EXPERIMENT DESIGN: Stage 1: Conceptualize the Core Experimental Procedure [37328]

The goal of this stage is to get the basic procedure designed. There are 3 major steps to Stage 1, as shown below.

 

Stage 1: Conceptualize the Core Experimental Procedure

1) Preparation Work • Provide an operational specification of the base procedure • Create a folder for the experiment

2) Conceptualize a Basic Trial • Specify the experimental design, independent variables, stimuli, and expected responses • Specify the trial procedure • Specify the non-default and varying properties of the trial events

3) Data Logging • Specify what data will be logged for analysis • Verify the core experiment • Visualize the data logging of the core experiment

 

Stage 1, Step 1: Preparation Work

Provide an operational specification of the base procedure

The process of putting your experiment into words forces you to clearly define all aspects of the experiment. In order to design a scientifically sound experiment, precisely define what is being manipulated (independent variables), the expected outcome (dependent variables) and the hypothesis before any programming takes place. Then write a draft of the abstract for the experiment being implemented, particularly detailing the procedure.

 

Hypothesis

Participants will recognize a string of letters that make up a word faster than they recognize a non-word.

 

Abstract

The experiment will measure the time a participant takes to make a lexical decision. The independent variable is the string type: word or non-word. The participant will be presented with a fixation (+) displayed in the center of the screen for one second. Then a Probe Display will present a letter string stimulus in the center of the screen for up to 2 seconds. The stimulus display will terminate when the participant responds. Participants are to respond as quickly as possible as to whether the stimulus was a word or a non-word by pressing the “1” or “2” key respectively. The dependent measures are the response (i.e., key pressed), response time, and response accuracy of the Probe Display. The stimuli will be words and non-words, presented in random order in black text on a white background.

 

Using this rough abstract, you can define the basic procedure for your experiment.

 

Create a folder for the experiment

It is good practice to organize your design notes and your experiment specification files. Experiment development is highly iterative in nature, and you can quickly develop many different versions of a similar experiment. We recommend developing your experiments in the My Experiments folder that was created in your Documents folder when E-Prime was installed. Within that folder organize your notes in folders that are named by the experiment series, e.g., LexicalDecision. We also recommend that as you make changes to your experiment you increase the number each time you save major changes to that version of the experiment, Stage1-LexicalDecision001.es3, Stage2-LexicalDecision001.es3, etc. and that you number each version of the experiment (e.g., LexicalDecision001). Do not minimize the importance of organizing your design work and experiments. A little thought at the start of your experiment development work will make retrieving specific versions of your design notes, and ultimately experiment programs, much easier in the future.

 

Stage 1, Step 2: Conceptualize a Basic Trial

Specify the experimental design, independent variables, stimuli, and expected responses

In order to specify the design of the experiment you will need to list all of the conditions, stimuli, and expected responses. The abstract of the lexical decision experiment includes the following details about the design: The independent variable is whether a letter string is a word or a non-word… the stimuli will be text strings of words and non-words… Participants are to respond as quickly as possible as to whether the stimulus was a word or a non-word by pressing the “1” or “2” keys respectively. The underlining in the abstract highlights key terms that will directly influence the names or settings in the E-Prime experiment specification. This portion of the experiment design can be organized in a table, such as the following:

 

Condition	Stimulus	Correct Response
Word	cat	1
NonWord	jop	2

 

In this case, we start with one independent variable (Condition) having two levels or cells (Word, NonWord). For each cell, the table contains a separate line. We need to specify the stimulus that determines what the participant sees in that condition and the correct response to the stimulus, which determines how the response is scored. Later, more independent variables (e.g., word frequency and priming), stimuli, and associated responses can be added.

 

Specify the trial procedure

The core trial procedure of an experiment is a minimal, repetitive portion in which different conditions are selected, stimuli are presented, and responses are collected. The core trial procedure typically defines the sequence of events the participants experience. It is useful to make a diagram of the sequence of events and connect the design to the events.

 

For example, in the lexical decision experiment, the core trial procedure was described in the abstract: The participant will be presented with a fixation (+) displayed in the center of the screen for one second. Then a Probe Display will present a letter string stimulus in the center of the screen for up to 2 seconds. The stimulus display will terminate when the participant responds. The lexical decision procedure can be specified in a list of events as follows:

• Select the Stimulus from the list of trial stimuli.
• Present the Trial Procedure, which contains: Fixation, then ProbeDisplay, and collect the response.

 

Specify the non-default and varying properties of the trial events

The specific nature of each stimulus can be specialized by setting critical properties of the stimuli. E-Prime provides defaults for dozens of properties for each stimulus (e.g., font style, forecolor, background color, location, duration, etc.). The properties that do not have a default value must be set (e.g., text in the fixation display to be a “+”), as must any properties that differ from the default. Also, the properties that vary in the experiment (e.g., based on the condition) must be set. The abstract specified a series of properties that must be set:

 

The stimuli will be words and non-words, presented in random order in black text on a white background.

 

The participant will be presented with a fixation (+) displayed in the center of the screen for one second. Then a Probe Display will present a letter string stimulus in the center of the screen for up to 2 seconds. The stimulus display will terminate when the participant responds. Participants are to respond as quickly as possible as to whether the stimulus was a word or a non-word by pressing the “1” or “2” key respectively.

 

For the lexical decision experiment, the fixed and varying properties can be summarized in the following table:

 

Object	Fixed Properties	Varying Properties
Fixation	• Present “+” in the center of the screen • Duration = 1 second (default) • Foreground color = black (default) • Background color = white (default)	(none)
Probe	• Display in center of the screen (default) • Duration = 2 seconds • Input keys = “1” and “2” • Terminate display upon response • Foreground color = black (default) • Background = white (default)	• Stimulus (e.g., “cat”, “jop”, etc.) • Correct response = “1” or “2”

 

Stage 1, Step 3: Data Logging

Specify what data will be logged for analysis

For a research experiment, the experimenter has to record and analyze data. This requires clarity on what variables will be logged. Be able to answer, “What dependent measures will be recorded in reference to specific objects in the experiment?” The abstract above stated: The dependent measures are the response, response time, and response accuracy of the Probe Display. After a trial, expect to record the following information for the Probe Display: What was the response that the participant entered (i.e., 1 for word and 2 for nonword)? What was the response time? What was the accuracy (i.e., 1 for correct and 0 for wrong)?

 

Verify the core experiment

At this point, the minimal core of the experiment is designed. Once you have this much of the experiment designed, it is a good idea to verify your design. Of course, when you are implementing the experiment in E-Prime, you would run the experiment at this point to see if it works as intended. Once you have the experiment programmed, you would simply run the experiment to see if it works as intended. Now we will walk through the proposed layout to make certain it is logical.

 

Recall the experimental procedure abstract

The participant will be presented with a fixation (+) displayed in the center of the screen for one second. Then a Probe Display will present a letter string stimulus in the center of the screen for up to 2 seconds.

 

It is helpful to draw out the displays the participant would see. In this case, expect to see two trials, with the first presenting a fixation “+” for one second and then the word “cat” for 2 seconds, or until a response is entered. Then the “+” is presented again, and the second stimulus “jop” is displayed until a response. For example:

 

 

Prior to running the experiment you should also be able to envision the ways in which the trial event ends. In this case, when the participant presses a valid response key, the text string will disappear, and the experiment will go to the next trial.

 

Visualize the data logging of the core experiment

An experiment is only useful if the critical data are recorded for later analysis. Be clear about what you expect to see as input data for the analysis. From the abstract:

 

The independent variable is whether a letter string is a word or non-word… The dependent measures are the response (i.e., key pressed), response time, and response accuracy to the Probe Display.

 

In this case, for each trial you would expect to see the condition and unique stimulus selected, plus the dependent measures (participant response, response time, and accuracy of the response) in the data file. Prior to actually collecting any data, you should be able to visualize the information that is being logged and how you will analyze the data to investigate the experimental questions of interest. For example, the trials described above might appear like the following:

 

Independent Variables			Dependent Variables
Condition	Stimulus	Correct Response	Probe Response	Probe RT	Probe Accuracy
Word	cat	1	1	586	1
NonWord	jop	2	1	1224	0

 

For trial 1 of this run of the experiment, the independent variables were Condition = Word, Stimulus = cat, and Correct Response = 1. The dependent variables were the response key (Probe Response = 1), response time (Probe RT = 586), and response accuracy (Probe Accuracy = 1, indicating a correct answer). For trial 2, the Condition = NonWord, Stimulus = jop, Correct Response=2, Probe Response =1 (which mismatches the expected response), Probe Reaction Time =1224, and Probe Accuracy=0 (indicating an error).

 

Next Article: EXPERIMENT DESIGN: Stage 1: Conceptualize the Core Experimental Procedure [37328]