe the general methods of designing an experiment. The conceptual work that needs to be completed prior to implementing an experiment is reviewed in detail. The process that is described can be followed for any experiment. These articles also identify design issues that should be considered prior to attempts to compete an experiment for data collection. For users who are new to experimental design, using computers to implement experiments, or to E-Studio itself, we recommend working through these articles in their entirety.
For the purposes of the next several articles, the lexical decision experiment will be used as an example. Read the following overview of the experiment to become familiar with the basic experiment:
The lexical decision experiment requires the participant to make a decision about a string of letters. In this example, they need to decide if a string of letters is a word or nonsense. First the participant sees a prime presentation (e.g., “word” or “nonword”). This is followed by a the presentation of string of letters which is either an actual word or nonword. The participant responds and is given feedback. Several trials run and then the experiment ends with displaying a Goodbye screen to the participant.
Define Your Experiment in Stages
An enormous amount of detail is required to program an experiment, and in order to create a scientifically valid experiment these details must be clearly conceptualized before implementation in E-Studio. Each of the variables, and the experimental procedures must be precisely defined. Additionally, careful consideration must be given to the data collection, and how the data is analyzed. Once these steps are complete, the experiment should be tested and the data analyzed. It is bad practice to collect data before testing the experiment or visualizing the analysis.
Conceptualizing and testing both the experiment and analysis, will save time and result in better research. It is best to design the experiment in a series of stages, each containing a few steps. Each step describes one aspect of the experiment, and builds on the previous step like a ladder. Taking the time to map out a few trials on paper before programming begins will substantially speed development and reduce errors. The table below contains the recommended stages for developing an experiment: