Mvo alles

Survey

Asking people questions face-to face, on the phone, on written questionnaires, on online.

Question formats

writing well-worded questions

encouraging accurate responses

Open-ended items

respondents can answer how they like

considerations

Closed-ended items

Provide responses by choosing between two options

considerations

Likert scale

using rating scale to reflect degree of agreement

Rosenberg selfesteem inventory

Semantic Differential format

anchored on meaning

BU books

Effective questionare

• Brief

• relevant

• unambiguous

• specific

• objective

Leading Questions

questions that lead to participant to answer in a certain way

Double-Barreled questions

two questions in one

Negatively Worded Questions

Questions that use strong works that lead to a negative outlook. Words like: Impossible and never

Question order

Importance in the order the questions are in as to not prematurely effect the participants response

Encouraging Accurate responses

People can give meaningful responses to survey. Techniques used to ensure accuracy

Acquiescence

Adding two questions that are clear contradictions to ensure that participants are being honest and reading the questions

Fence-sitting

When a participant only choses the most middle options because they don’t want to commit to one side. This can be avoided by eliminating a middle option.

Socially Desirable Responding

People respond based on what they think to be the correct answer

Self-reporting memories of events

memory is flawed

avoid direct recall of personal events

Sampling

Sample members of the population to be part of our sample

hope that the results from our sample can apply to the entire population

Probability sampling

• each individual in the population must have a specified probability of selection

• the selection process must be unbiased; must be a random process

Non-probability sampling

Odds of selecting a particular individual are unknown

Simple random sampling

Assign every population member a number

randomly select a number

Cluster sampling

randomly select groups (clusters) from the population

use all members of those groups (clusters)

Multistage sampling

Stage 1: random sample of clusters is selected from population of interest

Stage 2: random sample from clusters make up whole sample

Stratified random sampling

population divided into strata

strata are all equal

randomly sample from strata in proportion with population

Oversampling

researcher overrepresent one or more groups in the sample

usually to endure traditionally underrepresented groups are present in a sample

Convenience sampling

sampling participants who are easily available

quota sampling

subgroups are identified to be included

select individuals in each subgroup via convenience sampling

allows a researcher to control the compositions of a convenience sample

systematic sampling

every nth participant is selected from a list of the entire population

a random starting position is chosen

Sampling bias

bias is a major threat

a biased samples characteristics are noticeably diffrent from those of the population

participants or subjects are selected in a manner that increases the probability of obtaining a biased sample

self-selection

certain people choose to participate

Bivariate Correlations

associations that involve exactly two variables

make claims or inferences about the relationship or association between two variables

Studies are correlational if

both variables are measured

no manipulated variables

Correlation Coefficient

a number that represents the degree of association between two continuous variables

two pieces of information from correlation coefficient

strength of the relationship

direction of the relationship

strength of the relationship

1=perfect correlation

0=no correlation

direction of the relationship

• positive(+)= as one variable increases/decreases, the other variable increases/decreases in the same direction

• Negative(-)= as one variable increases/decreases, the other variable decreases/increases in the opposite direction

Effect size

when everything else is equal, a larger effect size is usually considered more important than a small one. but there are some exceptions

statistical significance

whether there is less than a 5% chance we would hace found this correlation coefficient if the null hypothesis is true

• if p<.05 (typical alpha value), the correlation is significant

• if p>.05, the correlation is not significant

Factors affecting r

• outliers

• restriction of range

• curvilinear association

Outliers

one or more extreme scores

influence is larger with smaller samples

Restriction of range

the values of one or more variables has been reduced

can make the correlation appear smaller

Curvilinear association

the relationship between two variables is not a straight line

r it 0

Directionality problem

can’t tell which one is going in which direction

Third variable problem

a missing third variable that has an influence on both previous variables

Covariance

can show relationship between variables

temporal precedence

if the variables are measured at the same time, no temporal precedence

don’t know which variable comes first

internal validity

we’re only measuring two variables

other factors that can play a role

Regression

regression and correlation are mathematically the same

• the coefficients are in diffrent units

• we use regression when we want to predict one variable from another

• we use correlation when we want to look at the association between two variables

regression X

X= participants observed score on the predictor variable

• independent variable

• we know this, use it to predict y

Regression Y

Y= participants observed score on the outcome variable

• dependent variable, sometimes called criterion variable

• Y’= the participant score on Y that is predicted from the regression equation

Regression Line

Similar to the equation for a straight line Y’=a+bX

• a is the y-intercept

• b is the slope of the line

• X is the participants score on the predictor (independent) variable

• Y’ is the participants predicted score on the outcome(dependent) variable

Coefficient of Determination

The squared correlation coefficient (r²)

the proportion of variation in the Y variable that can be accounted for by the X variable

can be used to interpret the association between two variables

Correlation does not imply Causation

saying to remember

Multivariate Designs

Involve more than two measured variables

Criteria for establishing causation

covariance

temporal precedence

internal validity

Controlling for variables

holding them constant

want to see if the relationship between two variables still exists after controlling for a third variable

Conceptual Example

is exposure to sexual TV content associated with pregnancy risk among 16-20 year old girls

potential third-variable problem?

observational studies

systematic recording of human or animal behavior

Naturalistic observation

observing behavior in the natural environment; no laboratory

Structured observation

observing behavior in the natural environment but the researcher has intervened in the situation

Observer bias

observers see and code what they expect to see

observer effects

Behaviors are disrupted or influenced by the presence of an observer

reactivity

Solutions to observational studies

• observer training

• masked design

• conceal the observer

• habituate to the observers presence

• measure the behaviors results

Things to keep in mind for observational studies:

• replication is incredibly important

• construct validity is essential

• quantify your observations whenever possible

Case studies

in-depth description of one particular individual

Importance of replication

Observational or case studies can be anecdotal

replication+extension

Baseline

measured before any treatment is introduced

reversal design

baseline, treatment, baseline

ABA design

ABAB…design

data analysis

single-subject research relies heavily on visual inspection

• Level

• trend

• latency

Level

How high or how low the level is

trend

do the data points tend to trend upwards, downwards, or stay level

Latency

how quickly is the behavior changing when adding or taking away treatment

Multiple-treatment reversal design

Phases that introduce different treatments that are alternated

ABCACB design

Potential Problems

may be unethical to remove treatment

dependent variable may not return to baseline when the treatment is removed

multiple-Baseline Designs

multiple baselines are established

treatment is introduced as a diffrent time for each baseline

Data analysis

Can use statistical procedures

Problems with data analysis

• Can we detect weak effects

• can be unreliable

• the results of a visual inspection cannot be clearly and efficiently summarized or compared across studies

Internal validity

can be very high if the study is carefully designed

External validity

Can be problematic depending on the goals of the study

Construct validity

can also be very high if definitions and observations are precise

What are the four types of experiment?

• Lab

• Field

• Natural

• Quasi

Which type of experiment is similar to lab experiments?

Field experiments

Which type of experiment is similar to natural experiments?

Quasi experiments

Key points of Laboratory experiments

• Researcher manipulates IV

• Controlled setting

• Standardised procedure

• Artificial task (designed for the purpose of the study)

• Participants aware (susceptible to demand characteristics)

Key points of Field experiments

• Researcher manipulates IV

• Natural (uncontrolled) setting

• Procedure may be somewhat standardised

• Real life task or one that appears to be real

• Participants usually unaware (unlikely demand characteristics)

Key points of Natural experiments

• Researcher does not manipulate the IV - it is a naturally occurring change in real life e.g., if the IV is winter/summer, P1/P5 or before/during lockdown.

• Participants can be tested in either a lab style or a field style.

Key points of Quasi experiments

• Researcher does not manipulate the IV - it is an existing difference between people e.g., if the IV is age, gender, mental illness.

• Participants can be tested in either a lab style or a field style.

Give two strengths of laboratory experiments

• High in reliability, because of standardisation and control. Therefore, the results should be replicable.

• High in internal validity, because of standardisation and control. Therefore, we can be surer that any effect on the DV is due to the IV.

Give a weakness of laboratory experiments

• Low in ecological validity, as the experiment is an unnatural environment and participants are being tested, so may behave unnaturally (demand characteristics)

Give a strength of field experiments

• High in ecological validity, because the task is natural, the participants are naïve and the setting is natural. Therefore, we can assume the participants will be behaving naturally.

Give two weaknesses of field experiments

• Low in reliability. Not a lot can be standardised, so there are lots of situational variables affecting the experiment.

• Low in internal validity due to lack of standardisation, control and situational variables. Therefore, we cannot be sure that any effect on the DV is due to purely to the IV.

Give a weakness of natural and quasi experiments

Participants cannot be randomly allocated to the conditions . So, there may be confounding participant variables, so the research may lack validity as any effect on the DV may not actually be due to the IV,

Give a strength of natural and quasi experiments

They provide the opportunity to study things that would otherwise be impossible to study because it would be unethical or impossible to manipulate the IV e.g., f the IV was abused/non abused children.

What do other strengths and weaknesses of natural and quasi experiments rely on?

Whether the experiment has a more of a laboratory or field style of testing the participants.

aim

• a statement of what the researcher intends to find out in a research study

• identifies the purpose of the investigation

hypothesis

• a precise and testable statement about the assumed relationship between variables

• should be fully operationalised

operationalised

• the variables and how they will be measured must be clear

• makes a statement testable

independent variable (IV)

• something that can be manipulated (changed)

dependent variable (DV)

• something that can be measured

directional hypothesis
(one-tailed)

• states the kind of difference between two conditions or groups

• direction of the predicted difference

non-directional hypothesis
(two-tailed)

• simply states there will be a difference between two conditions or groups

• does not specify the difference

null hypothesis (H0)

• a hypothesis that states there will be no change / impact