Chapter 2: The Research Enterprise in Psychology
Hypothesis - An unconfirmed statement of prediction relating two or more variables.
Variables - Measurable conditions, events, characteristics, or behaviour that are controlled (independent variable) or observed (dependent variable) in a study.
Theory - Set of interrelated ideas used to explain a set of observations. A good theory will generate a host of testable hypotheses.
Empirical - A method of research that can be verified by observation and experiment.
THE SCIENTIFIC APPROACH TO BEHAVIOUR
The scientific approach assumes that there are laws of behaviour that can be discovered through empirical research.
Goals of the Scientific Enterprise
- Psychologists & scientists share three sets of goals for research:
- Measurement & Description - Identify & develop method of measurement.
- Understanding & Predicting - Hypothesize about the relationship between two variables. Understand the explanation behind relationship between two variables.
- Application & Control - Information should increase our control on a phenomenon & apply it to practical situations.
- Another goal of research is work with existing theories, or to propose new ones. Theories help scientists integrate their information into a coherent whole (the big picture). Theories also help guide future research questions. Proving or disproving a hypothesis that is derived from a theory can either strengthen or weaken the theory, respectfully.
Advantages of the Scientific Approach
Clarity & Precision in Communication
- Scientific approach requires people to specify exactly what they are researching.
- Reduces ambiguity and vagueness.
Intolerance of Errors
- Scientists are trained to be skeptical.
- They scrutinize one another’s findings critically.
- They demand verifiable & objective data before accepting any conclusions.
- They continue to ask & answer questions, even about “settled” science.
Steps in a Scientific Investigation
- Formulate a Testable Hypothesis
- Hypothesis must be precise & variables must be operationally defined.
- Operational definition describes the actions that will be used to measure/control a variable.
- Select the Research Method & Design the Study
- Figure out how to empirically test the hypothesis.
- Various Methods - Experiments, case studies, surveys, naturalistic observation, etc.
- Collect Data
- Various Methods - Direct observation, questionnaire, interview, archives, physiological recordings (blood pressure, heart rate, muscle tension, brain activity, etc.), psychological tests (mental abilities & personality traits).
- Analyze Data & Draw Conclusions
- (Usually) convert data into numbers.
- Use statistics to analyze. Statistics is a major part of the scientific enterprise.
- Report Findings
- Write a concise summary of the study & its findings.
- Submit to journal(s)
- Potentially prepare a report for a scientific meeting or conference
- All of this is done so others may evaluate and critique findings, as well as verify.
LOOKING FOR CAUSES - EXPERIMENTAL RESEARCH
Experiments - Are the most popular research method where the investigator manipulates a variable under carefully controlled conditions & observes whether any changes occur in a second variable. This allowed researchers to determine cause-and-effect, which is usually difficult to prove.
Independent Variable - Is a condition that an experimenter varies in order to see its impact on the dependent variable.
Dependent Variable - Is a condition that is thought to be affected by the manipulation of the independent variable. It depends on the independent variable.
Experimental & Control Groups
In a typical experiment, there are two groups of subjects/participants who are treated different with regard to the independent variable:
Experimental Group - Subjects receive special treatment in regard to independent variable.
Control Group - Subjects receive no special treatment.
It is crucial that both groups are alike before the experiment begins. If the two groups are alike, except for the variation created by the manipulation of the independent variable, then any differences between the two groups must be due to changing independent variable.
Researchers will randomly assign individuals to a control group or an experimental group!
It is impossible to make sure both groups are alike because there will always be differences. Instead, researchers will only make sure that extraneous/secondary/nuisance variables are alike between the two groups.
Extraneous Variables - Are any variables (other than the independent variable) that seem likely to influence the dependent variable.
This is done to avoid any confounding of variables. Confounding of variables occurs when two variables are linked together in a way that makes it difficult to attribute specific effects to one or the other. By making sure that both groups’ extraneous variables are alike, researchers are assured than any changes to the dependent are the result of manipulation to the independent variable.
Variations in Designing Experiments
- Within-Subjects Design - Using the same set of subjects as the experimental & control group.
- Require fewer participants
- Does a better job of standardizing control variables & extraneous variables.
- Between-Subjects Design - Using different set of subjects as the experimental & control group.
- Manipulation of 2+ IVs - Manipulation of two or more variables to examine joint effect on DV.
- Study of 2+ DVs - Researchers often study a number of dependent variables to get a better picture. (Questionnaire & behaviour measure, for example).
- Permits conclusions of cause-and-effect because price control allows for isolation
- Experiments are often artificial due to the fact that great amount of control is needed. This can make it difficult to generalize your findings. This can be countered using field studies where researchers lose some control to gain generalizability.
- Not all research questions can be answered due to ethical limitations & feasibility. For example, you cannot take babies away from families to bring them up in certain ways.
LOOKING FOR LINKS - DESCRIPTIVE/CORRELATIONAL RESEARCH
These research methods cannot prove cause-and-effect & researchers have no control over variables. However, these research methods make it easier to answer questions that would be unethical to answer using the experimental method.
In naturalistic observation a researcher engages in careful observation of behaviour without intervening directly with research subjects or participants who behave in a natural environment.
- Less artificial
- Good starting point when little is known about the behaviour that is being studied
- Can be used to study animal behaviour (Jane Goodall’s work on chimpanzees’ social life)
- Reactivity – when a subject’s behaviour changes due to the presence of an observer
A case study is an in-depth investigation of an individual subject.
- Well suited for psychological disorders & neuropsychological issues (Henry Molaison’s brain was used a case study to help understand how memory works within the brain).
- Provide compelling, real-life illustrations that bolster a hypothesis or theory
- Highly subjective
- Information must be knit together in an unsystematically subjective manner
- Easy for investigators to see what they expect to see (due to how information is collected)
Researchers use questionnaires/interviews to get info about specific aspects of subjects’ behaviour
- Easy to collect lots of qualitative data (such as opinions & experiences)
- Can obtain info on behaviour that is difficult to observe directly (sexual harassment)
- Relies on self-reported data
- Intentional deception, wishful thinking, memory lapses & poorly worded questions can all make survey data unreliable.
- Potential of sample bias (not representative of the population)
- Gives researchers a way to explore questions that could not be examined via experiments. It does this by giving more leeway to understanding questions that would have otherwise harmed subjects
- BROADENS THE SCOPE OF PHENOMENA THAT RESEARCHERS CAN STUDY
- Investigators cannot control events to isolate cause-and-effect.
- RESEARCHERS CANNOT PROVE CAUSATION
LOOKING FOR CONCLUSIONS - STATISTICS & RESEARCH
Statistics - Is the use of math to organize, summarize & interpret numerical data, which allows researchers to draw conclusions based on their observations.
Descriptive Statistics - Used to organize & summarize data, to provide an overview of the data. There are 3 key descriptive statistics:
- Central Tendency - Averages (mean, median & mode).
- Variability - Refers to how the scores in a data set vary from each other.
- Standard Deviation is a common way of computing variability. 68% of data point lie within ±1 while 95% lie within ±2.
- Normal curve, normal distribution or the bell curve is a symmetrical curve that represents the pattern in which many human characteristics (height, nose length, etc) are dispersed in the population.
- Correlation - When two variables are related to each other.
- Correlation coefficients are numbers depicting the degree of relationship between 2 variables. Indicates direction of relationship & strength of relationship. Stronger correlation coefficients (closer to ±1) increases the ability to predict.
Inferential Statistics - Used to interpret data & draw conclusions using the laws of probability. Researchers use inferential statistics to evaluate whether results might be due to chance. If it’s concluded that it is not likely that chance caused the result, then the results are statistically significant (when fewer than 5 chances in 100 could be due to chance).
LOOKING FOR FLAWS: EVALUATING RESEARCH
Replication is the repetition of a study to see whether the earlier results are duplicated. Replication is key to scientific research.
Meta-Analysis - Is the combination of the statistical results of many studies that were answering the same question, to yield an estimate of the size & consistency of a variable’s effects. It can test the generalizability of findings, strength of correlation, check the reliability of a finding, etc.
Sampling Bias - Is when a sample is not representative of the population from which it was drawn.
Placebo Effects - Occur when participants’ expectations lead them to experience some change even though they receive empty, fake or ineffectual treatment.
Nocebo effects occur when participants’ expect a certain variable to not cause them to experience a change, which can lower the influence of a real treatment.
Distortions in Self-Report Data
- Social Desirability Bias - A tendency to give socially approved answers to questions
- Response Set - A tendency to respond to questions in a way that in unrelated to the content of the questions. For example, people tend to agree with nearly everything on a survey.
- Halo Effect - A tendency to allow an impression of a certain aspect of someone to influence the impression in another aspect. For example, smart people are often thought to be good.
Experimenter Bias - Occurs when a researcher’s expectations or preferences about the outcome of a study influence the results obtained.
Mistakes during data collection tend to heavily favour the researcher’s hypothesis. Countered by not telling the experimenter nor the subject which group is the control or experimental (or by not telling them the hypothesis).
LOOKING AT ETHICS - DO THE ENDS JUSTIFY THE MEANS
Researchers should not torture, should be honest, etc. Those are easy to understand.
But can researchers deceive subjects? Can researchers inflict any harm on subjects? Should researchers use animals to conduct harmful and unethical tests?
Ethic Principles for Psychology Research in Canada
- Respect for Dignity of Persons.
- Responsible Caring of rights, privacy, liberty, self-determination, etc.
- Integrity in Relationships (students, subjects, researchers, etc).
- Responsibility to Society to increase knowledge & promote welfare.
Featured Study - Can Fear Increase Sexual Attraction?
Participants were males between the ages of 18-35 who chose to cross alone either of 2 bridges spanning a river. One bridge was more fear-arousing than the other. After crossing the bridge, they were randomly approached by either a male or a female (actors). The actors offered their phone number & explained to the subject that they may call later in the day if they have any questions about the study.
Participants who crossed the high bridge & who were met by the female confederate evidenced more sexual arousal in their TAT stories than did those who crossed the low bridge (average scores of 2.47 versus 1.41). In addition, a higher proportion of males who took the confederate’s phone number called the female confederate back if they had crossed the high as compared to the low bridge. This shows that the men mislabeled their fear as sexual attraction.
Personal Application - Finding & Reading Journal Articles
- Journals publish technical and scholarly material. Usually they are written for other professionals in a narrow area of inquiry. In psychology, most journal articles are reports of original research.
- PsycINFO is a computerized database that contains brief summaries of newly published journal articles, books, and chapters in edited books. Works on specific topics and publications by specific authors can be found by using the search mechanisms built into the database.
- Computerized literature searches can be much more powerful and precise than manual searches.
- Journal articles are easier to understand if you are familiar with the standard format. Most articles include six elements:
Elements of Articles
- Abstract - A concise summary at the beginning of each article. This allows readers to quickly decide whether articles are relevant to their interests or not.
- Introduction - The intro presents an overview of the problem studied in the research. It mentions relevant theories, quickly reviews previous research, cites shortcomings in previous research that justifies the present study, details the hypothesis, etc.
- Method - A thorough description of the research methods used in the study. Includes participant information, procedures followed, data collection techniques used, etc. This allows for others to replicate the study.
- Results - Data is reported in this section, and includes complex statistical analyses, figures, tables & graphs. It does not include inferences based on the data.
- Discussion - Includes conclusions by the author(s), and any further interpretation & evaluation of the data. Implications for theory & factual knowledge in the discipline are discussed. Limitations may be acknowledged, and may include suggestions for future research on the issue.
- References - A list of bibliographical references for any studies cited.
Critical Thinking Application
Anecdotal Evidence - Consists of personal stories about specific incidents & experiences.