Research Event

Research Events are actions taken by researchers during a research study to measure, explain or make predictions about a phenomenon of interest or to manage these events. Types of events include Treatments, Measurements, and Control Events. A research protocol is an assembly and ordering of research events.

Trial

A Trial is a treatment event or combination of treatment, measurement, and control events. Specifically, a trial is an application of a treatment, stimulus, or intervention event to a subject, respondent, or other research object. It is an extension of the flipping of a coin in statistics. In performance research, a trial is often, but not necessarily, preceded or followed, by a measurement. For example, a car door may be slammed 1000 times and measured for metal fatigue after every slam or every k-th interval as in every 100 times. Since a measurement event frequently involves an instruction or other control event, a trial may also include a control event. The design called “clinical trials” involves the administration of a series of treatments accompanied by measurements or measurements taken periodically over a period of time. In some cases, a treatment requires or implies a measurement and so in this sense, the measurement is coupled to the treatment. An example would be the presentation of a passage of text with the instruction to read it and then answering a question about it or timing how long it took the subject to read it.

Sessions

An important concept involved in scheduling events is a “session.” Loosely defined, a “session” is the time spent with one subject, respondent, or other research object in one sitting or one day (generally with intermissions). However, a session can last for days, weeks, or longer as the computer is left running and the protocol remains open. Therefore, a more precise definition of “session” is the period between scheduled openings and closings of a protocol. Depending on design, a protocol may consist of one session or many sessions. When there is more than one session, sessions are demarcated by the insertion of “Session Begin” and “Session End” control events in the Events Window. The time periods between sessions in which protocols are closed are called “Off-Line Times” and may range from minutes to years. If there are no End and Begin Session control events listed in the Events Window, then the protocol will assume that there is only one event. In other words, the default condition of a protocol with regard to the number of sessions in a case is a Single Session.

Event Iterations

Event Iterations are repeated treatments, measurements, and control events or combinations of treatments, measurements, and control events within a Session. For example, in a laboratory design, a subject may be presented in one sitting a display stimulus 100 times with a measurement taken every time or every K iterations, or in a two-year clinical trial session, an event may be administered to a subject every day. In these cases, rather than entering an event in the Event Window for every time that a treatment or measurement is repeated, the number of repetitions, the order of repetitions, and the intervals between repetitions can be set in the Scheduling Dialog.

Research Case

A Case consists of one or more Sessions for one subject, respondent, or other research object. The recording of the data during a Case is a case record (see below). A completed Case is one containing all scheduled events for a specific subject, respondent, or other research object. A Case may be completed in one unbroken Session as in one day, or it may take days or weeks or years to complete, in which case, the protocol will generally have to be closed when subjects are not available.

Case Record

A case record contains data for one subject, respondent, or other research object at every point in a Case. In conventional data files, each row will represent a case record and each column will represent a variable measurement.

End Session Event (Protocol Closing)

An End Session Event is a control event that involves the closing of the research protocol at the end of a session. All of the events specified in the protocol may or may not have been completed before the End Session Event. For example, in a pretest-posttest control group design, the first session may involve only the pretest, which might be one or a large number of measurements. An end session event tells the experimenter that the protocol has been scheduled to close at this point in the protocol and asks the experimenter for confirmation that he/she would like to close it. The end session event also tells the protocol that it should save the data input during that session and that it should be ready to be “resumed” at future time or date by a Begin Session Event. This guarantees the continuity of the data input and organization and eliminates the need for merging files at the end of the project.

Begin Session

A Begin Session Event is a control event that involves the opening of a protocol and the beginning of a session. This can be unspecified as to time and date or the time and date or it can be specified. If the protocol previously has been closed by an End Session Event, then, the protocol starts where it left off at the end of the previous Session.

Off-Line Time Intervals

The time periods between sessions in which the protocol is closed are called “Off-Line Time Intervals” and may range from minutes to years. Off-Line Time Intervals are the intervals between End and Begin Session control events during which the protocol is closed and the case lies dormant. Since there may be more than one End Session Event, there will be more than one Begin Session Event and therefore, more than one Off-Line Time interval. The length of these off-line time intervals are set by the scheduling of End Session and Begin Session control events by time and date. Intervals may be short, as in hours, or they may be long, as in days, or weeks. In fact, intervals may span years and generations, as in typical cohort and longitudinal designs. Intervals may be equal or vary widely, as in one lasting a single day and the next one lasting a week or month.

Research Project

A research project involves a segment of scientific investigation generally consisting of data collection that employs one design or research method (such as experiment or survey) on a number of subjects, respondents, or other research objects. A project therefore involves multiple Research Cases. Exceptions to the single method criterion would be designs that nest one method in another. Exceptions to the multiple cases criterion would include the single organism design, which involves only one subject generally over an extended number of treatments and measurements. Of course, large projects can use more than one protocol, each involving one method or design or integrated method design. For example, a large project may collect data using an experimental design protocol, a survey design protocol, and an observational design protocol.

Event Sequencing (Ordering)

Event Sequencing refers to the order in which events are presented to subjects, respondents, or other research objects. In the general case, events are presented in the same order as they appear in the Events window of the ProtoGenie composer. If for example, you have defined a treatment, a measurement, another treatment, and another measurement listed in that order, then they will be presented in that order when the protocol is executed. The order of events in the events window can easily be changed by clicking the event that you want to move and then clicking on the UP or DOWN arrow immediately to the right of the Events window. There are two exceptions to this sequential execution pattern. They involve Conditional Branching and Trial Grouping (below)

Conditional Branching

Conditional Branching enables you to depart from the sequential execution of events in the Events window and to direct execution to a particular event either earlier or later in the list of events contingent on the value of an event that has been executed. This is done by inserted a conditional branching event at the point where you wish execution to go in the list if an earlier measurement event returns a certain value.

Execution by Event Groupings

This feature enables you to repeat a subset of events in the Events Window a specified number of times before moving on to a second Event Grouping.

Event Scheduling

Event Sequencing in the protocol (above) sets the order in which events are to be executed, but it does not specify when the events should be executed. This requires Event Scheduling. Event scheduling can be done in Real Time Scheduling or in Form Input Scheduling, or a combination of the two.

Real Time Scheduling

Real Time Scheduling involves the specification of exactly where End Session and Begin Session Events should occur in the list of events in the Events Window. One option is to set the Begin Session and End Session Events and then choose to make all groups the same by clicking on a check box. This choice will be quite common because of concern for the differential effects of history. A simple example of Real Time Scheduling is contained in the ProtoGenie New User Tutorial involving a typical Pretest-Posttest Control Group design. This design calls for two groups (treatment and control). The treatment group gets a pretest measurement, a treatment, and a posttest measurement. The control group gets the pretest measurement, a Null Event (no treatment), and the posttest measurement. Scheduling involves the placement of Begin and End Session control events to delineate sessions, which also sets the Off-Line Time Intervals between sessions. For details on scheduling events in a concrete example, click here for the New User Tutorial.

Form Input Scheduling

Form Input Scheduling involves entering session data on a paper form over a number of sessions and then entering them into the computer at designated intervals and times. This process may be required if subjects self-administer their treatments, as in clinical trials involving a new medication. In this kind of situation, subjects may be asked to check off days on a calendar-like form when they take the medication and to indicate how they feel on a five-point scale. Then, they submit these forms to a researcher or technician every month or other prescribed time, at which time, the researcher or assistant transcribes the data on the paper form to the protocol. Since the advent of handheld computers, there are now situations in which data can be entered in the field by one or more people and downloaded directly to the ProtoGenie protocol.