Choosing a Study for your IA

The study you carry out for your IA will be a modification of an existing study.  This can be a study you have learned in your IB Psychology class, or any other Psychology study you are familiar with.  You are allowed to make changes to the original study to make it easier to carry out - for example, you may have fewer participants, or just two levels of the independent variable.  

Choosing an appropriate study to replicate can make or break your IA.  Some studies are very complicated or difficult to replicate - even the most motivated students may find these nearly impossible to carry out.  Furthermore, if your study violates the IB code of ethics, you will automatically fail your IA.  Make sure you discuss your choice of study with your teacher, and listen to any advice before you proceed.

See below for four great choices for your IA.  Of course, you might decide to be adventurous and choose a different study - but the list below will give you an idea of what type of study is appropriate.  If you do decide to choose a different study, it must meet all the criteria listed below.

  • One independent variable, which can be manipulated.  This is something that is supposed to have an effect on the dependent variable.  For example, you might predict that a different type of processing task - a "deep" rather than a "shallow" task - will result in better memory for a list of words.  The independent variable must be something that you can manipulate - so gender, culture, age, and so forth don't qualify, because you can't change a participant's gender, culture, or age. 

  • One dependent variable, which can be quantitatively measured.  This is the variable that you are measuring, which you believe may be influenced by the independent variable.  In order to perform statistics on the dependent variable, it has to be something that you can measure with a number.
  • No placebos, ingestion, or deprivation.  For ethical reasons, you are not allowed to ask participants to consume any substance (ex. caffeine or alcohol) or deprive them of anything (ex. require that they not sleep for a certain length of time before the experiment)

  • No anxiety, stress, or pain to participants.  Any study which causes physical harm or pain to participants is obviously not allowed.  Furthermore, you must also consider the emotional well-being of participants - any experiment which may cause participants to feel embarrassed or anxious is not allowed

  • No conformity or obedience studies.  These studies involve potentially embarrassing or stressful situations, and are therefore not allowed

  • Minor deception allowed, but must be justified.  For some experiments, it may be important to keep the true aim and purpose of the experiment a secret from participants until after the experiment is finished.  For example, if your experiment is about the effects of leading questions on memory, you won't want participants to know this in advance!  You may even wish to tell participants that the study is about memory, when in fact it is really about leading questions.  Such minor deception is allowed, providing you explain why it is necessary, and it does not cause any embarrassment or harm to participants.

  • No animal studies.  Animal research is a controversial issue, and so no animal research is allowed.

  • Not involving children under the age of 12.  Children under this age may not understand what it means to participate in an experiment, and so you are not allowed to use them as participants.  Any study which requires the use of very young children may not be replicated.  Furthermore, if you wish to use participants aged 12-15, you must obtain parental consent in advance.

Study #1: Loftus & Palmer

Theory:  This study is based on the theory of reconstructive memory.  Rather than functioning like a video camera, which records every detail of the past,  memory is inexact, malleable, and open to suggestion.  Therefore, leading questions can alter memories of an event.

The study:  Participants watched a video of a car accident.  Afterwards, they were asked a number of questions about the accident, one of which was "How fast were the cars going when they hit / collided / bumped / smashed into each other?", with each group of participants asked the question with a different verb.  Participants who were asked the question with the more forceful verb (eg. "smashed") estimated a higher speed than participants who were asked with a less forceful verb (eg. "bumped").

Independent variable:  The verb in the key question about the speed of the cars

Dependent variable:  Estimate of speed

Suggested procedure:  It is easiest to use an independent samples design with two groups of participants, so you will choose just two of the verbs to use in the questions (eg. "smashed" vs. "bumped").  You can use any video of a car accident, for instance, one that you find on You Tube.  In order to prevent participants from guessing the aim of the experiment, you could ask a number of other questions (eg. "What color were the cars?", "Whose fault do you think the accident was?") before asking the critical question.

Hints:  In your debriefing, you must explain the true aim of the experiment.  Make sure to remind participants not to discuss this study with anyone else, in order to prevent other participants from knowing the aim of the study in advance.  The results of this experiment may be influenced by whether participants have driving experience or not - those without driving experience may not have a good sense of driving speed.  You may wish to restrict the study to participants with driving experience, or if this is impractical, at least collect data on whether participants have driving experience or not (eg. ask participants how many years they have been driving).

See video below for another example of a Loftus & Palmer replication.

Study #2: Kahneman & Tversky

Theory:  This study is based on dual processing theory, which states that there are two modes of processing - System 1 thinking, which is fast, automatic, and unconscious, and System 2 thinking, which is deliberate, conscious, and effortful.  In order to make rapid decisions, System 1 makes use of mental shortcuts, caused heuristics.  One of these heuristics is called the "anchoring effect", in which having a particular number of mind (the "anchor") acts like a starting point for making estimates.

The study: There are a large number of studies carried out on the anchoring effect, but all have some common features.  First, an anchor is randomly established.  It could be the last two digits of participant's telephone number, or a number chosen at random from spinning a roulette wheel.  Then, participants will be asked a question which they probably don't know the exact answer to, requiring them to estimate the answer.  For example, in one study, participants were asked to estimate the percentage of African nations in the UN, and in another study, participants were asked to estimate the age of Gandhi when he died.  Participants with a higher anchor tend to give a higher estimate than participants with a lower anchor.

Independent variable:  The anchor (either high or low)

Dependent variable: Participants' estimated answer in response to a question

Suggested procedure:  For simplicity, it is best to use an independent samples design with just two groups:  a high anchor group, and a low anchor group.  You can instruct participants to pick a slip of paper from a hat, which will contain a number between 1 and 99.  In reality, you can ensure that all the slips have the same number, for instance 87 (in the high anchor condition) or 17 (in the low anchor condition).  Then, you need to think of a question - any question - which participants will have to estimate the answer to.  (For instance - "How many films has Bradd Pitt acted in?")  First, ask participants, "Is the answer to the question higher or lower than the number you have chosen?"  Then, afterwards, ask participants for their best estimate.

Hints:  In your debriefing, you will need to tell participants the true aim of the experiment - make sure to remind them not to discuss the study with others.  Choose a question for which participants probably won't know the exact answer, but will at least be able to make a rough estimate.  The correct answer to the question should be roughly between the value of the "low anchor" and "high anchor".

See the video below for more information on the anchoring effect.

Study #3: Cunitz and Glazner

Theory:  This study is based on the multi-store model of memory, which states that memories are kept in one of three stores.  When you first encounter information, it is briefly held in the sensory store, and if you pay attention, it will enter your short term memory.  Your short term memory can only hold around 7 items for a span of a few seconds, and requires constant rehearsal in order to retain the memory.  With enough rehearsal, the information will enter long-term memory, where it is stored indefinitly.  Researchers have found that when you read a list of words aloud to participants, they will have better memory for the first few words in the list (known as the "primacy effect"), as well as the last few words in the list (known as the "recency effect").  The explanation for these findings is that the first few words have been rehearsed long enough to enter long term memory, while the last few words are still kept fresh in short term memory.

The study:  Cunitz and Glazner carried out a number of related studies to investigate the multi-store model of memory.  In one study, one group of participants were read aloud a list of words, and then immediately recalled as many words as possible.  The second group were read aloud the same list of words, however they had to perform a "distraction task" for 30 seconds before recalling the words.  The "distraction task" eliminated the recency effect, as the words at the end of the list were no longer available in short term memory.  On the other hand, the primacy effect was still evident - participants still had better memory for the first few words of the list.

Independent variable: Distraction task vs. immediate recall

Dependent variable:  Number of words recalled (suggestion:  to focus on the recency effect, you may choose to measure recall of only the last 10 words)

Suggested procedure: The easiest way to carry out this experiment is to use an independent samples design, with one group of participants given a distraction task, and the other group recalling the words immediately after they are read.  You should use the same list of words - perhaps 30 or so - for both groups.  It might be a good idea to record yourself reading the words, and then play the recording, in order to eliminate any variation in how the words are read.  For the distraction task, you can ask participants to count backwards from 100 by 7s (eg. 100, 93, 86, 79, etc). or any similar task that requires some degree of mental effort.  Finally, ask participants to write down as many of the words as they remember (in any order), and see how many of the last 10 words in the list are recalled.

Hints:  You will want to choose words that are neither too easy nor too difficult to remember.  Make sure to carry out the experiment in a quiet place, free from distraction.  The results of the experiment may be influenced by the English proficiency of participants - you may want to ask participants for their level of English in order to ensure that neither group is more fluent than the other.

Study #4: Craik & Lokhard

Theory:  This study is based on levels of processing theory, which predicts that the "deeper" you process information, the better you will recall it.  Three levels of information processing include structural, in which you focus on how a word looks like; phonetic, in which you focus on how a word sounds; and finally semantic, in which you focus on the meaning of the word.  Structural processing is shallowest, whereas semantic processing is deepest.

The study: Participants are shown a sequence of words, and given either a structural, phonetic, or semantic task to perform on each word.  For example, for the structural task, participants are simply asked to reply whether the word is written in upper case or lower case letters.  For the phonetic task, participants are asked whether or not the word rhymes with a second word (eg. "Does each word rhyme with 'fight'?).  Finally, for the semantic task, participants are asked a question which requires them to think of the word's meaning (eg. "Would you find this item in a school?")  Afterwards, participants are asked to write down as many of the words as they can remember.

Independent variable:  Type of processing task 

Dependent variable:  How many words are recalled

Suggested procedure:  The simplest way to carry out the procedure is to use an independent samples design, with two groups of participants - one given a structural task, and the other given a semantic task.  You should use the same list of words - perhaps 20 or so - for both groups. It might be a good idea to display the words on a Power Point presentation or something similar, which is set to display the next word at a fixed time interval, to ensure that all participants are shown the words in exactly the same manner.  

Hints:  You will want to choose words that are neither too easy nor too difficult to remember.  Make sure to carry out the experiment in a quiet place, free from distraction.  You will need to explain the true aim of the experiment during the debrief, but you should remember to ask participants not to disclose this to anyone else.  Finally, the results of the experiment may be influenced the level of English proficiency of participants - you may want to ask participants for their level of English to ensure parity between the two experimental groups