means of simplifying and understanding complex realities

-all wrong, some are useful

strong theory one from which a hypothesis can develop

explain relationships among phenomena 

hypothesis- clear concise testable statement, educated guess

EX: Newton's theory
EX: make guess that high sat score predicts college performance 

info about the frequency or amount of something

ex: sports stats

descriptions about the difference between groups

ex: male pilot fighter vs female

descriptions of the relationship between or among variables

-1 group, 2 variables

ex: study time and grades

strategy to produce strong evidence for a cause and effect relationship between two variables

-types

true: random assignment
 quasi: non-random
single- subject 

-consistency of measurement

-factors that influence:

a. what you measure: physical, cognitive, affective

b. instrument used

c. technique

-deals with ensuring an experimental process that generates consistent results

validity

-external vs. internal

-how meaningful, useful and appropriate our conclusions are

-characteristic of the test RESULTS (not the actual test)

-internal vs. external

-internal: extent to whichindependent variable and not other variables produce the observed change in the dependent variable 

a. confident tht our manipulation of variable had the effects rather than the other variables

ex: Gatorade not the the water that improved athletes performance

external: extent to which results can be generalized to other subjects, settings, time

ex: could us behavior strategies for students could not be generalized to other settings, subjects and time 

categorical (graph= bar or pie chart, stats= mode)

a. nominal: named categories (race, gender)

b. ordinal: ordered categories (grade level, race finish)

continuous- could plot on a number line. (graph= histogram, line. stats: mode, median, mean)

a. interval: equal intervals, no real 0s (temperature)

b: ratio: equal intervals with an absolute 0 (height, date/time)

-relationship of one score to the norm group in terms of standard units

-z = (raw score- control group mean)/ (control group SD)

-used to find difference bw an indiv score and the group mean

measure of the magnitude and difference of the means of two groups

-ES= (exp group mean- control group mean)/ (control group SD)

-used to find the effect of one variable on another

-magnitude and direction of the difference bw 2 scores

-predictability, how related 2 variables are. measure the strength of the relationship

-sensitive to outliers, move r close to 0

-strength: scale 0 to 1

-direction: positive or negative

graph shapes:

a. straight line= 1

b. corn on the cob= .75

c. football= .5

d. circular= 0

role: review and ensure human subjects involved in research adequately protected

1. autonomy: respect for persons

a. protect with informed consent. must be informative, comprehensive, and indicate

2. voluntariness

beneficence: maximize benefits, minimize harms

a. protect with a risk/benefit assesment involving nature and scope of farm, use a systematic assesment of favorable ratio

3. justice: distribute benefits and burdens fairly

a. protect with considerations in selection of subjects

study of culture (groups of people who share behaviors and beliefs)

ex: fast times at ridgemont high

-exploration of a bounded system (group, indiv, event or setting)

ex: Woodstock

phenomenon that transcends culture, place, and time

ex: <3 and war

-using inductive logic to generate a model or theory

-observe, think. look for patterns and provide simple way to explain

ex: Piaget's stages of development

choose number that reflects answer preference 

ex: rank chocolate 0 to 10

put in order

ex: what was favorite, 2nd, etc. 

strong directional statment; to what extent do you agree or disagree.

strongly agree, agree, neither, disagree, strongly disagree

ex: clickers?

-someone tries to influence you with a questionaire; purpose is not to get opinion but rather change mind about something

-signs:

a. do not collect or tabulate survey results

b. doesnt make demographic inquiries, not based on subset

c. list is used to select respondents belonging to pop subset

d. prefaces question with untrue statement

-effect: happens in both quant and qual research; presence of the observer influences the natural behavior of those observed. In exp design, hawthorne effect

-observer bias: seek confirmatory info, invalid info or conclusions due to researcher's perspective 

a. personal experience

b. common sense and logic

c. tradition (conventional wisdom)

d. authority

e. systematic inquiry

-best because it focuses on examining the process itself as well as the results 

-collects and logically analyzes info to provide explanations

a. every reasonable statement about human behavior seems obvious but it cannot all possibly be true

b. if given a reasonable explanation, every statement seems obvious

ex: list of claims, their opposite was true

a. correlation: measure of strength and direction bw 2 variables

-simply observe the relationshio, no manipulation

b. causality: proposes a cause and effect relationship, determines what the relationship is rather then how strong it is.

-Studied through experimentation

-relationship studied through experimentation involving manipulating one variable to see its effect on another

c. benefits of correlation: predictive power

-notice a relationship bw two variables w/o having to engage in causal experimentation 

-demonstrates trends over time, shows ones that increase and ones that decrease

*all variables are correlated to some degree

similarities

-forms of systematic inquiry

-subject to IRB approval (research ethics)

-results are tentative

-involve subjects and participants

-agreed upon ways of collecting data

differences (on specific group of people)

differences

-sampling: qual= most knowledgable sample, quantitative= most representative sample

-role of the researcher: qual= subjective measuresments, individual interpretation, quant= objective, seeks precise measurement and analysis of target concepts

-design: quant: predetermined design, qual: change as gom evolve *expected outcomes present in quant not qual

-data presented: quan= stats, qual= quotes, words

-goals of research: qual= create an understanding, often through gt, quant= test theory, establish cause and effect relationships and describe things through descriptive stats

a. belief must be correct

b. belief must be justified

c. person must be certain of belief, enough to make decisions 

ex: people drive on right side of road in uk

a. ...they do

b. i went to england and saw it and read about it

c. if i went to england i woul drive on the right side of the road

a. choose the right graph for your data

b. title, label, include units

c. no color, shading/ gradients, 3d

d. consistent from graph to graph

e. keep it simple, the less ink, the better

a. quant

i. variables 

ii. relationships bw variables 

iii. subjects

ex: what is relationship bw final grade and overall course evaluation for vandy students in 2011?

b. qual

i. central phenom

ii. particpants

iii. site

why do teachers with less than 3 years of teaching experience choose to leave the field? 

sampling error: sample value is never exactly equal to the pop value

-the larger the sample, the smaller the sampling error

sampling fluctuation- occurs when we measure a value for samples repeatedly drawn from the same pop. the value for each sample is different from the others (and different from the true value of the pop)

statistical significance- mathematical test that gives yes or no answer to the question "are the differences we see larger than we would expect from sampling fluctuation alone?

relationship: results are stat significant if differences occur for reasons other than sampling error and sampling fluctutations 

1. stat significance: mathematical test that gives yes/ no answer to the question, "Are the differences we see larger than we would expect from sampling fluctuation along?"

-based on sample size

-tells if differences we see are larger than we would expect from sampling fluctuations

2. effect size: represents the distance bw the mean of the exp group and the mean of the control group in units of the sd

-tells magnitude and direction of the differences

3. practical significance- tells us how important the differences are in terms of what people value

-answers so what

qual

goals:

Select knowledgable participants about topic/ phenomenon being researched
strategies:

1. snowball- networking, select participants based off reccomendations from other participants

2. extreme cases- select a unique or atypical particpant 

3. maximum variation- select at least 2 participants that are extreme cases

quant

goals:

-select a sample that is representative of the popilation you are trying to generalize your results to 

strategies:

-simple random- assign a number to each subject and randomly select numbers via computer, out of a hat etc

-cluster sampling: select naturally occurring groups first and randomly select students from these groups (ex: school or class for educational research)

-convenience sampling- select samples that are easy to access (ex: psych students). worst bc cant generalize results

a. use a control group- reduce threats to internal validity

b. randomly assigns subjects- ensure equivalanve bw two groups, on dpdt measures as well as all others

c. pretest- to test for equivalence for the groups at the start and give the baseline data to calculate pretest/postest data

d. treatment that subjects get: experimental or something unrelated to the dv

e. postest- measure difference bw pre and post. determine if stat significant/ effect size

f. delayed retention test- see how long-lasting effects are

1. ask questions in a way that people understand in the way you would like them to be understood

2. convey the kind of answer that you would like

3. ask questions that people are capable of answering

4. ask questions people are willing to answer

5. able to be administered in standardized (developed, used, analyzed) way

1. helps refine your question

2. establish the significance of the problem for self and reader

3. understand or learn the methods you could use to do your research

4. can find contradictory research then help explain the contradictions

5. learn new info

5 strategies to increase reliability of observations

1. target specific behaviors

2. use low-interference measures

3. use multiple judes and train them

4. make judges blind to conditions

5. strive for inter-rater reliability

1. target specific behaviors

2. use low-inference measures

3. use multiple judges and train them 

4. make judges blind to conditions

5. strive for inter-rater reliability

1. triangulation: different methods of collecting data (interviews, observations, document analysis)

2. researcher reflexivity: tell reader who you are and be sensitive to your own perspective 

3. seek disconfiriming evidence: develop ways to break theory and strengthen it

4. prolonged and persistent field work: more hours, more people

5. member checking: check what people said, check facts

1. physical traits (ex: height)

2. cognitive traits (info you know, for exam etc)

3. affectove traits (feelins, personality, most variable)

used to descrive a relationship bw something that has occurred and correlated variables controlled through matching in order to show cause and effect

-useful for ethical dilemma

use pair wise matching: match subjects bw groups to minimize extraneous variables 

conducting smoking/ pregnancy example:

get sample, split, do pairwise matching, tally effects

-graphed aba or abab if concluding with second treatment

-a= baseling, b= treatment

-line graph

requirements for graphing:

a. reliabile measure (at least 4 days)

b. repeated measure

c. clear description and control of conditions

d. baseline and treatment conditions

e. one variable at a time is investigated 

ex: emily throwing tantrums, will treats help?

INTERVIEWS

-strengths: flexible and responsive (can ask new questions based on convo), build rapport

-limitations: costly, interviewers lead respondents

OBSERVATIONS

-strengths: eliminates recall error (saying um during speech can be pointed out by observer), observe soemthing subjects are not capable or willing to answer (ex: little kids)

-weaknesses: dont know root cause, takes a long time to do well

FOCUS GROUPS. 6-12 people w/ moderator

strengths: very fast, relatively cheap

limitations: produces an environment for wish fulfillment, one person may dominate the focus group and some may self-censor

a. what you believe

b, why you believe it

c. change mind would need evidence

d. discuss research