Introductory Psychology PSYC 1000
Popular in Course
Popular in Psychlogy
verified elite notetaker
This 9 page Class Notes was uploaded by Lane Schuster on Sunday October 11, 2015. The Class Notes belongs to PSYC 1000 at East Carolina University taught by Karl Wuensch in Fall. Since its upload, it has received 25 views. For similar materials see /class/221340/psyc-1000-east-carolina-university in Psychlogy at East Carolina University.
Reviews for Introductory Psychology
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 10/11/15
Research Methods Employed In Psychology Psychologists use scientific research methods to investigate events of interest to them Sometimes these research methods involve only a single variable for example a psychologist may be interested in determining how many people dream in color rather than in black and white She could survey a random sample of people asking them questions that should reveal whether they dream in color or not and from the sample data estimate what percentage of people in the general population do dream in color and what percentage do not More interesting than singlevariable research is research that attempts to determine how variables are related to one another For example our psychologist having determined that some humans dream in color and others do not might now try to discover what other variables are related to dreaming in color For example are women more or less likely to dream in colorthan are men ls age related to dreaming in color Are colorful dreams more likely early in one s daily cycle of dreaming or later in that cycle No I don t have answers to the questions but we could answerthem with scientific research Sometimes we conduct research on two or more variables without having any strong feelings about how the variables are related to one another We just want to determine whetherthe variables are related and if so how they are related I call such research exploratory research Such exploratory research produces the ore from which we may be able to re ne a theory Scientific Theories A scientific theory is a mental often mathematical model of how basic constructs are related to one another and to measurable variables Theories are constructed by a mental process known as induction inferring the general from the speci c in which we take a fairly large number of observations and from them produce general principles that seem to explain the known observations or facts Our hope is that such a theory can be used to predict other events by the mental process of deduction inferring the specific from the general Such a prediction is sometimes referred to as an hypothesis The ideal scientist tests her theories by trying to falsify them That is she deduces hypotheses from her theory and then conducts research to determine whether or not the hypotheses appear to be correct lfthe data support the theory then her con dence in the theory is increased If not then she must either abandon the theory or attempt to revise it to account for the new data Operationalization Theories predict relationships between constructs these predictions are known as hypotheses Constructs are hypothetical entities such as political conservatism and happiness To test predicted relationships we must operationallze the constructs that is decide how to measure or manipulate them I might operationallze political conservatism by asking who you voted for in the last presidential election or with a survey having several questions about issues on which Gray2doc Page 2 conservatives typically differ from liberals I might operationalize happiness by having an observer count the number of times you smile during some period oftime or I might ask you to answer a series of questions about your emotional life or I might measure electrical activity in a brain area associated with feeling happy Once I have operationalized the constructs I can measure them and then see ifthey are correlated I can also operationalize how to manipulate change a construct and then do experimental research How might I manipulate your happiness It is very important that you distinguish between the meaning of theory as used in science and theory as used in everyday language In everyday language theory means a hunch a guess typically one with little empirical support In science a theory typically is much more than a crude guess it is a mental model that adequately explains aH ofthe known facts and which may or may not adequately explain facts that have yet to be discovered Since it is always possible that new facts may be discovered that cannot be explained by a theory scientists never claim to have created a theory that is beyond doubt in fact ifa theory could not conceivably be shown to be false by some new data then that theory is considered not to be scienti c For example your friend has a theory that God has a gray beard Since this theory cannot be tested and possibly refuted by any scienti c research in this lifetime it is not a scienti c theory Please do not confuse scienti c theories with such common hunches Scienti c theories excepting those that have been refuted are very likely to be close to the truth Clever Hans Dr Gray used the story of Clever Hans to illustrate the advantages of the scientific method as contrasted with the process used by nonscientists to evaluate hypotheses The hypothesis tested was that a horse if properly trained by the same methods used to teach German children can learn both spoken and written human languages math history and so on the same as humans Amazingly the performances of this horse and his owner convinced the general public and a number of putative scientists that horses are as intelligent as humans It took the scientific methodology of psychologist Oskar Pfungst to show that Clever Hans was just very good at detecting subtle hints from humans that told him which way to shake his head or when to stop tapping his hoof Lest you think that we are any less gullible now than in the early 1900 s when Clever Hans was a star let me tell you about a television show that appeared in the US late in the 1900 s A horse was brought on stage and asked assorted questions It answered by tapping its hoof and it always got the correct answer The audience went wild In a repeat performance the trainer demonstrated that this horse had ESP it could read minds The trainer asked one question like what is 3 times 2 but thought another question what is 3 plus 2 When the horse tapped 5 times the gullible audience was apparently convinced that this horse had ESP Experimental versus Nonexperimental Research In experimental research one manipulates an independent variable and then observes whether or not this causes changes in a dependent variable lfthe researcher has properly controlled extraneous variables which if not controlled could Page 3 contaminate the research then the researcher should be able to tell whether or not there is a causal relationship between the independent variable and the dependent variable Experimental research can be withinsubjects or betweensubjects ln withinsubjects research the same subjects serve in both experimental conditions For example I want to determine if drinking pint of my homemade beer affects my subjects reaction times I test their reaction times twice once after drinking a nonalcoholic beverage and once after drinking my beer When the results show that their reaction times are longer after the beer than after the other beverage I conclude that drinking beer causes longer reaction times If I wanted to conduct this research by the betweensubjects method I would start by randomly assigning each subject to serve in the beer group or the nonalcoholic group The appropriate treatments would be applied to each group and reaction times measured In nonexperimental research one observes whether changes in one variable are accompanied by changes in another variable but no variable is actively manipulated by the researcher For example I might simply ask my subjects whether or not they had been drinking priorto coming to have their reaction time tested If my results demonstrated that those who had been drinking had longer reaction times than those who had not been drinking then I can conclude that people who drink have longer reaction times than do those who do not drink Please notice that is not the same as concluding that drinking alcohol causes longer reaction times The correlation I observed could be explained in a number of different ways it could be that drinking does cause longer reaction times or it could be that there is some third variable causing changes in both reaction times and alcohol consumption producing a correlation between reaction time and alcohol consumption in the absence ofany direct relationship between reaction time and alcohol consumption For example there may be some brain disorder which causes affected people to have slow reaction times whether or not they drink alcohol and which also causes them to be likely to drink alcohol when it is available Let me give you another example of a third variable problem Observational research has demonstrated an association between smoking tobacco and developing a variety of health problems One might argue that this association is due to a third variable ratherthan any causal relationship between smoking and ill health Suppose that there is a constellation of third variables think ofthem as genetic or personality variables that cause some people to smoke and whether or not they smoke also cause them to develop health problems These two effects ofthe third variable could cause the observed associated between smoking and ill health in the absence ofany direct causal relationship between smoking and ill health geneticpersonality v ariables health problems Page 4 How can one rule out such an explanation It is not feasible to conduct the required experimental research on humans randomly assigning newborns to be raised as smokers or nonsmokers but suchresearch has been done on rats Rats exposed to tobacco smoke develop the same sort of health problems that are associated with smoking in humans 80 the tobacco institute has promised not to market tobacco products to rats By the way there has been reported an interesting problem with the rats used in such research When con ned to a chamber into which tobacco smoke is pumped some ofthem take their fecal boluses and stuffthem in the vent from which the smoke is coming Another example of a third variable problem concerns the air traf c controllers strike that took place when Reagan was president The controllers contended that the high stress of working in an air traffic control tower caused a variety of health problems known to be associated with stress It is true that those working in that profession had higher incidences of such problems than did those in most other professions The strikers wanted improved health bene ts and working conditions to help with these stress related problems But the government alleged that it was not the job that caused the health problems it was a constellation ofthird variables personalitygenetic that on the one hand caused persons ofa certain disposition Type A folks perfectionists to be attracted to the air traf c controllers profession and that same constellation ofthird variables caused persons with such a disposition to have these health problems whether or not they worked in an air traf c control tower One FAA of cial went so far as to say that working in an air traf c control tower is no more stressful than driving the beltway around DC Personally I nd such driving very stressful The government won the union was busted I suppose they could solve the problem by hiring as air traf c controllers only folks with a different disposition Type B lay back what me worry so what if those two little blips on the screen on headed towards one another Dr Gray page 3536 read these pages carefully gives another good example of a the dif culties encountered when attempting to interpret the results of nonexperimental research Diana Baumrind s research demonstrated that children whose parents have certain parental styles are more cooperative and friendly than children whose parents employ other parental styles Findings such as these have often been interpreted as showing that parental behavior causes offspring behavior but it could just as well be the other way around or there could be a third variable problem Please note that l have used the term nonexperimental ratherthan correlational to label research where no independent variable is manipulated I do this to prevent confusion between nonexperimental research methods and the statistical technique of correlation Demonstrating a statistical correlation between two variables that is demonstrating that when the one changes so does the other does establish that they are causally related ifthe data were obtained by experimental means and with adequate control of extraneous variable but when the data were obtained by Pages between the yariables possibly one inyolying a third yariable Statistics When discussing the results ot psychological research this Semesten l shall iiii u statistics we shall take a tew minutes now tor a quick lesson in research statistics Frequency Distributions A table or tigure that shows how trequent each score is in a group o scores is reterred to as a trequency distribution the hypothetical beer drinking ex erirnerit we discussed earlier Suppose that thetwo cuiyes shown below depict thetreouency distributions tor the groupthat drankthe nonalcoholic beverage the curve on the let andthe groupthat drankthe beer the curve on the right The vertical dimension p that the i a ii iiiiiii ior iiiaiithosetor the people who did not drink beer Measures of Location also called Central Tendency From the tigure aboye l curve tor the beer drinking group is located to the right ot that tor the other group that is min u iuiiiiking gr that quite clearly on generally will also want to report statistics that summarize the ditterences in thetw groups locations One such statistic is the arithmetic mean which is simply the sum ot the scores diyided by the number ot scores Alternatiyely 39 Whlch p in iihaltott e 8 u e 0 scores tall ui textbook is badly skewed lopsidedt the mean and median can ditter considerably in yalue in that case it is a good idea to report both the mean andthe median hen the distribution is nearly symmetric like those in thetigures aboye the mean andthe median will be nearly identical in that case it isthe mean which is usually reported Measures of Dispersion also called Variability When describing a distribution ot scores one should also indicate how he scores ditter rom one another Scores may all be yery A Scam Pquot m close to one anothen or they may be dispersed sprea out across a wide range Consider the two treouenc distributions graphed at the right The two distribution haye essentially the same mean andthe same shape butt e scores tor the pink one are spread out more than those tor the blue one Once agalm whilethe tigure says it all it is customary to present summary statistics One conyenient E Beverages 2 a o Burgers Page 6 measure is the variance which is the average squared amount by which the scores differ from the mean see the statistical appendix for an illustration of how to compute this statistic Because it is awkward to have this measure in squared units one usually takes the square root ofthe variance to obtain the standard deviation Correlation Coefficients A correlation coef cient can be employed to measure the strength ofthe association between two variables For example suppose that we have data on the number of burgers and number of beverages consumed by each person at a picnic The data are illustrated in the scatterplot here The plot clearly shows that beverage consumption is positively correlated with burger consumption as the one variable increases so does the other When the relationship is well described by a straight line as is the case here one can compute the Pearson product moment correlation coefficient r for short Computation of such a coef cient is illustrated in the statistical appendix of your textbook For these data r 8 Pearson r can range from 1 a perfect negative correlation to 0 no correlation to 1 a perfect positive correlation A perfect Pearson r correlation is one where all of the points fall perfectly on a straight line In that case if you knew a subject s score on the one variable you could predict her score on the second variable without error A positive r simply means that as scores on the one variable increase so do scores on the other variable A negative r simply means that as scores on the one variable increase scores on the other variable decrease Pearson r can also be used to measure how much two groups of scores differ from one another For our hypothetical experiment on beer consumption and reaction time we could give every subject in the beer group a score OH on the Beer variable and every subject in the nonalcoholic group a score ofO on the Beer variable We would then compute Pearson r between the Beer scores 0 and 1 and the Reaction Time scores When that r came out to be positive we would conclude that drinking beer causes increased reaction time The correlation between two variables may not be cum39inea Re39atimsmp described well by a straight line In that case one should compute a correlation coef cient that is appropriate for a 6 curvilinear relationship To the right is a scatter plot showing such a curvilinear relationship nding a relationship of this shape relationship with very dif cult tasks but not with simple tasks is referred to as the YerkesDodson law discussed on pages 220 and 221 of your textbook Performance lnferential Statistics So far we have discussed 0 1 2 3 4 only descriptive statistics statistics that describe the scores that we have on hand If we can consider those scores to be random sample ofthe population of scores in which we are actually interested then we can employ inferential statistics to estimate the values of means standard deviations correlation coefficients and so forth that would be obtained if we had the entire population of scores available lnferential Test Anxiety Page 7 statistics is the name given to the branch of statistics that is involved in making such estimates Statistical Significance When you have computed a correlation coef cient using data from a sample of scores it is almost certain that the correlation coef cient will not be exactly 0 even if there were absolutely no correlation between the variables in the population Accordingly before we use sample statistics to infer that two variables are actually correlated in the population we typically subject them to a test of statistical signi cance Such a test produces a value ofa statistic called the level of significance p for short The p value is a probability the probability of getting a correlation coefficient as large as the one we got if in fact the two variables were absolutely uncorrelated in the population from which our sample of scores was randomly drawn Think of the p value as measuring how well the data agree with the hypothesis that there really is no correlation between the variables lfthat p value is very low then we become willing to reject the hypothesis of no correlation and assert that there is probably a correlation between the variables in the population as well as in our sample By convention some would say a very poor convention the p value has to be less than or equal to 05 5 percent before we conclude that there probably is a correlation between the two variables in the population Such a result is described as a statistically significant result Being statistically signi cant does not mean that the result is ofany practical importance in fact a very tiny correlation will be found to be statistically signi cant ifone has enough data in the sample Likewise a result that is not statistically signi cant is not necessarily an result of no importance sometimes it is very important to demonstrate that there is little if any correlation between two variables One interpretation of a nonsignificant correlation coefficient is this Given our sample data we cannot dismiss the possibility that the variables are absolutely uncorrelated in the population or close to uncorrelated and ifa correlation does exist we cannot even predict with con dence in which direction the variables are correlated positively or negatively Dr Gray Goofed Dr Gray does not make many errors but on page 41 he made one that statisticians consider serious A p value is what statisticians call a conditional probability the probability of one event A given that another event B is assumed to have occurred Dr Gray has reversed the target event A with the given event B He said there that the p value is the probability that the variables are not correlated given the observed data That is incorrect To get such a probability we would have to use Bayesian statistics and there are problems with such statistics that prevent most psychologists from using them The p value is the probability of getting a sample correlation as large as the one obtained given that the variables are not correlated in the population I ll let Dr Gray know about this error and the fact that he should take comfort in knowing that most researchers and even many authors of textbooks in psychological statistics make exactly the same error Issues in Research Methodology There are certain pitfalls that should be avoided if possible when conducting psychological research I shall brie y discuss those mentioned by Dr Gray Page 8 Confounds Gray calls this Bias A confound is said to have occurred when an extraneous variable has become entangled with the independent variable in such a way that it is impossible to separate the effect ofthe independent variable from that of the extraneous variable For example a professor wants to compare the effectiveness of two different teaching methods He uses Method A with one class Method B with another and all get a common nal examination Statistical analysis shows that the mean score on the nal was signi cantly greater with Method B than with Method A So far so good but then we learn that the class where Method A was used met at 8 in the morning and the class where Method B was used met at 10 in the morning We have a confound now we don t know whether it was the method of instruction or the time of the meetings that caused the differences between the two classes Sometimes confounds are caused by our two groups of subjects already being different on the dependent variable before we manipulate the independent variable To avoid this problem we randomly assign subjects to groups when we can External Validity lfthe subjects employed in the research are not representative ofthe population of interest a biased sample then the results found with that sample may not generalize well to the population of interest Such research is said to have poor external validity Dr Gray gives a classic example ofthis problem on page 43 of your textbook the Literary Digest poll of voters in 1936 External validity may also be a problem when research is conducted in the laboratory rather than in the field Conditions is the laboratory might be so different from those in the eld that the results of the laboratory research do not generalize well to the field Reliability and Validity A reliable instrument is one which produces the same scores when used repeatedly to measure the same unchanged objects or events When research data are collected with an instrument of low reliability random error is introduced into the scores While this will not create a confound it can make it much harder to detect the effect ofthe independent variable A valid instrument is one which measures what it says it measures There are a number ofways to estimate the validity of an instrument Gray mentions face validity the instrument looks like it is measuring what it says it measures and criterionrelated validity scores from the instrument correlate well with another measure ofthe target attribute Expectancy Effects When the data collector knows which experimental group a subject is in she may unwittingly bias the results Dr Gray discusses an interesting example of such an observerexpectancy problem facilitated communication in autistic persons see pages 44 45 in your textbook To avoid such problems the data collector may be kept ignorant with respect to which group each subject is in Likewise ifthe subject knows which group he is in his behavior may be affected by that expectation a subjectexpectancy problem Accordingly the researcher may take steps to prevent the subjects from knowing what group they are in For example in drug trials subjects who are in the group that does not get the drug treatment may be given a placebo drug to take a substance that has no effect on the dependent variable This is intended to make it harder for the subject to figure out whether he is getting the experimental drug or not When both the data collector and the subjects are Page 9 kept in the dark with respect to who got what treatments the experiment is said to be a doubleblind experiment Research Ethics In recent years psychologists have struggled a lot with the questions regarding what is and what is not ethical to do in psychological research In fact many of the classic studies you will read about in your textbook studies done many years ago would be considered unethical by today s standards Here is a quick summary ofthe type of issues that have received the most attention Issues of Privacy most people do not want the results of psychological tests conducted on them to be revealed to others Deception most people get upset if they nd out that the researcher was not truthful with them with respect to things like the purpose ofthe research Discomfort or Harm if there is going to be any likely discomfort or harm caused by participating in an experiment most people would like to know about that before deciding to participate in the experiment These days most psychological research must be approved by an IRB Institutional Review Board prior to being conducted The IRB is charged with assuring that the research is conducted in an ethical fashion The typical steps taken to assure that there are no serious ethical problem include o Gathering and recording the data in ways that preserve subjects anonymity o Obtaining informed consent from the research participants and o lfthe research does have risks taking reasonable steps to reduce the amount of harm done Informed consent typically involves advising the participants regarding what will take place during the research any possible harm or discomfort how their con dentiality will be maintained and that they can quit the experiment at any time One interesting question here is How does the informed consent procedure affect the behavior of the research participants There is the possibility that the informed consent procedure could bias the results of the experiment Another interesting issue if that of conducting research on nonhuman animals How can you get informed consent from a laboratory rat for example Some would argue that it is never appropriate to conduct research on a participant that cannot grant consent such as a nonhuman animal while others would argue that it is appropriate if it can be shown that the bene ts of the research justify any harm or discomfort done to the nonhuman subjects Revised January 2009 Back to the PSYC 1000 Lecture Notes Paqe
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'