Exploration - The Procedure

Once you have chosen which study you will carry out, you are ready to begin designing your procedure.  Think of your experimental procedure like a carefully scripted scene in a movie.  From the moment the experiment begins, every detail - from the room in which the experiment takes place, to the materials you will use, to the exact words you say to participants - has to be planned carefully and deliberately in advance.  Otherwise, your results may well be influenced by extraneous variables, and your data will not be reliable.


Step 1: Choose your study design

For most studies, an independent samples design is the most practical.  In this design, participants are randomly divided into two groups, and each group is assigned to a different experimental condition.  For instance, if you are replicating Craik and Lockhart's study on levels of processing, one group of participants will complete a shallow processing task, while the second group will complete a deep processing task.  This contrasts with a repeated measures design, in which there is only one group of participants, and everyone participates in both experimental conditions (for example, completing both the shallow and deep processing task).

There are a number of advantages and disadvantages to using an independent samples design, which are outlined below:

  • Advantage - Reduces demand characteristics:  An independent samples design makes it less likely that participants will guess the hypothesis of the experiment.  If participants complete the experiment twice, they might guess the purpose of the experiment, and change their behavior accordingly.  For instance, in the Craik and Lokhart study, participants who have just completed the deep processing task will realize that the study is actually a test of word memory.  Therefore, they may try harder to remember the words during the second, shallow processing task.  

  • Advantage - Avoid order effects:  An independent samples design eliminates order effects.  If participants complete the experiment twice, different results might be obtained between the first and second trial, simply because participants have already grown accustomed to the experiment.  For instance, if the dependent variable is word memory, participants will likely remember more words on the second trial, simply because they have already seen the same words on the first trial.  Of course, you could use two different sets of words, but that introduces the possibility that one set of words might be easier to remember than the other.  Having two groups of participants eliminates order effects entirely, as each participant only completes the experiment once.

  • Disadvantage - More participants are needed:  An independent samples design requires two groups of participants, so twice as many participants are needed.  For your IB Psychology IA, you are required to have twenty participants using an independent samples design, while only ten participants are required for a repeated measures design.

  • Disadvantage - Participant variability.  Since an independent samples design involves two groups of participants, there is always the chance that one group will be different than the other.  For instance, participants in one group might be, on average, more intelligent than participants in the other group.  To reduce participant variability, it is important to randomly assign participants to the experimental conditions.  This helps to reduce researcher bias in choosing which participants should belong to which group.  Furthermore, increasing the number of participants also reduces variability, as larger numbers lower the chance that one group of participants will be much different than the other.


Step 2: Choose participants

Who will take part in your study?  You should consider how variables such as age, gender, language, and culture might influence your results, and what criterion you will set for participant eligibililty.  Consider the following:

  • If your study requires a good command of the English language, you might choose to restrict your study to native or near-native English speakers

  • To reduce participant variability, you might choose to focus on a particular demographic group - for instance, students aged 16-19.

  • It is generally advisable to include participants from both genders and a variety of cultural groups, unless there is a particular reason to the contrary

  • You should consider how other participant variables might influence the results.  For instance, if your participants are Psychology students, they may already be familiar with the study, possibly creating demand characteristics.  Or, if you are replicating Loftus & Palmer, participants who have driven a car for many years may give more accurate speed estimates than those without driving experience.  You may wish to exclude people with certain characteristics from participating in the study if you feel they may distort the results

You should record all the relevant characteristics of each participant.  These could include age, gender, level of English proficiency, Psychology student / non-student, and so forth.  This is also useful to ensure that there are not substantial differences between the participants in each group.  


Step 3: Recruit Participants

Now that you have decided on the sort of participants you are looking for, you now must decide how to actually recruit them to the study.  Review the information on Sampling Methods, and decide which method is most feasible for your study.  The most practical sampling methods are convenience sampling and self-selected sampling.

  • In a convenience sample, you approach people at a particular location (for instance, outside the school cafeteria during lunchtime), and ask if they would like to participate in your study.  This is the easiest way to obtain participants.  However, you will have to consider how the location or time in which you recruit participants might influence who ends up participating in the study

  • In a self-selected sample, you inform people about the study, and ask who is willing to participate.  You might inform potential participants through a class announcement, a poster, or an e-mail message.  Using a self-selected sample ensures that participants are motivated and willing to take part.  However, you may struggle to recruit enough participants, especially since you cannot offer payment as an incentive for participation.  Also, you should consider if the people who volunteer may be different from those who don't volunteer


Step 4: Control variables

In an experiment, the goal is to manipulate the independent variable while keeping all other variables the same.  That way, any difference in the dependent variable can be attributed to the independent variable, rather than some other factor.  You should think carefully about any other variables that might influence the results, and do your best to keep these fixed for all participants.  Here are some examples:

  • Obviously, the instructions you give to participants, and all materials used in the study, must be exactly the same for each participant

  • Think carefully about how the environment in which the study takes place might influence the results, and do your best to standardize this.  For example, you might choose to carry out the experiment in an empty classroom.  Factors like lighting, noise, and the presence of other people might influence the results, so it is best to keep any distractions to a minimum

  • Even the time of day might be an important factor.  Participants who are hungry and inpatient might produce different results than participants who have just eaten.  Or participants who are tired at the end of a long school day might produce different results than participants who are well rested.  You might choose to set a particular window of time for carrying out your study

  • Since you will be carrying out the experiment in a group of 2-4, you may each decide to recruit participants individually.  However, that raises the possibility that the age or gender of the researcher might influence participant behavior.  For instance, a male participant might be more motivated to impress a female researcher. Furthermore, if some researchers recruit participants with whom they already have a relationship (as friends, family members, and so on), this may also influence the findings.  These are variables that are difficult to control, but should be discussed in the research report



Step 5: Consider Ethics


When a person volunteers to participate in a Psychology experiment, they are trusting you to treat them fairly and ethically.  The field of Psychology has a controversial history of experiments that did not always fully consider the well-being of participants.  That is why the IB program insists on stringent ethical guidelines, to ensure that you don't traumatize your IA participants!

There are three steps you must complete to comply with ethical guidelines:

  • Briefing

These are the verbal instructions you give to participants at the beginning of the study.  You should explain the purpose of the study, and what will be required of participants if they choose to participate.  You must also explain that participation in the study is voluntary, and describe the right of withdrawal - the option to end participation in the study at any time, for any reason.  Finally, you must explain that the results of the study will be kept confidential - you will not share the data with anyone else.  To ensure confidentiality, you should assign a number to each participant, so that no actual names appear in the results spreadsheet. Depending on the study, you may not want to give participants too much information on the true aim of the study in advance - that is fine, as long as you debrief participants in the end.  For example, if you are carrying out a replication of Loftus and Palmer, it is okay to simply tell the participants that the study is about "perception and memory" in the briefing.

  • Signed letter of consent

After the briefing, you will give participants a copy of a "Letter of Consent" for them to sign.  The letter of consent will be similar to the briefing, but in written form.  The letter of consent should contain information on:  (a) the purpose of the experiment, (b) what participants will be required to do, (c) any risks associated with the experiment, (d) the right of withdrawal, and (e) confidentiality.  You must keep signed letters of consent from all participants.

  • Debriefing

The debriefing takes place at the end of the study, and is done verbally.  If you have used any deception or withheld any information on the true aim of the study, you must now give a full and complete account of what the experiment was really about.  Once participants know the true aim of the study, you must give them the option of withdrawing their data.  In other words, if participants are no longer comfortable with the study, they have the right to tell you not to use their data in the final analysis, and the data must be destroyed.  Finally, you may ask participants not to tell anyone else about the true aim of the experiment, in order to avoid demand characteristics in other participants.  Thank everyone for their participation, and you are one step closer to completing your study!



Checklist

  • I understand the strengths and limitations of different study designs, and can explain why I have chosen either an independent samples or a repeated measures design

  • I have decided which criteria to use in selecting participants

  • I have decided how to recruit participants to my study, and can explain the strengths and limitations of the sampling method I have chosen

  • I can discuss a number of extraneous variables which may influence my results, and have developed a plan to control as many of these as possible

  • I have prepared briefing and debriefing notes, as well as a letter of consent, which fully informs participants on the nature of the study, confidentiality, and the right to withdraw
Quiz Yourself!

1.  Which statement is true regarding an independent samples design?

(a) It does not require as many participants as a repeated measures experiment

(b) It avoids the problem of participant variability

(c) Demand characteristics are often a problem with this design

(d) It eliminates order effects


2.  How can participant variability between experimental groups be reduced?

(a) Random assignment to experimental conditions

(b) Including a balance of male and female participants

(c) Including a complete briefing and debriefing

(d) Controlling extraneous variables, such as the environment in which the study takes place


3.  You replicate Loftus & Palmer, and one of your participants has already learned about the study in her Psychology class.  What problem might this cause?

(a) Lack of informed consent

(b) Demand characteristics

(c) Researcher bias

(d) Order effects


4.  If you recruit participants by sending out a mass e-mail to all students at your school which asks them to participate in your study, what sampling method are you using?

(a) Self-selected sample

(b) Convenience sample

(c) Random sample

(d) Stratified sample


5.  In the briefing you give to participants, what are you NOT required to reveal?

(a) The right to withdraw from the study at any time

(b) Information about how confidentiality will be ensured

(c) Any possible risks of participating in the study

(d) The full aim and objectives of the study


6.  Extraneous variables are variables which.....

(a) Have no influence on the results

(b) Must be manipulated

(c) Must be measured 

(d) Must be controlled
Answers


​1 - D, 2 - A, 3 - B, 4 - A, 5 - D, 6 - D