NonExperimental designs Surveys Correlational Psych 231 Research Methods

  • Slides: 53
Download presentation
Non-Experimental designs: Surveys & Correlational Psych 231: Research Methods in Psychology

Non-Experimental designs: Surveys & Correlational Psych 231: Research Methods in Psychology

n n Mean = 74. 14 Median = 77 Max = 94 Min =

n n Mean = 74. 14 Median = 77 Max = 94 Min = 50 n Most common errors n Between vs. within designs Independent vs. dependent vars Scales of measurement n n Exam 2 results n n Confounds vs. extraneous variables Main effects vs. interactions

n Sometimes you just can’t perform a fully controlled experiment n n Because of

n Sometimes you just can’t perform a fully controlled experiment n n Because of the issue of interest Limited resources (not enough subjects, observations are too costly, etc). • • • n Surveys Correlational studies Quasi-Experiments Developmental designs Small-N designs This does NOT imply that they are bad designs n Just remember the advantages and disadvantages of each Non-Experimental designs

n n n Stage 1) Identify the focus of the study and select your

n n n Stage 1) Identify the focus of the study and select your research method Stage 2) Determining the research schedule and budget Stage 3) Establishing an information base Stage 4) Identify the sampling frame Stage 5) Determining the sample method and sampling size n Review Probability and Non-Probability methods • Voluntary response method n Importance of sample size Stages of survey research

n Sampling error - how is the sample different from the population? Often focus

n Sampling error - how is the sample different from the population? Often focus on this part But this part is important too n Confidence intervals • An estimate of the mean or percentage of the population, based on the sample data • “John Doe has 55% of the vote, with a margin of error ± 3%” • Margin of error (that “± 3%” part) • The larger your sample size, the smaller your margin of error will be. • Which would you be more likely to believe • “We asked 10 people …” • “We asked 1000 people …” Importance of sample size

n Sampling error - how is the sample different from the population? n Response

n Sampling error - how is the sample different from the population? n Response rate • What proportion of the sample actually responded to the survey? • Hidden costs here - what can you do to increase response rates • Non-response error (bias) • Is there something special about the data that you’re missing (From the people who didn’t respond)? Importance of sample size

n Stage 6) Designing the survey instrument n Question construction: How the questions are

n Stage 6) Designing the survey instrument n Question construction: How the questions are written is very important • Clearly identify the research objectives • Do your questions really target those research objectives (think Internal and External Validity)? • Take care wording of the questions • Keep it simple, don’t ask two things at once, avoid loaded or biased questions, etc. • How should questions be answered (question type)? 10 Stages of survey research

Poor Was the FDC negligent by ignoring the warnings about Vioxx during testing and

Poor Was the FDC negligent by ignoring the warnings about Vioxx during testing and approving it for sale? Yes Problem: a)emotionally b) No charged c) words Unsure Do you favor eliminating the wasteful excess in the public school budget? a) b) c) Yes No Unsure Good If the FDC knew that Vioxx caused serious side effects during testing, what should it have done? a)Ban it from ever being sold b)Require more testing before approving it c)Unsure Do you favor reducing the public school budget? a)Yes b)No c)Unsure Good and poor questions

Poor Should senior citizens be given more money for recreation centers and food assistance

Poor Should senior citizens be given more money for recreation centers and food assistance programs? a) Yes b) No Problem: asks c) Unsure two different questions Good Should senior citizens be given more money for recreation centers? a) b) c) Yes No Unsure Should senior citizens be given more money for food assistance programs? a) b) c) Yes No Unsure Good and poor questions

Poor Are you against same sex marriage and in favor of a constitutional amendment

Poor Are you against same sex marriage and in favor of a constitutional amendment to ban it? Good What is your view on same sex marriage? a) a) b) c) Yes No Unsure Problem: Biased in more than one direction b) c) I think marriage is a matter of personal choice I’m against it but don’t want a constitutional amendment I want a constitutional amendment banning it Problem: Asks two questions Good and poor questions

n Question types n Open-ended (fill in the blank, short answer) • Can get

n Question types n Open-ended (fill in the blank, short answer) • Can get a lot of information, but • Coding is time intensive and potentially ambiguous n Close-ended (pick best answer, pick all that apply) • Easier to code • Same response alternatives for everyone • Take care with your labels • Decide what kind of scale • Decide number/label of response alternatives Survey Questions What is the best thing about ISU? (choose one) n 1. Location n 2. Academics n 3. Dorm food n 4. People who sell things between Milner and the Bone

n Decide what kind of rating scales • Rating: e. g. , Likert scale

n Decide what kind of rating scales • Rating: e. g. , Likert scale PSY 231 is an important course in the major. 1 Strongly Agree 2 Agree 3 Neutral 4 Disagree 5 Strongly Disagree • Semantic differential: Rate how you feel about PSY 231 on these dimensions Important _____: _____: Unimportant Boring _____: Interesting _____: • Nonverbal scale for children: Point to the face that shows how you feel about the toy. Survey Questions: Close-ended

n Decide number/label of response alternatives • Use odd number (mid point and equal

n Decide number/label of response alternatives • Use odd number (mid point and equal # of responses above and below the mid point) • Questions should be uni-dimensional (each concerned with only one thing) • Labels should be clear Survey Questions: Close-ended

n Stage 7) Pre-testing the survey instrument n n Stage 8) Selecting and training

n Stage 7) Pre-testing the survey instrument n n Stage 8) Selecting and training interviewers n n n Fix what doesn’t seem to be working For telephone and in-person surveys Need to avoid interviewer bias Stage 9) Implementing the survey Stage 10) Coding and entering the data Stage 11) Analyzing the data and preparing a final report 10 Stages of survey research

n Sometimes you just can’t perform a fully controlled experiment n n Because of

n Sometimes you just can’t perform a fully controlled experiment n n Because of the issue of interest Limited resources (not enough subjects, observations are too costly, etc). • • • n Surveys Correlational Quasi-Experiments Developmental designs Small-N designs This does NOT imply that they are bad designs n Just remember the advantages and disadvantages of each Non-Experimental designs

n Looking for a co-occurrence relationship between two (or more) variables n We call

n Looking for a co-occurrence relationship between two (or more) variables n We call this relationship a correlation. n 3 properties: form, direction, strength Correlational designs

Linear Form Non-linear

Linear Form Non-linear

Positive Negative Y • X & Y vary in the same direction Direction Y

Positive Negative Y • X & Y vary in the same direction Direction Y X • X & Y vary in opposite directions X

r = -1. 0 “perfect negative corr. ” -1. 0 r = 0. 0

r = -1. 0 “perfect negative corr. ” -1. 0 r = 0. 0 “no relationship” r = 1. 0 “perfect positive corr. ” 0. 0 The farther from zero, the stronger the relationship Strength +1. 0

n Looking for a co-occurrence relationship between two (or more) variables n Used for

n Looking for a co-occurrence relationship between two (or more) variables n Used for • Descriptive research • do behaviors co-occur? • Predictive research • is one behavior predictive of another? • Reliability and Validity • Does your measure correlate with others (and itself)? • Evaluating theories • Look for co-occurrence posited by theory. Correlational designs

n Looking for a co-occurrence relationship between two (or more) variables n Example 1:

n Looking for a co-occurrence relationship between two (or more) variables n Example 1: Suppose that you notice that the more you study for an exam, the better your score typically is n At a descriptive level this suggests that there is a relationship between study time and test performance. n For our example, which variable is explanatory and which is response? And why? n It depends on your theory of the causal relationship between the variables n n Explanatory variables (Predictor variables) Response variables (Outcome variables) Correlational designs

Y 6 Hours study Exam perf. X Y 5 6 1 6 2 4

Y 6 Hours study Exam perf. X Y 5 6 1 6 2 4 5 6 2 3 4 1 3 2 Scatterplot n 3 1 2 3 4 For this example, we have a linear relationship, it is positive, and fairly strong 5 6 X

n n For descriptive case, it doesn’t matter which variable goes where n Correlational

n n For descriptive case, it doesn’t matter which variable goes where n Correlational analysis For predictive cases, put the response variable on the Y axis n Regression analysis Y 6 Response (outcome) variable 5 4 3 2 1 1 2 3 4 5 6 X Explanatory (predictor) variable Scatterplot

n Advantages: n Doesn’t require manipulation of variable • Sometimes the variables of interest

n Advantages: n Doesn’t require manipulation of variable • Sometimes the variables of interest can’t be manipulated n n Allows for simple observations of variables in naturalistic settings (increasing external validity) Can look at a lot of variables at once Example 2: The Freshman 15 (CBS story) • • Is it true that the average freshman gains 15 pounds? Recent research says ‘no’ – closer to 2. 5 – 3 lbs Looked at lots of variables, sex, smoking, drinking, etc. Also compared to similar aged, non college students Correlational designs

n Disadvantages: n Don’t make casual claims • Third variable problem • Temporal precedence

n Disadvantages: n Don’t make casual claims • Third variable problem • Temporal precedence • Coincidence (random co-occurence) n Correlational results are often misinterpreted Correlational designs

n Example 3: Suppose that you notice that kids who sit in the front

n Example 3: Suppose that you notice that kids who sit in the front of class typically get higher grades. n This suggests that there is a relationship between where you sit in class and grades. Daily Gazzett Children who sit in the back of the classroom receive lower grades than those who sit in the front. Possibly implied: “[All] Children who sit in the back of the classroom [always] receive worse grades than [each and every child] who sits in the front. ” Better: “Researchers X and Y found that children who sat in the back of the classroom were more likely to receive lower grades than those who sat in the front. ” Misunderstood Correlational designs Example from Owen Emlen (2006)

n Sometimes you just can’t perform a fully controlled experiment n n Because of

n Sometimes you just can’t perform a fully controlled experiment n n Because of the issue of interest Limited resources (not enough subjects, observations are too costly, etc). • • • n Surveys Correlational Quasi-Experiments Developmental designs Small-N designs This does NOT imply that they are bad designs n Just remember the advantages and disadvantages of each Non-Experimental designs

What are they? n n n Almost “true” experiments, but with an inherent confounding

What are they? n n n Almost “true” experiments, but with an inherent confounding variable General types 1) An event occurs that the experimenter doesn’t manipulate • Something not under the experimenter’s control • (e. g. , flashbulb memories for traumatic events) 2) Interested in subject variables – high vs. low IQ, males vs. females 3) Time is used as a variable Quasi-experiments

n Advantages n n n Allows applied research when experiments not possible Threats to

n Advantages n n n Allows applied research when experiments not possible Threats to internal validity can be assessed (sometimes) Disadvantages n n n Threats to internal validity may exist Designs are more complex than traditional experiments Statistical analysis can be difficult • Most statistical analyses assume randomness Quasi-experiments

Program evaluation n – Research on programs that is implemented to achieve some positive

Program evaluation n – Research on programs that is implemented to achieve some positive effect on a group of individuals. – – e. g. , does abstinence from sex program work in schools Steps in program evaluation – – – Needs assessment - is there a problem? Program theory assessment - does program address the needs? Process evaluation - does it reach the target population? Is it being run correctly? Outcome evaluation - are the intended outcomes being realized? Efficiency assessment- was it “worth” it? The the benefits worth the costs? Quasi-experiments

n Nonequivalent control group designs n with pretest and posttest (most common) (think back

n Nonequivalent control group designs n with pretest and posttest (most common) (think back to the second control lecture) Independent Non-Random Dependent Variable Assignment Measure Experimental group Dependent Variable Measure participants Measure Control group Measure – But remember that the results may be compromised because of the nonequivalent control group (review threats to internal validity) Quasi-experiments

n Interrupted & Non-interrupted time series designs n Observe a single group multiple times

n Interrupted & Non-interrupted time series designs n Observe a single group multiple times prior to and after a treatment Obs Obs Treatment Obs Obs • Look for an instantaneous, permanent change • Interrupted – when treatment was not introduced by researcher, for example some historical event n Variations of basic time series design • Addition of a nonequivalent no-treatment control group time series OOOTOOO & OOO_OOO • Interrupted time series with removed treatment • If treatment effect is reversible Quasi-experiments

n Advantages n n n Allows applied research when experiments not possible Threats to

n Advantages n n n Allows applied research when experiments not possible Threats to internal validity can be assessed (sometimes) Disadvantages n n n Threats to internal validity may exist Designs are more complex than traditional experiments Statistical analysis can be difficult • Most statistical analyses assume randomness Quasi-experiments

n Sometimes you just can’t perform a fully controlled experiment n n Because of

n Sometimes you just can’t perform a fully controlled experiment n n Because of the issue of interest Limited resources (not enough subjects, observations are too costly, etc). • • • n Surveys Correlational Quasi-Experiments Developmental designs Small-N designs This does NOT imply that they are bad designs n Just remember the advantages and disadvantages of each Non-Experimental designs

n Used to study changes in behavior that occur as a function of age

n Used to study changes in behavior that occur as a function of age changes n n Age typically serves as a quasi-independent variable Three major types n n n Cross-sectional Longitudinal Cohort-sequential Developmental designs

n Cross-sectional design n Groups are pre-defined on the basis of a preexisting variable

n Cross-sectional design n Groups are pre-defined on the basis of a preexisting variable • Study groups of individuals of different ages at the same time • Use age to assign participants to group • Age is subject variable treated as a between-subjects variable Age 4 Age 7 Age 11 Developmental designs

n Cross-sectional design n Advantages: • • Can gather data about different groups (i.

n Cross-sectional design n Advantages: • • Can gather data about different groups (i. e. , ages) at the same time Participants are not required to commit for an extended period of time Developmental designs

n Cross-sectional design n Disavantages: • Individuals are not followed over time • Cohort

n Cross-sectional design n Disavantages: • Individuals are not followed over time • Cohort (or generation) effect: individuals of different ages may be inherently different due to factors in the environment • • • Are 5 year old different from 15 year olds just because of age, or can factors present in their environment contribute to the differences? • Imagine a 15 yr old saying “back when I was 5 I didn’t have a Wii, my own cell phone, or a netbook” Does not reveal development of any particular individuals Cannot infer causality due to lack of control Developmental designs

n Longitudinal design n Follow the same individual or group over time • Age

n Longitudinal design n Follow the same individual or group over time • Age is treated as a within-subjects variable • • Rather than comparing groups, the same individuals are compared to themselves at different times Changes in dependent variable likely to reflect changes due to aging process • Changes in performance are compared on an individual basis and overall time Age 11 Age 15 Age 20 Developmental designs

n Example n Wisconsin Longitudinal Study (WLS) • Began in 1957 and is still

n Example n Wisconsin Longitudinal Study (WLS) • Began in 1957 and is still on-going (50 years) • 10, 317 men and women who graduated from Wisconsin high schools in 1957 • Originally studied plans for college after graduation • Now it can be used as a test of aging and maturation Longitudinal Designs

n Longitudinal design n Advantages: • Can see developmental changes clearly • Can measure

n Longitudinal design n Advantages: • Can see developmental changes clearly • Can measure differences within individuals • Avoid some cohort effects (participants are all from same generation, so changes are more likely to be due to aging) Developmental designs

n Longitudinal design n Disadvantages • Can be very time-consuming • Can have cross-generational

n Longitudinal design n Disadvantages • Can be very time-consuming • Can have cross-generational effects: • Conclusions based on members of one generation may not apply to other generations • Numerous threats to internal validity: • Attrition/mortality • History • Practice effects • Improved performance over multiple tests may be due to practice taking the test • Cannot determine causality Developmental designs

n Cohort-sequential design n Measure groups of participants as they age • Example: measure

n Cohort-sequential design n Measure groups of participants as they age • Example: measure a group of 5 year olds, then the same group 10 years later, as well as another group of 5 year olds n Age is both between and within subjects variable • Combines elements of cross-sectional and longitudinal designs • Addresses some of the concerns raised by other designs • For example, allows to evaluate the contribution of cohort effects Developmental designs

n Cohort-sequential design Cross-sectional component Time of measurement 1975 Cohort A 1970 s Cohort

n Cohort-sequential design Cross-sectional component Time of measurement 1975 Cohort A 1970 s Cohort B 1980 s Cohort C 1990 s Age 5 1985 1995 Age 15 Age 25 Age 15 Age 5 Longitudinal component Developmental designs

n Cohort-sequential design n Advantages: • Get more information • Can track developmental changes

n Cohort-sequential design n Advantages: • Get more information • Can track developmental changes to individuals • Can compare different ages at a single time • Can measure generation effect • Less time-consuming than longitudinal (maybe) n Disadvantages: • Still time-consuming • Need lots of groups of participants • Still cannot make causal claims Developmental designs

n What are they? n n Historically, these were the typical kind of design

n What are they? n n Historically, these were the typical kind of design used until 1920’s when there was a shift to using larger sample sizes Even today, in some sub-areas, using small N designs is common place • (e. g. , psychophysics, clinical settings, expertise, etc. ) Small N designs

n n One or a few participants Data are typically not analyzed statistically; rather

n n One or a few participants Data are typically not analyzed statistically; rather rely on visual interpretation of the data Observations begin in the absence of treatment (BASELINE) Then treatment is implemented and changes in frequency, magnitude, or intensity of behavior are recorded Small N designs

n Baseline experiments – the basic idea is to show: 1. when the IV

n Baseline experiments – the basic idea is to show: 1. when the IV occurs, you get the effect 2. when the IV doesn’t occur, you don’t get the effect (reversibility) § § Before introducing treatment (IV), baseline needs to be stable Measure level and trend Small N designs

n Level – how frequent (how intense) is behavior? n n Are all the

n Level – how frequent (how intense) is behavior? n n Are all the data points high or low? Trend – does behavior seem to increase (or decrease) n Are data points “flat” or on a slope? Small N designs

n ABA design (baseline, treatment, baseline) – The reversibility is necessary, otherwise something else

n ABA design (baseline, treatment, baseline) – The reversibility is necessary, otherwise something else may have caused the effect other than the IV (e. g. , history, maturation, etc. ) ABA design

n Advantages n n n Focus on individual performance, not fooled by group averaging

n Advantages n n n Focus on individual performance, not fooled by group averaging effects Focus is on big effects (small effects typically can’t be seen without using large groups) Avoid some ethical problems – e. g. , with nontreatments Allows to look at unusual (and rare) types of subjects (e. g. , case studies of amnesics, experts vs. novices) Often used to supplement large N studies, with more observations on fewer subjects Small N designs

n Disadvantages n n Effects may be small relative to variability of situation so

n Disadvantages n n Effects may be small relative to variability of situation so NEED more observation Some effects are by definition between subjects • Treatment leads to a lasting change, so you don’t get reversals n Difficult to determine how generalizable the effects are Small N designs

n n Some researchers have argued that Small N designs are the best way

n n Some researchers have argued that Small N designs are the best way to go. The goal of psychology is to describe behavior of an individual Looking at data collapsed over groups “looks” in the wrong place Need to look at the data at the level of the individual Small N designs