Techniques of Experimental Designs PDF

TECHNIQUES OF EXPERIMENTAL DESIGNS Tooba Arshad Institute of Professional Psychology, Bahria University- Karachi Campus PSYCHOLOGICAL EXPERIMENTS  Experiments in Psychology- to generate reliable reports about psychological life of people and animals- development of theories and models  Psycho...

TECHNIQUES OF EXPERIMENTAL DESIGNS Tooba Arshad Institute of Professional Psychology, Bahria University- Karachi Campus PSYCHOLOGICAL EXPERIMENTS  Experiments in Psychology- to generate reliable reports about psychological life of people and animals- development of theories and models  Psychological experiments- (mostly) study cause- and-effect relationship  Psychological Phenomenon- multiple cause and multiple effects  Development of sophisticated experimental designs and analysis EXPERIMENTAL DESIGNS  The design of controlled conditions under which one might make empirical observations within the context of testable hypothesis, that leads towards anticipated r/s b/w the conditions of an experiment and its outcome TECHNIQUES OF EXPERIMENTAL DESIGNS  Factorial Design  Complex Factorial Designs  Quasi-Experimental Designs (ex post facto research)  Functional Designs  Small n Designs (AAA and ABA Designs) SINGLE-FACTOR DESIGNS  Effect of only one IV on one DV  Only one factor is manipulated-rest of the factors are controlled or held constant  Example: Effect of Audio-Video (AV) Aids on the learning of students  Rare situations  Example???? Factorial Designs IVS  Reinforcement = Positive and Negative  Experience = Intermediate & Expert  DV= Performance 2(IV1)x2(IV2) 2x3x3 EFFECT OF REINFORCEMENT AND EXPERIENCE ON THE PERFORMANCE  Reinforcement (Positive & Negative)Levels  Experience (Intermediate & Master, mod, mod)  2 x 4=8 Factorial Design  Positive & Intermediate =5  Negative & Intermediate =0 Conditions  Positive & Master =10  Negative & Master =2 Effect of Gender and academic Background on the Learning of students 2x3 factorial Effect of Grades, Gender and Academic Year on the Learning of Students 6x2x4=36 factorial Effect of shape (round & square), size (Big & small) and color (Red, Blue & Black) of the cake on the perception of the taste 2x2x3=12 FACTORIAL DESIGNS  Psychological Phenomenon are not that simple that could be measured by only one cause  Experimenters not only try to find out the cause but they also try to find out the combination of cause that would lead towards a behavior  To determine the effects of two more than two IVs on a DV factorial designs are used  May be manipulated at the same time FACTORIAL DESIGNS  IV= Degree of Parenting (Kindness or Hostility) and Degree of Control (High & Low)  DV= Behavior of the children  No. of IVs 1x1  Level of IVs 2x3x2  Ehrenfreund and Badia (1962)- examined the performance of rats under varying food- deprivation and incentive conditions  Apparatus: 5 foot long straight alley-start box and goal box  DV= the running speed of rat  20 rats 2X2 FACTORIAL DESIGN 1. High deprivation- High Incentive 2. High deprivation- Low Incentive 3. Low deprivation- High Incentive 4. Low deprivation- Low Incentive  5 rats in each condition  Performance = running speed of rat traversed in the middle 2feet of alley  Median of running speed/ 10 each  Higher Score= faster running time  Main Effect When you see the effect of an IV on DV and ignore the other IV  Interaction Effect When you see the effect of one IV on the DV in the presence of Other IV RESULTS  Deprivation treatment Mean of High incentive > Mean of Low incentive  Incentive treatment Mean of High deprivation group > Mean of Low deprivation group EFFECT OF SHAPE (ROUND & SQUARE) AND SIZE (BIG & SMALL) OF THE CAKE ON THE PERCEPTION OF THE TASTE EXPERIMENT ON DISSONANCE THEORY  Dissonance Theory  Level of Dissonance depends on the amount of money  High Dissonance= ?  Low Dissonance= ? LINDER, COOPER AND JONES (1976)  Reinforcement theory is a applicable only in no choice condition  Dissonance can be created when people have choice; not when they are forced  Experiment: students have to write an essay on speaker-ban law  Factorial Design????? EXPERIMENTAL DESIGN  Independent variables  Choice (No Choice- Free Choice)  Incentive (50 cents- $2.50)  Dependent Variable= the amount of attitude change STATISTICAL SIGNIFICANCE  Level of Significance: as the statistical point above (or below) which one can infer the operation of no chance factor  0.05 (1 in 20)  0.01(1 in 100)  Statistically Significant: When data meet the pre-established criteria  Chance of variation MAIN EFFECT  Effect on one IV on DV ignoring the other IV INTERACTION EFFECT  One IV is effecting the DV in the presence of other IV  Interaction= One IV effects the DV differently in both levels of other IV  https://www.youtube.com/watch?v=6o7wyyh8Lu8  https://www.youtube.com/watch?v=9hnkvr3oyQM MORE (1969)  Compare the effects of immediate and delayed information feedback using learning material  8th Grade students from 4 different schools  IV1- Feedback: length of delay b/w taking the test and learning of the results  Intervals: immediate, 2.5 hr, 1 day, 4 days  After receiving the feedback they took the same test again  IV2 – the students either took the test immediately after receiving feedback (acquisition treatment) or 3 days after receiving feedback (retention treatment).  DV: no. of students answered correctly the second time  High Performance= Acquisition group> Retention Group  Retention  No support for the assumption that immediate feedback maximizes learning  Some delay in feedback is better Complex Factorial Designs SOLSO AND SHORT (1979)  Measurement of mental processes-reaction time  Intellectual processes such as matching two signals take some time, albeit a very small amount of time  T-scope (tachistoscope):  an instrument used for exposing objects to the eye for a brief measured period of time.  Used to test the limits of human visual perception  Psychologists have been interested in the way information is coded in the memory after it is perceived  Previously, it was found that participants initially formed visual codes for letters and then formed name codes. 3X3X2 WITH IN SUBJECT DESIGN  Solso and Short performed experiment on color codes  IV (1)= type of relationship (color to associate, color to word, color to color)  IV (2)= delay b/w the stimuli (0, 500, 1500 msec)  IV (3)= match (red, blood) or miss-match (red- blue) 3X3X2 WITH IN SUBJECT DESIGN  Experiment on color codes  IV (1)= type of relationship (color to associate, color to word, color to color)  IV (2)= delay b/w the stimuli (0, 500, 1500 msec)  IV (3)= match (red, blood) or miss-match (red- blue)  DV= reaction time  With-in subject design: each subject was treated wit each of the experimental variables. RESULTS  Equal number of non-matched secondary stimuli  Fastest response= color-color pairings (red-red)  Slowest response= color- association pairings (red-blood)  Increase in interval (1500msec) decrease in Reaction time  Parallel development of codes to color PRACTICE  Design an experiment while using factorial design technique and indicate following:  IVs and DV  Levels of IVs  Subject Participation (With-in Group/ Between Group)  Main Effect  Interaction effect ADVANTAGES AND DISADVANTAGES OF FACTORIAL DESIGNS Click to edit Master text Click to edit Master text Advantages Disadvantages styles styles  Investigate multiple  Factorial designs take factors or variables much more work simultaneously  Factorial designs can  Multiple factors can produce interactions combine in different which are difficult or ways impossible to interpret QUASI-EXPERIMENTAL DESIGNS  Introduced by Campbell and Stanley (1966)  Studies which selects IVs from the natural environment  Overcome the problems faced by psychologists who wish to study behaviors occurring in less formally structured environments than labs  sometimes termed as “ex po facto research” or “as if” designs  The collection and the analysis of the data take place after an event has happened  Similar to naturalistic observation; difference introduction of IV  Responsible for experimental control  Selection of participants: random- well defined characteristics  Lab Setting- controlled environment- results are generalized to real life  Real life occur naturally- sometimes natural settings cannot be brought in labs  How can someone study a riot in laboratory?  Examples????  Observation based on microcosmic life (as might be observed in labs) may not be valid for macrocosmic life (real-life)  Closed System: dealing with highly control system/environment- the lab has many virtues and many psychology experiments require this rigid control over stimuli  A Open System: is an environment over which we have no or little control- real world CAMPBELL (1969)  Record number of automobile fatalities in Connecticut- harsh actions were taken against speeders  Once these measures were introduced, a decline in traffic fatalities was noticed  Campbell made a detailed study of this phenomenon utilizing a quasi-experimental design  After the introduction of harsh treatment for speeders was introduced, a decline in traffic fatalities was seen.  Other factors may be responsible: road conditions, better driving education  Compare the fatality rates among comparable states  Higher decline in Connecticut ADVANTAGES AND DISADVANTAGES  Lab experiment could not be possible in the real life situation  With less experimental control Quasi Experimental Design renders a close approx to valid conclusion that one could expect DIFFERENCE'S B/W TRUE AND QUASI EXP. FUNCTIONAL DESIGNS FUNCTIONAL DESIGNS  Within Subject Design  Factorial Designs: Mean/ Percentage/ Frequency is compared of the scores of the participants by using statistical analysis  Experimental Analysis of behavior  Originated by B.F. Skinner-Skinnerians  Uses functional definitions of terms and concepts  Developing a functional definition of a concept (e.g. punishment) is accomplished by specifying the relationship between a set of conditions and their effects on behavior  Example: Punishment  Precisely Measurable  Researchers in this area- atheoratical  Concerns with testing the variables that control behavior than testing a theory  Instead of viewing an experiment as a means of theory testing, experimenters systematically explore variables that control behavior with the assumption that theory will emerge inductively from the data  Inductive approach towards theory development INDUCTIVE APPROACH DEDUCTIVE APPROACH FRESTER AND PERROTT (1968)  Skinner Box- 14 sq inches  Small Plexiglas plate-food magazine  If the plate was pushed, a food pellet would be released from the food magazine  A pen point rested on a strip of paper- each time the pigeon pecked on the Plexiglas plate, the pen moved- automatically note down the pigeons rate of response  Pigeon- with prior experience  80% of the free feeding weight through out the experiment  1 hr per day for 6 weeks  Pigeon was reinforced for pecking the plate on a fixed-ratio schedule  Pigeon received food after n number of times pecking the plate  Several fixed ratios were used  Week 1: 70 pecks (FR 70)  Week 2: 185 pecks (FR 185)  Week 3: 325 pecks (FR 325)  Then the order was reversed for the next 3 weeks RESULTS  Pecked 3-4 times per second for FR70- almost continuous pecking- slight pause after each reinforcement  FR185: long pauses after each reinforcement; rapid pecking until the food was appeared  FR325: longer pauses after each reinforcement; pecking rate same as for FR70 and 185 CONCLUSION  No. of pecks necessary for reinforcement do not influence the rate of pecking but does affect the length of the pause b/w the dispensing of a reinforcement and the resumption of pecking DIFFERENCE B/W FUNCTIONAL AND FACTORIAL DESIGNS Click to edit Master text Click to edit Master text Factorial Designs Functional Designs styles styles  Deductive Approach  Inductive Approach  Big samples  Small n  Sophisticated  No particular Statistical Analysis statistical analysis- typical curves SMALL N DESIGNS SMALL N DESIGNS  Other than approach (atheoretical) the number of subjects differs in functional designs- compared to factorial designs  Small n Designs use only one or two subjects in each experimental treatment  Data presented in responsive curve rather than in forms of means and variances  Curve is a segment of subject’s behavior that deemed his performance in Experimental Conditions ABA DESIGNS A B A Observation Introduction Removal of of the of experimental untreated experimental variable and behavior variable(s) measure of and behavior measures of behavior  A: is observed first  Baseline Data: observation/scores of the subject of untreated behavior  Departures from which contrast effect of experimental behavior  B: Experimental variable is introduced and its effects are measured  A: Experimental variable is absent and behavior is observed EXAMPLE  Behavior therapist wants to treat overeating  IV= Over Eating  DV= Weight of the client  A= Weight before the Treatment  B= Treatment (Behavioral Therapy)  A=Withdrawal of the treatment and again Weight of the client  Withdrawal Designs MULTIPLE IVS  Earlier Example: behavior therapy+ positive reinforcement + exercise  Critical Situation: do not permit unequivocal Cause-and-Effect relationship (have to mention that both the IV’s combined, produced the effect)  Combination of Treatments: Synergistic Effect  Cooperative action of the treatments could result in more effective reaction- profound behavioral consequences  Alternating Treatment Designs AAA DESIGNS  Internal validity  Baseline data of a single subject serves as a Control for subsequent observation A A A Observation No Measure of of untreated treatment behavior Behavior introduced, measure of behavior  EXTENDED VERSIONS  ABABA Designs: Experimental Variable is reintroduced A B A B A Observation of Experimental Withdrawal of Again Withdrawal of the untreated Treatment the treatment Experimental the Treatment behavior and the and the Treatment and the measurement measurement and measurement of the of the Measurement of the behavior behavior of the behavior Behavior  EXTENDED VERSIONS  AB1AB2A Designs: Experimental Variable is reintroduced A B1 A B2 A Observation of Experimental Withdrawal of Again Withdrawal of the untreated Treatment 1 the treatment Experimental the Treatment behavior and the and the Treatment 2 and the measurement measurement and measurement of the of the Measurement of the behavior behavior of the behavior  AB1AB2AB3 Behavior  Extensive Control AYLLON (1963)  A strength of response may decrease with continuous reinforcement called satiation  When animals were given a continuous reinforcement for a long period of time they stop emitting the reinforced response  Used the same procedure to control hoarding behavior in a psychiatric patient  47-years old female used to collect towels and kept them in her room  Nurses used to retrieve the towel but 20 towels per day on an average  Baseline was established on the basis of observation of seven weeks  Satiation period: nurses stopped removing the towels instead they used to bring the towels and handed over to her without any comments  7 per day 1st week- 60 per day 3rd week  Satiation period lasted for 5 weeks  Accumulated 625 towels and started removing the towels  Removing of towel continued until week 22  Average of 1.5 towel in the room and lasted till 28 weeks  Periodic observation was made until the next year- average was continued  Never returned to hoarding behavior-no other problems were replaced