Experimental and Control Groups: The Logic of the Scientific Method

1. Experimental and Control Groups:The Logic of the Scientific Method Experimental group Control group Random assignment Manipulate independent variable for one group only Resulting differences in the two groups must be due to the independent variable Extraneous and confounding variables In an experiment, the investigator assembles two groups who are as alike as possible, an experimental group (who receives a special treatment in regard to the independent variable) and a control group (who do not receive the special treatment). Then, after they administer the treatment, if the two groups differ on the dependent variable, it MUST be due to the treatment. An extraneous variable is a variable, other than the independent variable, that may influence the dependent variable. Confounding of variables occurs when participants in one group of subjects are inadvertently different in some way from participants in the other group, influencing outcome. Random assignment of subjects is used to control for confounding variables. In an experiment, the investigator assembles two groups who are as alike as possible, an experimental group (who receives a special treatment in regard to the independent variable) and a control group (who do not receive the special treatment). Then, after they administer the treatment, if the two groups differ on the dependent variable, it MUST be due to the treatment. An extraneous variable is a variable, other than the independent variable, that may influence the dependent variable. Confounding of variables occurs when participants in one group of subjects are inadvertently different in some way from participants in the other group, influencing outcome. Random assignment of subjects is used to control for confounding variables.

2. Experimental Research:Looking for Causes Experiment = manipulation of one variable under controlled conditions so that resulting changes in another variable can be observed Detection of cause-and-effect relationships Independent variable (IV) = variable manipulated Dependent variable (DV) = variable affected by manipulation How does X affect Y? X = Independent Variable, and Y = Dependent Variable An experiment is a research method where there is manipulation of one variable under carefully controlled conditions so that resulting changes in another variable can be observed�key word �resulting.� Experiments are very powerful in that they allow for detection of cause-and effect relationships�Does X cause Y? The IV is the variable that the experimenter controls or manipulates�the DV is the variable thought to DEPEND (at least in part) on manipulation of the IV. If we wanted to know how X affects Y, X would be the IV, and Y would be the DV. An experiment is a research method where there is manipulation of one variable under carefully controlled conditions so that resulting changes in another variable can be observed�key word �resulting.� Experiments are very powerful in that they allow for detection of cause-and effect relationships�Does X cause Y? The IV is the variable that the experimenter controls or manipulates�the DV is the variable thought to DEPEND (at least in part) on manipulation of the IV. If we wanted to know how X affects Y, X would be the IV, and Y would be the DV.

3. Experimental Designs: Variations Expose a single group to two different conditions Reduces extraneous variables Manipulate more than one independent variable - Allows for study of interactions between variables Use more than one dependent variable - Obtains a more complete picture of effect of the independent variable Experimental designs can be quite complex. These are a few of the ways designs can vary. Sometimes, a single group can be used for both experimental and control conditions�for example, you might study the effects of having the radio on when people work on an assembly line�you�d collect data from the same group of workers twice, once with the radio on and once with it off. Researchers can also manipulate more than one IV to see what the combined effect is�sometimes, the effect of one variable depends on the effect of another�for example, you might find that having the radio on increases productivity in workers, but only in the morning�in this example, time of day interacts with the effects of the radio. Researchers can also use more than one dependent variable in a single study to get a more complete picture of the effect of the independent variable. For example, we might measure not only number of pieces workers finish when the radio is allowed to be on while they work, but also worker satisfaction, absenteeism, and attitude. Having 1 day less a month absenteeism might make up for a slight decrease in productivity. Experimental designs can be quite complex. These are a few of the ways designs can vary. Sometimes, a single group can be used for both experimental and control conditions�for example, you might study the effects of having the radio on when people work on an assembly line�you�d collect data from the same group of workers twice, once with the radio on and once with it off. Researchers can also manipulate more than one IV to see what the combined effect is�sometimes, the effect of one variable depends on the effect of another�for example, you might find that having the radio on increases productivity in workers, but only in the morning�in this example, time of day interacts with the effects of the radio. Researchers can also use more than one dependent variable in a single study to get a more complete picture of the effect of the independent variable. For example, we might measure not only number of pieces workers finish when the radio is allowed to be on while they work, but also worker satisfaction, absenteeism, and attitude. Having 1 day less a month absenteeism might make up for a slight decrease in productivity.

4. The Scientific Method: Terminology Operational definitions are used to clarify precisely what is meant by each variable Participants or subjects are the organisms whose behavior is systematically observed in a study Data collection techniques allow for empirical observation and measurement Statistics are used to analyze data and decide whether hypotheses were supported Psychologists use operational definitions to clarify what their variables mean�what exactly is sociability? Researchers use procedures for making empirical observations and measurements, including direct observation, questionnaires, interviews, psychological tests, physiological recordings, and examination of archival records. They depend on statistics to analyze data and decide whether hypotheses were supported�observations are converted into numbers, which are then compared.Psychologists use operational definitions to clarify what their variables mean�what exactly is sociability? Researchers use procedures for making empirical observations and measurements, including direct observation, questionnaires, interviews, psychological tests, physiological recordings, and examination of archival records. They depend on statistics to analyze data and decide whether hypotheses were supported�observations are converted into numbers, which are then compared.

5. Strengths and Weaknessesof Experimental Research Strengths: conclusions about cause-and-effect can be drawn Weaknesses: artificial nature of experiments ethical and practical issues The power of the experimental method lies in the ability to draw conclusions about cause-and-effect relationships from an experiment. No other research method has this power. Experimental research does, however, have limitations. Experiments are often artificial; researchers have to come up with contrived settings so that they have control over the environment. Some experiments cannot be done because of ethical concerns�for example, you would never want to malnourish infants on purpose to see what the effects are on intelligence. Others cannot be done because of practical issues�there�s no way we can randomly assign families to live in urban vs. rural areas so we can determine the effects of city vs. country living. The power of the experimental method lies in the ability to draw conclusions about cause-and-effect relationships from an experiment. No other research method has this power. Experimental research does, however, have limitations. Experiments are often artificial; researchers have to come up with contrived settings so that they have control over the environment. Some experiments cannot be done because of ethical concerns�for example, you would never want to malnourish infants on purpose to see what the effects are on intelligence. Others cannot be done because of practical issues�there�s no way we can randomly assign families to live in urban vs. rural areas so we can determine the effects of city vs. country living.

6. Evaluating Research:Methodological Pitfalls Sampling bias Placebo effects Distortions in self-report data: Social desirability bias Response set Experimenter bias the double-blind solution Sampling bias � when a sample is not representative of the population�poll only men, may get a different outcome if the population is both male and female. Placebo effects � when a participant�s expectations lead them to experience some change even though they receive empty, fake, or ineffectual treatment�cured by a sugar pill. Distortions in self-report data: Social desirability bias � a tendency to give socially approved answers to questions about oneself�did you vote? Response set � a tendency to respond to questions in a particular way (agree with everything, etc.). Experimenter bias � when a researcher�s expectations or preferences about the outcome of a study influence the results obtained�researchers see what they want to see � errors are usually in favor of the hypothesis�similarly, researchers may unintentionally influence the behavior of their subjects, possibly through body language, smiles, etc. To control for this problem, a double-blind procedure in which neither subjects nor experimenters know which subjects are in the experimental and which are in the control groups is used�a non-directly involved researcher keeps track of everything. Sampling bias � when a sample is not representative of the population�poll only men, may get a different outcome if the population is both male and female. Placebo effects � when a participant�s expectations lead them to experience some change even though they receive empty, fake, or ineffectual treatment�cured by a sugar pill. Distortions in self-report data: Social desirability bias � a tendency to give socially approved answers to questions about oneself�did you vote? Response set � a tendency to respond to questions in a particular way (agree with everything, etc.). Experimenter bias � when a researcher�s expectations or preferences about the outcome of a study influence the results obtained�researchers see what they want to see � errors are usually in favor of the hypothesis�similarly, researchers may unintentionally influence the behavior of their subjects, possibly through body language, smiles, etc. To control for this problem, a double-blind procedure in which neither subjects nor experimenters know which subjects are in the experimental and which are in the control groups is used�a non-directly involved researcher keeps track of everything.

7. Statistics is the use of mathematics to organize, summarize and interpret numerical data.

8. Descriptive/Correlational Methods:Looking for Relationships Methods used when a researcher cannot manipulate the variables under study Naturalistic observation Case studies Surveys Allow researchers to describe patterns of behavior and discover links or associations between variables but cannot imply causation When practical or ethical issues do not allow for variables to be manipulated, researchers rely on descriptive/correlational methods. Naturalistic observation is when a researcher engages in careful observation of behavior without intervening directly with the subjects�do more men than women run yellow lights? A case study is an in-depth investigation of an individual subject�profile of a serial killer, etc. In a survey, researchers use questionnaires or interviews to obtain specific information about subjects� behavior�the Kinsey Report on �normal� sexual behavior. A modern day example: Cooper (1999) set out to determine how much time people spend on online sexual pursuits�conducted an online questionnaire that was posted for seven weeks that invited internet sex pursuers to participate. A self-selected sample such as this is not representative of the population of the U.S., but it probably was representative of those who visit sexually explicit websites. Descriptive/Correlational methods allow researchers to discover links or associations between variables, but cannot imply causation. When practical or ethical issues do not allow for variables to be manipulated, researchers rely on descriptive/correlational methods. Naturalistic observation is when a researcher engages in careful observation of behavior without intervening directly with the subjects�do more men than women run yellow lights? A case study is an in-depth investigation of an individual subject�profile of a serial killer, etc. In a survey, researchers use questionnaires or interviews to obtain specific information about subjects� behavior�the Kinsey Report on �normal� sexual behavior. A modern day example: Cooper (1999) set out to determine how much time people spend on online sexual pursuits�conducted an online questionnaire that was posted for seven weeks that invited internet sex pursuers to participate. A self-selected sample such as this is not representative of the population of the U.S., but it probably was representative of those who visit sexually explicit websites. Descriptive/Correlational methods allow researchers to discover links or associations between variables, but cannot imply causation.

9. Descriptive Statistics: Correlation When two variables are related to each other, they are correlated. Correlation = numerical index of degree of relationship Correlation expressed as a number between 0 and 1 Can be positive or negative Numbers closer to 1 (+ or -) indicate stronger relationship A correlation exists when two variables are related to each other. The correlation coefficient is a numerical index of the strength and direction of association between two variables. A correlation is expressed as a number between 1 and 0, and the number may be positive or negative. The closer to 1 the number is, whether +1 or �1, the stronger the relationship between the variables�for example, a correlation of .17 is pretty weak, while a correlation of -.89 is pretty strong. The positive/negative dimension of the correlation coefficient expresses the direction of the relationship. If two variables are positively correlated, they co-vary in the same direction�as scores on one variable go up, scores on the other variable go up too�if two variables are negatively correlated, the variables co-vary in the opposite direction�as one goes up, the other goes down. A correlation exists when two variables are related to each other. The correlation coefficient is a numerical index of the strength and direction of association between two variables. A correlation is expressed as a number between 1 and 0, and the number may be positive or negative. The closer to 1 the number is, whether +1 or �1, the stronger the relationship between the variables�for example, a correlation of .17 is pretty weak, while a correlation of -.89 is pretty strong. The positive/negative dimension of the correlation coefficient expresses the direction of the relationship. If two variables are positively correlated, they co-vary in the same direction�as scores on one variable go up, scores on the other variable go up too�if two variables are negatively correlated, the variables co-vary in the opposite direction�as one goes up, the other goes down.

10. Correlation:Prediction, Not Causation Higher correlation coefficients = increased ability to predict one variable based on the other SAT/ACT scores moderately correlated with first year college GPA 2 variables may be highly correlated, but not causally related Foot size and vocabulary positively correlated Do larger feet cause larger vocabularies? The third variable problem As a correlation increases in strength (closer to � or + 1), the ability to predict one variable based on knowledge of the other variable increases. SAT/ACT scores are correlated with first year college GPA at a moderate .40 to .50...this may not be perfect, but it allows admissions committees to predict with some accuracy how well a prospective student will do in college. Although correlation may allow prediction, it does not infer cause-and-effect. For example, a strong positive correlation has been shown between foot size in children and vocabulary�as foot size increases, so does vocabulary. Do bigger feet make children learn more words? No. It is a third variable, age, which causes both feet and vocabulary to grow. As a correlation increases in strength (closer to � or + 1), the ability to predict one variable based on knowledge of the other variable increases. SAT/ACT scores are correlated with first year college GPA at a moderate .40 to .50...this may not be perfect, but it allows admissions committees to predict with some accuracy how well a prospective student will do in college. Although correlation may allow prediction, it does not infer cause-and-effect. For example, a strong positive correlation has been shown between foot size in children and vocabulary�as foot size increases, so does vocabulary. Do bigger feet make children learn more words? No. It is a third variable, age, which causes both feet and vocabulary to grow.

11. Figure B.8: Scatter diagrams of positive and negative correlations. Scatter diagrams plot paired X and Y scores as single points. Score plots slanted in the opposite direction result from positive (top row) as opposed to negative (bottom row) correlations. Moving across both rows (to the right), you can see that progressively weaker correlations result in more and more scattered plots of data points.Figure B.8: Scatter diagrams of positive and negative correlations. Scatter diagrams plot paired X and Y scores as single points. Score plots slanted in the opposite direction result from positive (top row) as opposed to negative (bottom row) correlations. Moving across both rows (to the right), you can see that progressively weaker correlations result in more and more scattered plots of data points.

12. STATISTICS

13. Descriptive Statistics-used to organize and summarize data Inferential Statistics- used to interpret data and draw conclusions

14. Descriptive Statistics:Measures of Central Tendency Measures of central tendency = typical or average score in a distribution Mean: arithmetic average of scores Median: score falling in the exact center Mode: most frequently occurring score Which most accurately depicts the typical? Measures of central tendency are used to describe the typical or average score in a distribution. The mean is the arithmetic average and is therefore sensitive to extreme scores. The median is the score that falls exactly in the center of the distribution. The mode is the most frequently occurring score. Which one is the most accurate depiction of the typical score? It depends on the data, as depicted on the next slide. Measures of central tendency are used to describe the typical or average score in a distribution. The mean is the arithmetic average and is therefore sensitive to extreme scores. The median is the score that falls exactly in the center of the distribution. The mode is the most frequently occurring score. Which one is the most accurate depiction of the typical score? It depends on the data, as depicted on the next slide.

15. Graphing data Frequency distribution- an orderly arrangement of scores indicating the frequency of each score Histogram-a bar graph (uses data from a frequency distribution) Frequency Polygon- a line figure (uses data from a frequency distribution)

16. Figure B.1: Graphing data. (a) Our raw data are tallied into a frequency distribution. (b) The same data are portrayed in a bar graph called a histogram. (c) A frequency polygon is plotted over the histogram. (d) The resultant frequency polygon is shown by itself.Figure B.1: Graphing data. (a) Our raw data are tallied into a frequency distribution. (b) The same data are portrayed in a bar graph called a histogram. (c) A frequency polygon is plotted over the histogram. (d) The resultant frequency polygon is shown by itself.

17. Measures of Variability How much the scores tend to vary or depart from the mean (how much they vary). range= highest score minus lowest score standard deviation=amount of variability in a set of data (square root of the average squared difference scores from the mean) variance= SD squared

18. Normal distribution (bell curve) a symmetrical bell shaped curve that represents the pattern in which many human characteristics are dispersed in the population.

19. Figure B.7: The normal distribution and SAT scores. The normal distribution is the basis for the scoring system on many standardized tests. For example, on the SAT, the mean is set at 500 and the standard deviation at 100. Hence, an SAT score tells you how many standard deviations above or below the mean you scored. For example, a score of 700 means you scored 2 standard deviations above the mean.Figure B.7: The normal distribution and SAT scores. The normal distribution is the basis for the scoring system on many standardized tests. For example, on the SAT, the mean is set at 500 and the standard deviation at 100. Hence, an SAT score tells you how many standard deviations above or below the mean you scored. For example, a score of 700 means you scored 2 standard deviations above the mean.

20. Percentile Score (Rank)- the percentage of scores in a distribution that fall below a given score (or that score falls above the other scores)

21. Figure B.2: Measures of central tendency. Although the mean, median, and mode sometimes yield different results, they usually converge, as in the case of our TV viewing data.Figure B.2: Measures of central tendency. Although the mean, median, and mode sometimes yield different results, they usually converge, as in the case of our TV viewing data.

22. Figure B.5: Steps in calculating the standard deviation. (1) Add the scores (SX) and divide by the number of scores (N) to calculate the mean (which comes out to 3.0 in this case). (2) Calculate each score�s deviation from the mean by subtracting the mean from each score (the results are shown in the second column). (3) Square these deviations from the mean and total the results to obtain (Sd 2) as shown in the third column. (4) Insert the numbers for N and Sd 2 into the formula for the standard deviation and compute the results.Figure B.5: Steps in calculating the standard deviation. (1) Add the scores (SX) and divide by the number of scores (N) to calculate the mean (which comes out to 3.0 in this case). (2) Calculate each score�s deviation from the mean by subtracting the mean from each score (the results are shown in the second column). (3) Square these deviations from the mean and total the results to obtain (Sd 2) as shown in the third column. (4) Insert the numbers for N and Sd 2 into the formula for the standard deviation and compute the results.

23. Figure B.6: The normal distribution. Many characteristics are distributed in a pattern represented by this bell-shaped curve (each dot represents a case). The horizontal axis shows how far above or below the mean a score is (measured in plus or minus standard deviations). The vertical axis shows the number of cases obtaining each score. In a normal distribution, most cases fall near the center of the distribution, so that 68.26% of the cases fall within plus or minus 1 standard deviation of the mean. The number of cases gradually declines as one moves away from the mean in either direction, so that only 13.59% of the cases fall between 1 and 2 standard deviations above or below the mean, and even fewer cases (2.14%) fall between 2 and 3 standard deviations above or below the mean.Figure B.6: The normal distribution. Many characteristics are distributed in a pattern represented by this bell-shaped curve (each dot represents a case). The horizontal axis shows how far above or below the mean a score is (measured in plus or minus standard deviations). The vertical axis shows the number of cases obtaining each score. In a normal distribution, most cases fall near the center of the distribution, so that 68.26% of the cases fall within plus or minus 1 standard deviation of the mean. The number of cases gradually declines as one moves away from the mean in either direction, so that only 13.59% of the cases fall between 1 and 2 standard deviations above or below the mean, and even fewer cases (2.14%) fall between 2 and 3 standard deviations above or below the mean.

24. Statistical Significance Is said to exist when the probability that the observed findings are due to chance is very low, 5 in 100 P<.05

25. Certainty What level of certainty is most scientific research based on?

26. 95% (5 chances in 100 that the observed findings are due to chance)

27. Ethics in Psychological Research:Do the Ends Justify the Means? The question of deception The question of animal research Controversy among psychologists and the public Ethical standards for research: the American Psychological Association Ensures both human and animal subjects are treated with dignity The question of deception: Is it OK to make subjects think they are hurting others? Have homosexual tendencies? Think they are overhearing negative comments about themselves? The question of animal research: Controversy regarding humane treatment of animals vs. no use of animals in research. These and other ethical issues have led the American Psychological Association (APA) to develop a set of ethical standards for research, to ensure that both human and animal subjects are treated with dignity. The question of deception: Is it OK to make subjects think they are hurting others? Have homosexual tendencies? Think they are overhearing negative comments about themselves? The question of animal research: Controversy regarding humane treatment of animals vs. no use of animals in research. These and other ethical issues have led the American Psychological Association (APA) to develop a set of ethical standards for research, to ensure that both human and animal subjects are treated with dignity.

29. PERSPECTIVES IN PSYCHOLOGY How many perspectives or different approaches to explaining behavior are there? Do all psychologists agree? Do all people agree?

30. SIX 6 PERSPECTIVES: Psychoanalytic (Freud) Emphasizes unconscious impulses, conflicts: views behavior as clashing forces in personality: negative, pessimistic view of human nature Behavioral- Emphasizes the study of observable behavior and the effects of learning: external rewards & punishments: neutral view of human nature: mechanistic

31. Humanistic- Emphasizes subjective experience, human potentials, �self�: positive view, philosophical view of human nature Biological- explains behavior through the role of the brain, nervous system, and genes: neutral, reductionistic view

32. Cognitive- �thinking� view that behavior is influenced by conscious thinking: memory, decision making, judgment: behavior as information processing, neutral, computer-like view of human nature Eclectic- a combination of two or more perspectives to explain behavior