Social Research Methods Exam 2

population

who you are studying

census

full counting of the population with basic info, meaning you ask EVERYONE in the population.

for this, you must talk to EVERYONE in the US

every 10 years bc expensive

• time

• expensive

• willingness to participate

• ability to find everyone

limits to census study

sampling

process of who you are going to observe in ur study

non-probability sampling

A type of sampling

• we do NOT know the likelihood (the probability) that any one subject is in our sample

• does NOT allow us to generalize to our population

probability sampling

A type of sampling

• we DO know the likelihood (the probability) that any one subject is in our sample

• DOES allow us to generalize to our population

• Goal: to draw conclusions about target population (i.e. to generalize!)

probability sampling and pnon-robability sampling

opposite types of sampling.

generalizable

I can generalize from my sample to my population. My findings are ____

external validity

the conclusions we draw from our study accurately reflect or are generalizable to the real world.

Do your findings speak to anything outside of your study?

convenience

type of non-probability sample.

• people who are really easy to talk to

• people standing at the store with a clipboard

• Benefits: chip and easy

• Drawbacks: not a probability sample so we cannot generalize it to the whole population.

surveys

what is the most common type of research but often done very poorly?

• to describe a big population thats hard to talk to everyone in

• can be a probability sample if done well

when are surveys helpful?

• constraints peoples responses, you must choose one of the choices even if youre between 2

• not good with rich descriptions and context

• inflexible, once you send out the first survey you cant change it.

weaknesses of surveys

primary design

type of design:

• design and field your own study

secondary design

type of design:

• rely on surveys that already exist in the world

• advantages: cheaper, faster, accessible

• disadvantages: data doesnt perfectly align with research purposes/questions.

survey instrument and survey strategy

what are 2 things you need for a GOOD sample?

sampling frame

Complete list of all people in our target population

the actual list you get your sample from

sample error

how close does my sample resemble my population?

our sample doesnt look like the population.

nonresponse bias

you might get this if you give a “prefer not to answer option” on a survey.

end

put demographic questions at the end or beginning of the survey?

open ended and closed ended

2 types of questions in a survey

social desirability bias

no leading questions, biased terms like “dont you agree”

experiments

what is the ONLY method that can establish causation? surveys cannot do this.

• to manipulate a variable

• to establish causation

when are experiments useful?

non-spurious, time ordering, and correlation

3 conditions for causation

experimental and control/independent and dependent

what are 2 groups you need in an experiment?

laboratory experiments

types of experiments:

• take subject and bring it to your lab/office

• you have the control

• highest internal validity

• Low external validity (setting of the lab is so articial)

• uncommon for crim/sociology

field experiments

types of experiments:

• set up experiment somewhere, playout in the real world

• much higher external validity then a lab, because you are in the real world

• very strong casuality

• Bad: even if you see causal relationship, you cant say why

• hard to get a truly random sample

survey experiments

types of experiments:

• easier to control conditions and use a probability sample with a random assignment

• weakness: low external validity, rely on self report, subject to social desirability bias

natural experiment

types of experiments:

high external validity since realistic

Weaknesses:

“Rely on finding things opportunistically”

“Not really as random as we’d like (place where opportunities arise are different)”

experiments

what is the ONLY study to show causation and a casual relationship?

variable

measured piece of info (age, height)

univariable analysis

describes at one variable

bivariate analysis

describes at 2 variables

multi-variate analysis

looks at more then 2 variables

frequency

how many people selected that answer

mode

most common value of the variable, comes up the most

mean

the average, add all the answers and divide by the number of responses.

skewed right

distribution of data: tail is on the right

skewed left

distribution of data: tail is on the left

symmetrical

distribution of data: hill is in the middle

median

what is the best estimate to be in the middle if the distribution of the data is skewed.

median

middle, half people are more then __ and half are less.

range

smallest possible number to biggest

standard deviation

how narrow or wide the distribution is

crosstabs

the frequency distributions of a bivariate analysis

generalize a population

goal of sampling

case oriented sample

types of sampling: ask deeper questions, range of concepts, which features appear together

key informants

strategy:

• select an individual or a few to provide a composite pictures of the group

  • Gangs: get heads of the crips

• benefits: can get some sense of whats going on in difficult to reach or resistant populations

• drawbacks: bias of informant, cannot generalize, sample still not representative.

whatever number you need to hit, your goal. number of respondents.

when are you done with your probability sample?

when you’ve reach saturation. when each additional interview does not provide anymore or new informations

when are you done with your non-prob sample?

probability sample

what sample is never generalizabke?

quantitative

conducting numerical data is what type of research?

qualitative

conducting rich research

• ethonography (go imbed yourself in what your studying, live in public housing)

• content analysis (coding reddits, newspaper)

• case studies

• in depth interviews (use case-orienteded sampling strategies with goal of saturation

unstructured

when you go into an interview with only themes, not exact questions

structured

when you go into interviews with a prepared list of questions. has follow ups. one question, yes or no

• deep rich understanding

• flexible, can modify research design at anytime

• natural setting, find them where they are at

• only requires one person

• allows for an inductive approach.

strengths of qualitative research

• never generalizable

• observations based on one researches perception, have to be very reflecivbe

• measure develiped apply to one siutiation only

• nake sure you are very reflexive, no bias

weaknesses of qualitative research

prompt question

this is a type of question where you rephrase the question, get back on track

this is a type of question to clarify if you understand, more detail/explanaition

rational systems

Types of theory building

• rule/regulations based on decision making on based on formal rules and procedures.

• if your job says policy is no criminals then we have no discretion

natural systems

Types of theory building

• individual assessments of the weight of a criminal record when making employment decisions so theres no laws or rules about hiring criminal records

coding

classifying or categorizing individual pieces of data

variable based coding

deductive process where you have developed codes in advance and then find them in the data

open coding/ in vivo coding

inductive process starting with the data and letting concepts emerge

saturation

no longer getting new information. all codes from previous intervuews have applies and there are no new codes.

memoing

writing notes to clarify thoughts and codes as you go. code notes.