Research and Sampling techniques

5 Sampling techniques

VORSP

* volunteer
* opportunity
* random
* snowball
* purposive
* *very often rats steal pizza*

volunteer strengths

* More target population control (compared to opportunity) because the researcher has the opportunity to change their advertisement’s location 


* Ethically sensitive: control in the audience & no forcing. Good for autonomy

volunteer weaknesses

* participants can ignore advertisement 


* Not as efficient because it depends on the individual to take initiative
* Necessary to consider what type of audiences will view your advertisement → may take more work from the researcher to categorize volunteers into target populations and not
* Aggressive and extremely persuasive advertising needed

opportunity strengths

* Very convenient and efficient
* Doesn’t have biased information
* Quick
* Cheap
* Not a lot of materials/preparation needed

opportunity weaknesses

* No control over selection


* people might not care and have poor quality or even unreliable information/answers
* cannot be generalized
* may be biased
* could be said that it's not effective when studying a specific group

random strengths

* Relatively simple to get sample as there was a name list given


* Wide sample, getting people who are different from one another
* Not much thinking involved when selecting people (computer generated)

random weaknesses

* contact details needed and may ignore contact
* less control over traits of participants
* research becomes dependent
* must represent the target population proportionately

\

purposive strength

* control the type of people involved in your study


* Time and cost-efficient
* Small margin of error (the people chosen fit a specific criteria 
* Beneficial for the understanding of a specific narrow group inside the target populations.
* Good for examination of rare topic/group

purposive limitation

* certain types of studies (eg. a rare or unique condition)


* Might not be a good vary representation of the target population 
* Bias heavy
* Data invalidation / ethical
* Not effective on a large scale 
* Difficult to collect participants as  it depends on whether your target population aligns with the people you know

quanitative data

* predicting & testing


* Numerical data 
* Non-probing (closed) questions 
* Large sample sizes
* Narrow focus (deliberately isolates variables)

qualitative data

* understanding and describing how/why


* Non-numerical data
* Rich, in depth and detailed 
* Uses probing (open) questions
* Small sample size often

Lab and Field experiments

* IV manipulated
* determines C+E between two variables
* often quantitative

“true” experiements

**four elements**: manipulation, control. random assignment, and random selection

Lab vs Field experiments

* both IVs manipulated
* Field “natural setting”
* Laboratory: controlled to minimize extraneous

In field experiments, why do researchers have to balance realism with control?

Realism is necessary to ensure it is actually applicable and generalizable to real-world applications. However, it needs control to be certain about C+E relationship without extraneous variables

Field experiment issues

* observation/experimentation without knowledge
* cannot dismiss or leave the experiment at any point
* lack of informed consent on goals, about, risks & longevity

CT of Field experiment

* lack of consent, deception, lack of debrief
* control of variables -- replicability
* relative generalizability and no participant bias

CT lab experiment

* replicability due to controlled variables
* lack of generalizability if too controlled
* consent may be obtained
* bias may occur

correlation

* relationship between two studies
* **co-variables**
* IV cannot be manipulated
* positive, negative and zero correlation

Why do correlations not showcase causation between two variables?

* cannot manipulate one variable to showcase C+E

Strengths of correlation

* understands complex relationships without unethical practices

What does it mean for two variables to correlate with each other?

the degree to which two variables move in coordination with each other.

correlational CT

* avoids ethical practices
* cannot establish C+E
* bi-directionality -- which affects?
* third variable: Another variable altogether that is not being measured affects the other variable(s)

Naturalistic observations

* no IV
* Observation that occurs in a natural setting, with limited control of variables to observe behaviors in natural and realistic settings

4 types of naturalistic observation

* **Participant observation:** The researcher is part of the group being observed
* **non-participant observation:** Researcher is not part of the group being observed


* **covert**: participants are unaware they are being observed
* **overt**: participants are aware they are being observed

How are observations different to experiments?

no IV manipulated

Which types of observation may naturally create ethical issues? Why?

* covert because participants are unaware they are being recorded -- cannot consent or withdraw

Why is this research method’s approach to studying behaviour generally considered qualitative (even if you gather some quantitative data as part of the process)?

* depends on the interpretations of the researcher.
* Quantitative data is often calculated through standardized formulas and algorithms however, observations require researchers to decide which information is worth noting and how that is interpreted and generalized into data.

CT naturalistic observations

* ethical: lack of consent and withdrawal
* bias and subjectivity
* validity: cannot stop extraneous variables
* Hawthorne effect: participants change to appear socially desirable

Interviews key features

* interview includes at least 1 interviewer and another interviewee
* semi structured questions
* qualitative
* interviewer must be able to establish a good rapport, have people skills & understand biases that may occur

survey key features

* collects self-reported data
* may be on a large scale
* can do both quantitative and qualitative
* structured questions

interview vs survey

* self-reported vs with interviewer
* semi vs fully structured questions
* interviews may be a focus group
* interviewers may lead with another action (look at pic)

How/why do surveys offer a more ethical approach to studying sensitive topics?

* no pressure to answer questions participant does not want to
* does not have to come face to face admitting something
* lowers social desirability bias

Why would a survey be used after an interview to help generalize?

it would all be the same question which can yield similar answers like all numbers which can be easily quantifiable

The biggest issue with self-reported data is the possibility that it isn't accurate. In what ways can answers be biased/ altered by:

* participant can lie or not tell full truth
* the researcher can misinterpret

CT interviews and surveys

* ability to generalize may be good if a large enough sample population
* interviews difficult to generalize if different questions
* validity affected by researcher and participant bias
* surveys can lessen participant pressure
* Wording of questions may not sound objective or may hint at a desired answer that causes the participant to answer in a way influenced by the question

quasi experiment

has an IV and DV, in which the IV is not manipulated by the researcher, but rather, the IV itself had already **pre-existed**."

natural experiment

has an IV and DV, in which the IV is produced by **environmental or external factors** that had occurred over time, and consists of a **before** and **after** period of observation."

CT natural & quasi

* **Extraneous variables**: (aka confounding variables) are undesirable variables that influence relationship of IV and DV
* **Demand characteristics:** participants act differently because they know they are in a study, may try to guess aims
* time
* expense
* Access to prospective target populations

4 points when discussing ethics

* what is the ethical issue?
* which study?
* where and how does the ethical issue arise?
* cost-benefit analysis