research methods and designs (class 1)

Name the five criteria for good measurement

1. Direct relevance to hypothesis

2. inter-rater reliability

3. test- retest reliability

4. internal validity

5. external validity

describe direct relevance to hypothesis

Ask the question: Does the measurement actually relate to what the hypothesis is asking?

The measurement directly measures what is being asked in hypothesis

ex: if the hypothesis is about a child’s preference for a certain toy, measuring which toy they choose is relevant. Measuring how fast they can run isnt

describe inter-rater reliability

Ask the question: do different observers/raters get similar results when measuring the same thing?

Consistency across people’s answers

PEOPLE

describe test-retest reliability

ask the question: does the measurement produce consistent results over time

Consistency across time of a measurement two or more times

TIME

describe internal validity

validity=accuracy/truthfulness of what a measure is studying/saying INSIDE the study

ex:

Researchers want to know whether sleep affects test performance.

Group A sleeps 8 hours.
Group B sleeps 4 hours.

BUT Group A also gets an easy test while Group B gets a hard test. 🚨

Now we don't know whether the difference in scores came from:

sleep OR test difficulty.

That hurts internal validity because something else could explain the result.

describe external validity

ask question of: can the findings generalize beyond this particular study like to other peope, settings, or situations

OUTSIDE

what does the 5 criteria of good measurement help with

its basically a checklist researchers use to judge whether their measurements and conclusions are reliable, valid, useful (double checking)

whats the difference between reliabiity and validity

reliability- is it reproducible (if i do this test on different peeps do i get the same result)

Validity- does our measure have truth to it? is it general can it apply to multiple people/situations?

What is naturalistic observation

where researchers observe childern in their natural, everyday environments without creating or changing the situations

what would be a benefit/cost for naturalisitc observation

benefit: you are seeing what children naturally do

cost: little control like you can’t control what happens or guarantee the behavior you are interested in will occur

what is structured observation

researcher creates the situation and observes how child responds to it

Benefit:

more control + easier to compare because everychild experiences the same sitution

less natural:

children may act differently in an arranged/research setting then they normally would

what is interviews and questionaries

researchers ask questions to gather info on child’s experiences, thoughts, feelings, or behavior

name and describe the two important sources/types of interviews and questionaries

• self report: the child reports about themselves

• other report: someone who knows the child reports about the like teacher or parent

name the pros and cons of a interview/questionaries

benefit: you can learn about thoughts, feelings, experiences, and behaviors that can be difficult to observe

cost: reports can be inaccurate or biased

name the three approaches for gathering data

1. naturalistic observation

2. structred observations

3. interviews/questions

what should you be asking yourself when choosing one of these types for gathering info

DON’T ASK: which one is better if they all have cons

ASK: which method is best for my research and what weakeness am i willing to accept

explain correlation:

a relationship/association between two variables; does not = casualtion

like these to things go together, but i can’t tell you excatly why cause there can be so many factors

why doesnt correlation prove causuation

because we dont know whether A causes B , or if B causes A or a third varible influences both

why are correlation studies common in child development

because many variables cannot ethically or pratcially be mainpulated so researchers study naturally occruing differences. this can sho relationship/ but not causuation

name an a design that can show causaiton

experimental design because its more like i chnaged ths and tested whether that chnaged bacuse of it

why doesn’t correlation prove causation

• direction of causation

• third variable problem

describe direction of causation

we dont know which varaible cuases which like does A cause b or does B cause A

describe third variable problem

an outside varible may be influencing both A and B which explains why they are correlated

what are the two types of research desgins

• longitudinal

• cross sectional

describe crossectional

• compare different age groups at the same point in time

name a benefit and cost for corss sectional

benefit: fast + cheaper you dont have to wait years for chid to grow up

cost: doesnt show how the SAME child changes over time

describe longitudinal

same people different times , follows the same person over an extended period of time

name a cost and benefit for longitudinal

benefit: shows actual development change within the same people

cost: takes a long time+expensive+people may drop from study

explain triangulation

Using multiple ways/sources to study the same thng to see if they point to the same conclusion

• checking something from different angles because if all angles/sides point to the same answer, then I will have a more stronger/trustworthy evidence

why use triganulation

every method has a weakness so using multiple sources helps compensate for those weaknesss

different methods like different ways of collecting info