There are many different ways that psychologists might collect information from or about participants' thoughts and behavior. Each one of these could be used in a variety of different kinds of studies - but we will start by distinguishing observations, self-reports and physiological measures.
Naturalistic Observation
A researcher unobtrusively collects information without the participant's awareness
Examples:
Research Psychologists for the National Highway Traffic Administration conduct seatbelt observations to determine whether use varies by age, sex, race, geographic location, and type of vehicle. Knowing this helps design programs that focus on those groups and areas with the lowest observed use.
Drain and Engelhardt (2013) observed six nonverbal children with autism's evoked and spontaneous communicative acts. Each of the children attended a school for children with autism and were in different classes. They were observed for 30 minutes of each school day. By observing these children without them knowing, they were able to see true communicative acts without any external influences.
READ: Chimpanzees -- for another example on naturalistic observation
Advantages:
The researcher does not influence the participants’ behavior, so it may be more representative than if people knew they were being watched
Can provide interesting descriptive data
Disadvantages:
Not everything can be observed in its natural environment
Generally time-consuming and expensive
Behavior has to be interpreted by the researcher, and two researchers might see things differently
The researcher generally knows very little about the ‘participants’
Individuals do not have the opportunity to decide whether or not they want to participate
Structured Observation
Researchers can set up a situation and observe that participant's behavior
Example:
READ: Bethesda Company Tests Honesty of Tea Drinkers- In this study, notice how it is structured, rather than naturalistic, because the experimenters are setting up a situation to see how people act - whether they pay for the drink or just take the drink for free.
Advantages:
The researcher has more control of the situation and can keep most variables under control
The researcher can collect other information from the participants
Disadvantages:
The participant is aware of the researcher and may be influenced by the artificial situation or his/her beliefs about what the researcher expects to find
The researcher may still need to interpret the behavior
Self-Report
Participants are asked to provide information or responses to questions on a survey or structure assessment
Example:
Educational psychologists can ask students to report their grade point average and what, if anything, they eat for breakfast on an average day. A healthy breakfast has been associated with better academic performance (Digangi, 1999)
Advantages:
Surveys are inexpensive and efficient ways to collect a lot of data
They are easy to create and score
Computers allow data collection from participants all over the world
Disadvantages:
There are often differences between what people really think and do and what they believe or want the researcher to believe
Participants may misinterpret a question or the answer choices
Participants may be unable or unwilling to answer the questions
Psychophysiological
We can also use technological devices to measure what is taking place in the body (e.g., heart rate, levels of hormones, areas of brain activity).
Example:
To measure the level of anxiety experienced by people, Leary (1986) measured heart rate while participants had a conversation with a stranger. Shy people were anxious during the interaction and had much faster heart rates. However, if participants were told that a noise in the room might make the conversation difficult they experienced less of an increase in heart rate.
Researchers have found that levels of oxytocin are associated with social bonding and empathy. By collecting saliva or blood researchers found that oxytocin levels increase after receiving a hug (e.g., Light et al., 2005).
Aron (2010) found, using an fMRI machine, that there are different areas in the brain that are active when we are experiencing lust vs. when we are experiencing love.
Advantages:
These are direct measures of the body that do not require subjective judgments from the researcher.
Participants cannot willfully influence or hide their body's reactions
Disadvantages:
These can be expensive and time consuming, requiring special machines or processing.
They generally cannot be collected unobtrusively.
Archival
Researchers can examine data that has already been collected for other purposes.
Examples:
Anderson (1987) tried to find the relationship between uncomfortably hot temperatures and aggressive behavior, which was then looked at with two studies done on violent and nonviolent crime. Based on previous research that had been done by Anderson and Anderson (1984), it was predicted that violent crimes would be more prevalent during the hotter time of year and the years in which it was hotter weather in general. The study confirmed this prediction.
Zullow and Seligman (1988) measured levels of pessimism and optimism in Presidential nomination acceptance speeches and succeeding campaigns. They found that the candidate who mentioned hope more in their speeches had won 9 out of the 10 elections that were studied.
Advantage:
Researchers can conduct an analysis of very large data sets without the time and expense of collecting all the data.
Disadvantages:
Data may be missing, incomplete, filled with errors, or difficult to access.
Sometimes researchers are interested in how things change over time. There are two general approaches to this, longitudinal studies and cross-sectional studies. Many of the above methods for collecting data can be applied in these kinds of studies.
Longitudinal Studies
This approach is to recruit a sample of participants and track them for an extended period of time.
Examples:
In a study of a representative sample of 856 children Eron and his colleagues (1972) found that a boy’s exposure to media violence at age eight was significantly related to his aggressive behavior ten years later, after he graduated from high school.
To study predictors of marital satisfaction Huston and colleagues (2009) have followed 169 couples who married in 1981. They found that 35% had divorced, 20% were still married but unhappy and, overall, satisfaction had decreased over time for the average couple.
Advantage:
Researchers can track changes over time.
Disadvantages:
Studies take a long time to conduct, sometimes decades.
“Attrition” – Not all participants who begin the study will complete it… some may drop out, die, or may be hard to find.
“Selective Attrition” – There might be something that causes particular participants to drop out of the study, so the sample of participants at the end of the study may be different in some important way from how it started. For example, if researchers wanted to know whether having kids affects marital satisfaction their results might be influenced if many of the happy families with children dropped out from the study because they were doing other fun things together.
Cross-Sectional Studies
Researchers can gather participants of different ages and look for differences between the groups.
Examples:
Psycholinguistic (language) psychologists could study groups of 1st through 8th grade students to determine the age at which children begin to use swear words.
In 1996, Russell surveyed people of varying age groups and found that people in their 20s tend to report being lonelier than people in their 70s.
Advantages:
Researchers can study the effect of time or experience in much less time than a longitudinal study would take.
Attrition from the sample is not an issue.
Disadvantage:
“Cohort Effects” – The differences between the two groups might have been caused by something other than time. Perhaps that generation had a different experience, or there was some event that affected that particular group of participants. For example, imagine that our swearing study found that swearing was much more common in 6th graders than in 4th graders. It could be that swearing naturally develops in that time. However, the 4th graders may, as a cohort, have been exposed to different parenting trends early on than the current 6th graders. Swearing might have been just as low for the 6th graders if they had been raised the same way. Thus, the results could have been caused by a cohort effect.
Just as there are different ways that researchers can collect data, there are also different kinds of studies that they can design. We will focus on three different designs - case studies, correlational studies and controlled experiments. Any of the methods we have already covered above might be used in any of these study designs.
Case Studies
Researchers can conduct a detailed analysis of a particular person, group, business, event, etc. We tend to use this approach to learn more about especially interesting or rare examples with the goal of describing that particular thing.
Examples:
Ted Bundy was one of America's most notorious serial killers who murdered at least 30 women and was executed in 1989. Dr. Al Carlisle evaluated Bundy when he was first arrested and conducted a psychological analysis of Bundy's development of his sexual fantasies merging into reality (Ramsland, 2012). Carlisle believes that there was a gradual evolution of three processes that guided his actions: fantasy, dissociation, and compartmentalization (Ramsland, 2012).
READ: Imagining Ted Bundy for more information on this case study.
Researchers might study a successful company to learn about its management, hiring and training practices in the hopes of discovering things that explain their success.
Advantages:
Researchers can study unusual, rare, or difficult-to-find participants or events.
In-depth study of a particular case can reveal interesting areas for future study.
Disadvantage:
What happens with one case may not generalize to other cases, so the results are generally limited to describing that particular thing.
Correlational Designs
In correlational designs two different variables are measured to determine whether there is a relationship between them. We will get into more detail on this in a later section, but the key here is that we are limited to measuring variables.
Examples:
Thornhill et al. (2003) had people rate how physically attractive they found other people to be. They then had them separately smell t-shirts those people had worn (without knowing which clothes belonged to whom) and rate how good or bad their body odor was. They found that the more attractive someone was the more pleasant their body odor was rated to be.
Controlled Experiments
Researchers create a controlled environment in which they can carefully manipulate at least one variable to test its effect on another. The key here is that the researchers can cause a change in one variable.
Examples:
Clinical psychologists can test a new pharmaceutical treatment for depression by giving some patients the new pill and others an already-tested one to see which is the more effective treatment.
You already learned about the testing effect - participants who were given a practice test remembered more a week later than those who were given more time to study.