asliyalcin_PSY 341_CHAPTER 2.pptx
Document Details
Uploaded by BraveJubilation
Full Transcript
CHAPTER 2 Research Methods in Industrial/Organizational Psychology Inst. Aslı Yalçın Çankaya University Psychology Department Social Scientific Research Methods What qualities make a person an effective manager? Ask others? (get conflicting answers) Observe a good manager? K...
CHAPTER 2 Research Methods in Industrial/Organizational Psychology Inst. Aslı Yalçın Çankaya University Psychology Department Social Scientific Research Methods What qualities make a person an effective manager? Ask others? (get conflicting answers) Observe a good manager? Knowledge of tasks? Relationship with subordinates? Situation? How do you know who is a good manager? How will you determine the characteristics ? Scientific research methods do not rely on hunches or beliefs, but on the systematic collection & analysis of data Social Scientific Research Methods To determine the characteristics of a successful work group manager; Define the criteria Success (successful- well-liked? successful- group productivity?) Measurement Accurate and precise (distinction btw. unsuccessful & successful managers can be made) Isolate the specific characteristics related to criteria (K, A, Ss or personality that make a successful manager) Social Scientific Research Methods Social scientific research methodology a set of procedures that allows us to investigate the HOWs and WHYs of human behavior and to predict when certain behavior will or will not occur. Social scientific research methods enable an I/O psychologist to study a specific issue objectively. Objectivity is the unbiased approach to observation and interpretations of behavior. Social Scientific Research Methods The goals of science; to describe, explain, & predict phenomena. I/O psychology; Describe ( e.g. describing the production levels of a company, the rates of employee turnover) Explain (e.g. employee turnover rates high bec. of dissatisfaction with the payments) Predict work beh. (e.g. use scores from tests to predict which employee would be the best candidate) As an applied science, I/O psychology also has the goal of attempting to control or alter behavior to obtain desired outcomes. ( e.g. İmplement a program to increase job satisfaction) Social Scientific Research Methods 1. The first step in conducting research is to specify the problem or issue to be studied. ( rel. btw. worker job satisfaction & loyalty to organization) II. The second step is to take those elements the researcher intends to measure and develop hypotheses. Variables are the elements measured in research investigations. (e.g. Job satisfaction, loyalty to organization, turnover, tenure) Hypotheses are statements about the supposed relationships between or among Social Scientific Research Methods A theory organization of beliefs into a representation of the factors that affect behavior. III. The third step; selecting the research design. IV. depends upon several factors: research setting, the degree of control the researcher has, the research questions, and the variables. (e.g. observation, survey, experiment) The fourth step; data collection. Sampling is the selection of a representative group from a larger population for study. It is impossible to study all members of population! ( generalize the results obtained from sample to Social Scientific Research Methods Random sampling refers to the selection of research participants from a population so that each individual has an equal probability of being chosen. (table of random numbers or a computer program ) Stratified sampling is the selection of research participants based on categories representing important distinguishing characteristics of a population. Randomly select a specified number of people in a way that research sample mirrors the actual breakdown of these groups in the total population. Social Scientific Research Methods V. The fifth step in the research process is analysis of data. Once the data are collected, they can then be analyzed. In most cases, quantitative data are analyzed using statistical analysis. (also qualitative data) Statistics are used to describe data & to test hypotheses. Simple or complex statistics Only the relationships among the variables or analyses showing causality Social Scientific Research Methods VI. The final step of the research process is the interpretation of the results. The researcher draws conclusions about the meaning of the findings and their relevance to actual work behavior, as well as limitations of the current study and directions for future research investigations. Example: Effects on work productivity of two managerial styles (directive and non-directive/participative) in a manufacturing company. Conclusion: directive style lead more productive groups Limitation: May only apply to managers of factory Major Research Designs Experimental Design: research design characterized by a high degree of control over the research setting to allow for the determination of cause-and-effect relationships among variables. Experiments can be conducted in both labs. & fields. Independent variable (IV): the variable that is manipulated by the researcher Dependent variable (DV) is the variable that is acted upon by the independent variable (the outcome variable). No elements except the IV are allowed to vary. (Any other variable that can affect the DV other than IV should be kept constant or controlled) Statistically or by design Major Research Designs treatment group: the group that is subjected to the change in the independent variable control group: the group that receives no treatment or irrelevant treatment. extraneous (confounding) variables: variables other than the independent variable that may influence the dependent variable. Hold all extraneous variables constant! Ex: Hawthorne studies attention of researchers Random assignment: assigning subjects to groups to control the effects of extraneous variables. Easier in lab settings. Major Research Designs Disadvantages of experimental design: Artificiality A situation quite different from actual work setting Generalizability problem (advantage of field experiment) Experimental Design Example: Effectiveness of a training program for salespersons 100 employees randomly assigned to receive customer training & 100 employees randomly assigned to no-customer training condition. Following the training program; look at the change in customer spending IV: The training program DV: customer spending Experimental Group: received training Control Group: received no training Major Research Designs Quasi-experiment In many cases, a researcher does not have the control over the situation needed to run a true experiment. Quasi-experiments involve comparison of preexisting groups, where random assignment of participants to groups is not possible. Differences from a true experiment: Random assignment and/or Manipulation of independent variable Example: The job of a bank teller is redesigned in one branch of a bank but not in another branch. After three months, customer satisfaction found to be higher at the branch where the job redesigned took place. Major Research Designs The Correlational Design: examines the relationship between variables as they occur naturally. The researcher observes two variables & measures their statistical association with each other. No manipulation of variable! The main advantage: easier to implement in a particular setting, including the workplace. Disadvantage: cause-effect rel. can not be drawn Reverse causality problem e.g. organizational Satisfaction- purchasing organizational stock Third variable problem e.g. employees with tenure organizational satisfaction – purchasing organizational stock Major Research Designs 1 1. İlişki Odaklı Liderlik 2 3 4 5 6 7 1 2. İş Odaklı Liderlik .01 1 3. Kişisel Yakınlık .49** -.01 1 4. Lidere Yönelik Etkinlik Algısı .58** .14 .35* 1 5. Lidere Duyulan Güven .42** .20 .43** .61** 1 6. İş Doyumu .49** .23 .36** .59** .54** 1 7. İş Performansı .01 .36* .12 .21 .18 .42** 1 Major Research Designs Meta-analysis Different research investigations of the same topic may reach inconsistent (or contradictory) results. So, what conclusion? Meta-analysis; technique that allows results from several different research studies to be combined &analyzed & summarized (at least 20 studies). Statistical procedure that combines the results of many independent research findings on a single topic used to identify moderating variables estimate true relationship Measures effect size of findings Uses archival data Major Research Designs Meta-analysis Meta-analyses rely on indicators of effect size & the number of participants in each of the independent studies. Effect size: an estimate of the magnitude of the relationship btw. X & Y (in corr. design ) or effect of IV on DV (in experimental design) found in a research investigation. Example: Job Satisfaction – Absenteeism Example: One meta-analysis confirmed that more physically demanding jobs were related to workers becoming stressed & burnout. Major Research Designs Case study Depth view of past events using observations, interviews & archival records Qualitative descriptions of behavior. may provide rich, descriptive info. about work beh.s and settings. Ex: In one case study; a psychologist found that company picnics & other social activities increased employees’ loyalty to organization. Benefits: Detailed account of why particular event occurred Disadvantages: Little generalizability (anectodal)! No hypothesis testing, no cause-effect rel.s Measurement of Variables Research variables are operationalized, or clearly defined Abstract Concrete Operational definition: a working definition of what the variable is & how a variable will be manipulated/ measured in a study. Ex: safe driving beh: wearing a seat belt, using a turn signal & coming to a full stop at an intersection Two basic methods of measurement in I&O psychology Observation Self-Report Measurement of Variables Observation: the researcher directly & systematically observes certain beh.s Ex: supervisory beh.s- demonstrating work techniques to subordinates, giving direct orders. Obtrusive observation: research observation in which the presence of the observer is known to the participants. Disadvantage: the part. may behave diff.ly ( Hawthorne effect) Unobtrusive observation: observation in which the presence of the observer is not known to the participants. Disadvantage: ethical concerns about protecting the privacy of the part.s Measurement of Variables Self-Report Techniques: measurement methods that rely on research participants’ reports of their own behaviors or attitudes. Interview: How do feel right now?. Survey: How often do you feel angry? Disadvantage: distortion of the responses ( workers may give socially desirable answers) Interpreting and Using Research Results Two key issues concerning measurement of variables: Reliability: consistency of the measure Test-retest reliability: people who take the test more than one time should get the same scores. Ex: the job applicant should get similar scores on the same tests taken at two different points of time. Inter-rater reliability: the extent to which different people agree on the characteristics they are measuring. Ex: 1. I/O psychologist- worker’s beh.s job satisfaction 2. I/O psychologist- worker’s beh.s loyaty to organization Interpreting and Using Research Results Two key issues concerning measurement of variables: Validity: extent to which a measure assesses what it claims to measure. Ex: a job performance test should really test an employee’s job performance Internal validity: the extent to which extraneous/ confounding variables are removed. the extent to which cause&effect conclusions can be drawn. External validity: the extent to which the results can be generalized to the different settings, different people. Reliability is necessary but not sufficient for Ethical Issues in Research and Practice in I/O Psychology The American Psychology Association (APA) lists several core principles that should guide the ethical conduct of researchers in psychology, including I/O psychology. Do no harm & strive for the benefit of participants Be honest & accurate in science, teaching and practice of psychology Respect the rights of people to privacy and confidentiality. Informed Consent: a research participant is fully informed of the nature, purpose, duration and procedures of the study and ensure that the part. has the right to not participate. (confıdentiality & anonymity of the responses) Debriefing: part.s should be fully debriefed & ensure that no harm has been caused. Statistical Analyses of Research Data The science & the practice of I/O psychology require a throughknowledge of research methods and statistics. Two types of statistics: Descriptive Statistics Inferential Statistics Descriptive Statistics: numerical characteristics of the nature of the data set. Arithmetical formulas for summarizing & describing research data What is the avarage tenure in your sample? How many female employees? How many male Statistical Analyses of Research Data o Three major types of descriptive statistics; 1. Frequency Distribution: A descriptive statistical technique that presents data in a useful format, arranging the scores by category In the form of bar graph or histogram Statistical Analyses of Research Data o Three major types of descriptive statistics; II. Central Tendency: where the group tends to cluster Mean, Median, Mode EXAMPLE EXAMPLE Data: Data:1,1,1,1,1,2,2,3,3,3,3,5,5 1,1,1,1,1,2,2,3,3,3,3,5,5 Mean Mean ==2.38; 2.38;Median Median==2;2;Mode Mode==11 Mean Arithmetic average Median The score at the midpoint of a statistical distribution Mode The most frequently obtained score in the distribution of data Statistical Analyses of Research Data o Three major types of descriptive statistics; III. Variability: How scores varies among the part.s Range - distance between highest and lowest score (Range = High score – Low score) Range = 17 – 5 = 12 (less susceptible to extreme scores!) Standard Deviation – average distance from the mean S= Σ(x – xm)2 / n – 1 (sample) S= (5-9) + (6-9) + (6-9) + (8-9) 2 + (9-9) 2 + (10-9) 2 + (11-9) 2 + (17-9) 2 / 7 S = 3.85 2 2 2 Statistics: The Language of Research mean ( Range vs. Standard Deviation Sample Distributions Normal Distribution Bell-shaped with most scores falling toward the middle with few at high and low extremes Mean, median and mode are equal Skewed Distribution Asymmetrical with more scores falling closer to high or low extremes Median is most useful measure of central tendency because it is less affected by extreme scores The Normal Distribution of IQ Inferential Statistics Methods for analyzing data that express relationships (e.g., differences between groups) in terms of probabilities E.g. Statistical significance The level of confidence we have in the results of an experiment based on probability values Probability The idea that a difference between the means of experimental and control groups could have occurred by chance If P=.01, a difference would occur by chance only 1 time out of 100. A safety training program reduced the accident rates? Control group no training Statistical Analyses of Experimental Method Data T-test: A statistical test for comparing the means of two groups whether a process or treatment has an effect ? whether two groups are different from one another? IVCat. DV Con. Analysis of Variance (ANOVA): Differences among more than two groups on a single variable IVCat. DV Con. Multivariate Analysis of Variance (MANOVA): Examines data from multiple groups with multiple DVs. whether there are any differences between groups on more than one dependent variable? IVCat. DV Con. Statistical Analyses of Experimental Method Data Statistical Analysis of Correlational Methods Correlation Coefficient Multiple Regression Multiple Regression Effects of one variable over another (used for prediction) Yi = ß0 + ß1Xi1 + ß2Xi2 (Y = a + bX) Regression Variables - Predictor Variable (X) – measure used to predict an outcome (similar to independent variable) E.g. selection test scores, years of experience, education level Statistical Analysis of Correlational Methods Multiple Regression Statistical Analyses of Experimental Method Data Multiple Regression Main effects Interaction effects When the effect of one IV on the dependent variable differs depending on the level of the second independent variable. E.g: Leadership context (military, business, politics) X Perceived leadership style (ROL, TOL) effectiveness (Goncu, Kastendieck, & Johnson, 2009)
