Research Methods to Assess Claims. Scientific Attitudes

Transcription

1 LP 1D descriptive and exp methods 1 Scientific Attitudes Science as a process for investigating claims Hypothesis o Description o Examples Operational Definitions o Description o Examples o Not all concepts can be operationalized Empirical Evidence o Description o Examples Testability o Description o Examples Falsifiability o Description o Examples o Examples beliefs can influence perception of falsifiability Problems With Casually Testing Beliefs Examples where beliefs influence what you see and don t see Why do we need a scientific method when the outcome seems obvious? Problems with Casually Testing Beliefs 1. We tend to look for information that confirms our beliefs 2. We use different standards for evaluating behavior 3. We minimize inconsistent information Applying your knowledge broadly what other beliefs can be understood because people casually test beliefs How does science minimize the problems with casually testing beliefs? Why is learning about the problems with casually testing beliefs important? Research Methods to Assess Claims Experimental Methods o Independent variables o Dependent variables o Random assignment o Experimental group(s) o Control group o Difficulties in conducting experiments o Examples of Experiments Descriptive Methods o Naturalistic Observations Description Examples o Case Studies Description Examples o Surveys Description Examples Problems when conducting surveys o Correlations Description Examples Correlations What does a correlation represent? The correlation coefficient o Ranges from -1 to +1 sign magnitude Limits of Correlations o Correlations do not suggest a causal relationship o If there is a causal relationship, correlations do not suggest a direction of the causal relation Why is learning about correlations and the limits of correlation important?

2 LP 1D descriptive and exp methods 2 Research Methods to Collect Empirical Data There are many research questions in science. How you phrase your research question will affect the method you use. Each method has its strength and weakness. The use of multiple research methods help compensate for each individual weakness. The research methods to collect empirical data discussed in your textbook are the following: Experimental methods: A method of investigation used to demonstrate cause-and-effect relationships by purposely manipulating a factor (the independent variable) thought to produce change in a second factor (dependent variable). Key terms: Independent variable (IV) Dependent variable (DV) Random assignment Control group Experimental group Descriptive methods: Scientific procedures that involve systematically observing behavior in order to describe the relationship among behavior and events. naturalistic observation case studies surveys correlational studies

3 LP 1D descriptive and exp methods 3 Experiments: Examining Cause and Effect A method of investigation used to demonstrate cause-andeffect relationships by purposely manipulating a factor (the independent variable) through to produce change in a second factor (the dependent variable). There must be at least two groups (an experimental and control group to compare).

4 LP 1D descriptive and exp methods 4 Difficulties in Conducting Experiments Like with many research methods, there are other factors affecting people s behavior than the ones you take into account. Some common problems that social scientists have to take into account are the following. Pre-existing differences: To reduce the effect of preexisting differences between groups, researchers randomly assign participants to the control or experimental group so that individual differences are not likely to be a factor in any differences in the outcome (the dependent variable). Demand characteristics: Cues that participants pick up about the hypothesis of a study or about how they are to behave. Participants belief about the study will unconsciously change their behavior to match that of the hypothesis. o For example, if participants are told to drink alcoholic drinks that don t smell or taste like alcoholic drinks will rate their behavior differently (eg. less inhibited) Expectancy effects: The subtle and unintentional ways researchers influence their participants to respond in a manner that is consistent with the researcher s hypothesis. o For example, researchers might be more empathic to participants who receive a drug that is suppose to help them. Double-blind studies are used to reduce the expectancy effects.

5 Experimental Methods LP 1D descriptive and exp methods 5 A method of investigation used to demonstrate cause-andeffect relationships by purposely manipulating a factor (the independent variable) through to produce change in a second factor (the dependent variable). There must be at least two groups (an experimental and control group to compare). Examples of Experiments: Does watching violent television lead to more aggressive behavior? Two groups randomly assigned: one group watches 24 (violent TV) and the other watches Sesame Street (nonviolent TV); compare the probability of violent behavior by members of the two groups o IV = type of television program o DV = violent behavior Does oat bran reduce cholesterol levels? o IV = presence or amount of oat bran o DV = cholesterol levels Does taking vitamin C reduce the number of colds? o IV = o DV = Which of two new anti-depressants is more effective in treating major depression? Developmental psychologists want to know if exposing children to public television improves their reading skills. A pizza parlor wants to know if changing the number of toppings increases sales.

6 LP 1D descriptive and exp methods 6 A Social psychologist wants to know whether being polite or rude to people tends to make them more cooperative. Strengths: Cause-and-effect relationships can be established. Precise control can be exerted over other, potentially confounding variables. Weaknesses: Results can be subject to several sources of bias (eg. experimenter effects or demand characteristics). Generalizability can be doubtful if behavior is observed under highly artificial conditions that may inhibit the natural behavior of subjects and may not generalize to the real-world. In some cases, experimentation may be unethical, such as with patients with brain damage.

7 LP 1D descriptive and exp methods 7 Descriptive Methods: Naturalistic observation The systematic observation and recording of behavior as it occurs in a natural setting.

8 LP 1D descriptive and exp methods 8 Examples of Naturalistic Observation Do people who know someone else is waiting for their parking spot take more time to leave than when no one else is waiting for their parking spot? How do people in elevators behave? How do teenagers behave at the mall? In fifth-grade classrooms, are boys or girls more likely to say answers aloud without raising their hands? To learn what changes take place through childhood, Jean Piaget observed the differences in his children s thinking (Chapter 9: Lifespan Development) Strengths: Behavior is observed in the setting where it normally occurs, compared to the experimental designs that occur in an artificial and controlled setting. It can provide the basis for hypotheses to be tested later. Weaknesses: Cannot be used to establish cause-and-effect relationships Often costly and difficult to perform. Researcher s expectations and beliefs can distort observations. The researcher needs to carefully separate their data from their interpretations of the data. The presence of the researcher may influence the behavior of the subjects being studied (see Hawthorn Effect).

9 LP 1D descriptive and exp methods 9 There is little or no control over the conditions of observation.

10 Descriptive Methods: Case Studies LP 1D descriptive and exp methods 10 A highly detailed description of an individual or small group of individuals or event. Case studies may be performed on a group of people who undergo extraordinary circumstances (such as victims of a bank hostage situation). Usually case studies are performed with extraordinary cases that cannot be studied by any ethical means (e.g. what are the effects of nuclear radiation on an unborn fetus, the effects of torture on obtaining information, or what are the effects when part of the frontal lobes are damaged).

11 Examples of Case Studies LP 1D descriptive and exp methods 11 Researchers prepare an intensive report of the childhood who became a murderer. The police have discovered a 6-year-old who has been kept in a room by herself since birth and not exposed to spoken language. Does the lack of early exposure to language result in an inability to learn language. Phineas Gage had a 3½-foot iron rod go through his skull. What were the effects on his behavior? H.M. had part of his hippocampus damaged. As a result, he could not form new explicit memories (such as meeting you), but he could form new procedural memories (such as how to type). Super-Size Me Strengths: Detailed information is gathered; individuals can be studied for a long period of time. Sources of information are from rare or unusual conditions or events such as serial killers, neglected children, or people with brain damage. Case studies can provide the basis for hypotheses to be tested later. Weaknesses: Objectivity of the research may be compromised. The case being studied may not be representative of the condition or event. Generalizability of the results is uncertain.

12 Descriptive Methods: Surveys LP 1D descriptive and exp methods 12 A questionnaire or interview designed to investigate the opinions, behaviors, or characteristics of a particular group.

13 Examples of Surveys LP 1D descriptive and exp methods 13 How many hours per week do most people watch violent TV programs? How many hours of television do 8-year-olds in the United States watch per day? How is our customer service? How do you feel toward your President? What do you give your boyfriend/girlfriend for St. Valentine s Day? An Industrial/Organizational psychologist tests to see if wearing nametags make the employees happier with their work. A Personality psychologist explores whether extroverted people have more fun at parties than introverted people. Strengths: A large amount of information can be acquired quickly. Accurately predictions of large-scale trends can sometimes be made. Weaknesses: Generalizability may be questionable unless persons surveyed are a representative sample of a larger population. Many call-in or magazine surveys do not have a representative population. Most are selfselected and represent a small slice of the population. Responses provided may be inaccurate due to a person s lack of awareness of their own attitudes and biases.

14 LP 1D descriptive and exp methods 14 The characteristics of the interviewer may influence the response. Interpretation of the answers may be difficult, especially with a biased survey. Surveys lack the ability to explain.

15 Problems with Surveys LP 1D descriptive and exp methods 15 Ideally, survey questions are worded so as not to suggest a desired answer where interviewees will answer in a way that they think the interviewer or sponsoring organization wants to hear, thus making the survey flawed for scientific purposes, (but good for political purposes). Such surveys include questions that are intentionally worded in a biased manner. Examples of good questions: Which candidate do you plan to support? Do you think that increased parking fees would be a good idea or a bad idea? Examples of bad questions: Do you agree that our President is doing the right thing in Don t you think it s terrible the way our political enemies are doing the following Do you think a woman should have the ability to choose Do you think the life of a child is important If you receive a biased survey, you can safely assume that the organization is not actually interested in recording your opinion. Such requests are accompanied by requests, and they serve only as fund-raisers or political gain, not as surveys.

16 LP 1D descriptive and exp methods 16 Problems with Surveys: Unrepresentative Samples You cannot survey the entire population to find out what they think. You can take a smaller sample and have a 95% confidence that your sample is similar to the larger population as long as your sample is a representative sample of the population the sample resembles the larger population. Most call in internet, television and radio show participants have strong beliefs and choose to volunteer their opinions and thus do not resemble the larger population. Examples of non-representative samples: A senator is interested in whether his constituents favor the death penalty. His staff reports that letters about the death penalty have been received from 854 constituents and 654 favor it. A cookie manufacturing company wants to know what percentage of Denver residents make cookies from scratch. A sample of 1,000 residential addresses is chosen and interviewers call these households during regular working hours on weekdays. A newspaper is interested in finding out what proportion of drivers in the city wear seat belts. Some reporters go to a GM plant and record the number of employees who fasten their set belts when they leave work.

17 LP 1D descriptive and exp methods 17 Consequences of an Unrepresentative Sample

18 LP 1D descriptive and exp methods 18 Descriptive Methods: Correlational studies A research strategy measures two factors* and allows the precise calculation of how strong two factors* are related to one another. Unlike experiments, correlations do not manipulate one variable (the independent variable). *correlations can examine more than two factors, but for an introductory course, we will restrict our discussions to two variables.

19 Examples of Correlations LP 1D descriptive and exp methods 19 Is there an association between hours of watching violent TV programs and some measure of violent behavior? Do people exhibit more aggressive behavior when it is hot? Does the score on the pre-quiz predict your final grade in the course? Does the amount spanking effect the frequency of a child s disobedient behavior? Is there a relationship between the number of books in a family s home and the child s school grades? Behavioral and Industrial/Organizational psychologists want to know whether reinforcing comments will make people work harder. Strengths: Large amount of information can be gathered quickly. The method can be used in field as well as laboratory settings. It can assess the strength of a relationship and provides the basis for other predictions. It can rule out some factors that are relevant for prediction and identify others for study. Weaknesses: Difficult to establish cause-and-effect relationships.